Abstract of 29th Regional Congress of the ISBT (2023)

SOUND OF MUSIC: QUANTUM MECHANICS OF INNOVATION IN AUSTRIA

C Gabriel^1,2

¹Department for Blood Group Serology and Transfusion Medicine, Medical University Graz, Graz ²Institute for Traumatology, Ludwig Boltzmann Society, Vienna, Austria

Austria had for centuries a rich history of culture. The growing scientific community in the fin de siecle was heavily concentrated in Vienna. Freud, Boltzmann, Schrödinger and Mach might be the first names to find, whenever one cites Austrian scientists. But more related to transfusion are the Noble Prize winners Max Perutz and Karl Landsteiner. Landsteiner′s fate illustrates the brain drain beginning in the early 30s escalating in 1939 with the “Anschluss”, which lead to the forced emigration of many scientists. A loss which was not regenerated in the post war years and was further aggravated by dubious and often undisclosed relations and scandals in the Nazi‐era. All together this leads to a severe loss of credibility and productivity of universities across decades. Opening university access in the early 80s and intensive historical work‐up of scandals transformed the Austrian universities to open and effective scientific institutions driving innovation in the country.

Austria has achieved a great economic deal in recent decades, which was accelerated by the EU membership in 1995. As a result of strong long‐term economic performance, the country's gross domestic product (GDP) per capita is the eighth highest among OECD countries and fourth in the EU28. Levels of poverty and income inequality are both below the OECD average. Investment in research and development (R&D) increased since the EU accession, when Austria's R&D intensity (aggregate R&D expenditure as a percentage of GDP) was well below the OECD average and significantly far lower than Switzerland ‐ a country to which Austria prefers comparison. The EU target of 3% R&D intensity was first met in 2014 and is 2018 the sixth highest among OECD countries and the second highest in the EU28. Austria showed the second highest increase in R&D intensity of all OECD countries, exceeded only by Korea.The rapid expansion was matched by a similar increase in human resources and scientific output of universities. Austrian science in quantum mechanics, quantum communication and information is world renown. Vienna is a major biotech hub, as is Linz in mathematics and mechatronics and Graz in automotive and production technologies.Austria has been a net resource recipient in the Horizon 2020 and the preceding 7th Framework Programme. Small and medium‐sized enterprises show a high propensity to co‐operate with universities and other research organisations and more and more included in scientific grant schemes.

Vienna is the largest student city in the German‐speaking world and consistently ranks among the top cities in the world on quality‐of‐life indices. As Austria possesses globally recognised cultural attractions ranging from famed Salzburg festival to the Vienna New Year concert its inhabitants are not aware of the progress made in R&D and how thriving innovation is going on in their country. They still love to show their cultural heritage and events and impress the world with some kind of eternal Sound of Music.

INNOVATION IN CELLULAR THERAPY: HYPE AND HOPE

N Worel

Transfusion Medicine, Medical University Vienna, Vienna, Austria

Patients with refractory B‐cell malignancies as non‐Hodgkin lymphomas (NHL) resistant to standard therapieshave a dismal prognosis. The outcome is even poorer in patients relapsing after autologous stem cell transplantation. Most of these patients do not qualify for an allogeneic hematopoietic cell transplantation (HCT) due to refractory disease, lack of a suitable allogeneic donor, higher age or cumulative toxicity of previous chemotherapy. Despite patients undergoing allogeneic HCT normally profit from a graft‐versus ‐lymphoma effect, overall survival in patients with NHL after HCT remains short. A similar situation can be observed for patients with acute lymphoblastic leukemia (ALL). Therefore novel treatment modalities are urgently needed.Chimeric antigen receptor (CAR)‐T cells, a new class of cellular immunotherapy involvingex vivogenetic modification of T cells to incorporate an engineered CAR have been used in clinical trials. In the majority of studies B‐cell malignancies treated with CD19 targeting CAR‐T cells have been analyzed. Austria had the advantage to participate in two international trials in the past and is currently involved in further CAR‐T studies.

Recently, results from CD19 directed CAR‐T cell trials with an increased follow‐up of patients led to FDA (Food and Drug Administration) and EMA (European Medicines Agency) approval of tisagenlecleucel and axicabtagene ciloleucel. Common adverse events (AEs) include cytokine release syndrome and neurological toxicity, which may require admission to an intensive care unit, B cell aplasiaand hemophagocytic lymphohistiocytosis. These AEs are manageable when treated by an appropriately trained teamfollowing established algorithm. In this presentation, results of four large phase II CD19CAR‐T cell trials for patients with NHL and ALL and focus on AEs is summarized.

INFLUENCE OF ANEMIA ON OUTCOME IN PATIENTS UNDERGOING ORTHOTOPIC LIVER TRANSPLANTATION

D Baron

Department of Anesthesiology and Intensive Care Medicine, Medical University of Vienna, Vienna, Austria

Preoperative anemia is a known risk factor for increased perioperative morbidity and mortality in patients undergoing major surgery. Previous studies have not only shown higher in‐hospital mortality, but also an increased hospital length of stay, greater postoperative admission rates to intensive care and prolonged use of mechanical ventilation and intensive care resources in patients with anemia compared to those with normal preoperative hemoglobin concentrations. About 30% of patients scheduled for major surgery suffer from preoperative anemia. This figure is even higher in patients requiring orthotopic liver transplantation, where up to 75% of all patients are diagnosed with anemia prior to surgery.

Transfusion of packed red blood cells (PRBCs) is commonly used to correct anemic hemoglobin values. However, transfusion of PRBCs has been associated with increased morbidity and mortality in patients undergoing cardiac, orthopedic, and abdominal surgery. Additionally, transfusion of PRBCs is associated with a greater incidence of postoperative acute kidney injury in patients undergoing orthotopic liver transplantation. As preoperative anemia might increase the perioperative use of PRBCs, negative effects observed after PRBC transfusions might even be augmented.

Data on the influence of preoperative anemia on morbidity and mortality after orthotopic liver transplantation are limited. Thus, we retrospectively analyzed the association of preoperative anemia and mortality in adult patients undergoing orthotopic liver transplantation at our institution. In addition, we examined the influence of anemia on perioperative parameters such as transfusion requirements, surgical complications, early allograft dysfunction, acute kidney injury, and the need for renal replacement therapy. Based on the results obtained in the retrospective analysis, an ongoing prospective randomized clinical trial was initiated.

TRANSFUSION MEDICINE FOR SICKLE CELL PATIENTS: FROM BLOOD DONOR ACCRUAL TO GENE THERAPY

F Pirenne

Hôpital Henri Mondor, Etablissement Français du Sang, Créteil, France

The two suspensive treatments in sickle cell disease (SCD) are hydroxycarbamide, inducing the production of the functional HbF, normally repressed at birth, and Red blood cell (RBC) transfusion, a critical component of SCD management. However, RBC transfusion is not without risk. Repeat exposure to allogeneic RBCs can result in the development of RBC alloantibodies which can make it difficult to find compatible RBCs for future transfusions. However, the main concern of alloimmunization is the development of hemolytic transfusion reaction, with in the most severe cases, hyperhemolysis, leading to multi organ failure and death in 4% of the cases. The prevention of this life threatening condition must be based on risk factors. However, although some risk factors, such as alloimmunization, have been identified, much of the mechanism underlying DHTR remains a mystery, particularly in severe cases presenting hyperhemolysis. Here we will describe the current and future development to prevent and treat this severe syndrome

In order to decrease exposure to transfusion in SCD but also improve red blood cell quality, some new products are developed. Oxidative damage is one of the parameter that could be diminished. Some work is also ongoing to prevent filter blockage during leucodepletion of precious RBCs units from Afro‐Caribbean donors carrying the sickle cell trait. Finally, in countries with higher risks of transmission of infectious disease, treatment of red blood cell units against infectious agents can be discussed.

The only current curative treatment of SCD is hematopoietic stem cell transplantation (HCS). However, the occurrence, frequency, and effects of immune hematologic complications in HCS remain and will be discussed. Finally, gene therapy is a real hope as a definitive curative treatment. Clinical trials are ongoing in France and will be discussed as well as the remaining place of transfusion in this therapeutic.

IMMUNOTHERAPY CART‐CELLS: FROM TARGET IDENTIFICATION TO THE CLINIC

C Ferrand¹, W Warda¹, F Larosa², M Deschamps¹

¹Research, INSERM UMR1098 ‐ EFSBFC ‐ UBFC, BESANCON CEDEX ²Clinical, Hematology, Besancon, France

Based on the clinical success of Chimeric Antigen Receptor engineering T‐cells (CART) targeting CD19 in B‐malignancies, such as Acute Lymphoid Leukemia, lymphoma but also demonstrated now in Multiple Myeloma, CART‐cells Immunotherapy is become one of the most promise alternative for patients in refractory/relapsed hematological diseases.

In the context of the Chronic Myeloid Leukemia (CML), we have hypothesized that quiescent Leukemic Hematopoietic Stem Cells (HSC) compartment, escaping to the current tyrosine kinase inhibitors (TKis) treatment, in part associated in the molecular relapse, may be targeted by CART‐cells immunotherapy. Gene expression profiling studies have established that a cell surface biomarker IL‐1RAP is expressed by the leukemic but not by the normal CD34+/CD38‐ HSC.

This talk will focus of the whole process of development of a CART‐cells starting from recombinant IL‐1RAP protein mice immunization to produce a specific monoclonal antibody (mAb), to the proof of concept demonstration, before moving into the clinic.

We produced and selected a specific anti‐IL‐1RAP mAb (#A3C3 clone, Diaclone SA, Besançon, France). After molecular characterization of antigen‐binding domain, nucleotide sequences were fused with 3rd generation T cell activation coding sequences and cloned as a single chain into a lentiviral backbone comprising a safety switch suicide gene iCASP9 (inducible Caspase 9) and a monitoring/selection cell surface marker ∆CD19.

We demonstrated in‐vitro and in an in‐vivo xenograft murine model that IL‐1RAP CAR T cells can be activated in the presence of IL‐1RAP+ cell lines or primary CML cells, secrete pro‐inflammatory cytokines, degranulate and specifically killing them. We also demonstrated that multi‐TKIs treatment over a 4‐year period does not affect transduction efficiency of CML patient T‐cells by IL‐1RAP CAR vector and that autologous CART‐cells are able to target IL‐1RAP+ leukemic primary HSC.

“Off‐tumor‐on target” toxicity prediction, by studying IL‐1RAP expression on a tissue macroarray comprising 30 normal human tissues (3 donors), with #A3C3, detected various IL‐1RAP intensity staining in only few tissues. Regarding the healthy hematopoietic system, #A3C3 flow cytometry staining did not detect hematopoietic cells, except monocytes that express poorly IL‐1RAP. As expected, monocyte subpopulation is targeted by autologous IL‐1RAP CART cells (ratio E:T=1:1), but at a lower level that IL‐1RAP CML cell line.

In‐vivo investigation of specific toxicities of autologous IL‐1RAP CART‐cells against HSC and/or immune cells on a human‐CD34+cord blood cell engrafted/NOG murine model, but also by an in‐vitro CD34+colony forming unit assay didn't reveal any significant toxicities in immunocompetent cell subpopulations, suggesting that healthy CD34+HSC are not affected.

Finally, to overcome potential toxicity, functionality of the iCASP9/ Rimiducid^® safety switch was demonstrated in‐vitro but also in‐vivo in a NSG tumor xenograft model, showing that, when activate, the system is able to eliminate more than 90% of CART‐cells, after exposure to AP1903.

In conclusion, based on CML model, we demonstrated that IL‐1RAP is an interesting target for CART‐cell immunotherapy, with a limited “on target, off tumor” predictable toxicity. Next step will be the up‐scaling of the process in order to match with the use in human regarding also the regulatory requirements. This strategy may be applied, in the future, in other hematological malignancies.

TRANSFUSION IN THE ACUTE BLEEDING SETTING: MASSIVE TRANSFUSION PROTOCOLS, INTRODUCING NEW “OLD” BLOOD PRODUCTS SUCH AS WHOLE BLOOD

S Ausset¹, P Tiberghien², S Gross², S Begue², T Pouget³, C Martinaud³

¹Anesthesia, intensive care and emergency medicine, Val de Grâce military medical academy, Paris ²EFS, Saint Denis ³CTSA, Clamart, France

Mortality ranges from 20 to 30% for trauma victims with severe bleeding and is largely dependent on the transfusion therapy from which they can benefit. The nature of this therapy has an impact on prognosis with a halving of mortality when the plasma/pRBC ratio is greater than ½ and a decrease of about 20% when the proportion of platelets transfused is close to that of whole blood. The speed with which such therapy is actually administered has a major impact as well with an increase in mortality of 5% for each minute of delay in making the entire therapy available. This can be explained mainly by the fact that the probability of death of these patients is greater within minutes of their admission to hospital with a median time to death of 2h after admission. To allow plasma, platelets and pRBC to be made available in a timely manner, North American trauma centers have mandated that trauma centers have massive transfusion packs at the patient's bedside within 15min. To further simplify and speed up logistics from distribution to transfusion, several trauma centers now use whole blood stored at 4°C. This return of an “old” product is largely inspired by military experience where whole blood is mainly used “warm” immediately after collection with compelling evidence of its effectiveness. Its return to civilian practice requires the ability to deleukocyte it while preserving platelets and to store it while maintaining their hemostatic functions. Good quality data shows this is achievable and several clinical studies are planned to begin in the coming months. In France, the French Blood Establishment and the French Army are cooperating to initiate the prospective randomized non‐inferiority STORHM trial (Sang Total pour la Réanimation des Hémorragies Massives) which will be comparing whole blood to separate blood components in an 1/1/1 ratio in severely bleeding trauma patients. The primary endpoint will be a thromboelastographic parameter (maximum amplitude) assessed at the 6th hour after admission. Secondary endpoints will include early and overall mortality, lactate clearance (reflection of the effectiveness of resuscitation) and 24h organ failure. This trial will be recruiting 200 patients in 6 French trauma centers and is planned to be initiated second half of 2019.

MESENCHYMAL STROMAL CELLS FOR REGENERATIVE THERAPY

H Schrezenmeier*¹, M Kalbitz², E Amann¹, R Lotfi¹, M Rojewski¹

¹Institute of Transfusion Medicine, Institute of Clinical Transfusion Medicine and Immunogenetics Ulm, GRC Blood Transfusion Service Baden‐Württemberg‐Hessen and University Hospital Ulm ²Department of Traumatology Hand‐, Plastic‐, and Reconstructive Surgery, University Hospital Ulm, Ulm, Germany

Mesenchymal stroma cells (MSC) can be used for immunomodulatory and regenerative treatment. These Advanced Therapy Medicinal Products (ATMP) can be obtained from various tissue sources (bone marrow, adipose tissue, cord blood, placenta and others) and various donor types (autologous, allogeneic). We previously demonstrated that bone marrow derived MSC (BM‐MSC) are a powerful tool for in vivo formation of bone and treatment of bone defects. However, adipose‐derived stem cells (ASC) seem to be more powerful in treatment of acute or chronic inflammation, especially when degeneration of tissue is involved (e.g. osteoarthritis). The evaluation of MSC in clinical trials has substantially increased during the last decade.

We will review the development of large‐scale GMP‐grade ex vivo expansion protocols for MSC from bone marrow (BM‐MSC) and adipose‐derived stem cells (ASC) which were developed, optimized and standardized in several consortia within the Programme of the European Commission (7th Framework Programme: CASCADE, REBORNE; HORIZON 2020 Programme: ADIPOA‐2; ORTHOUNION, MAXIBONE). A two‐step protocol allows the ex‐vivo expansion of>200×10⁶ cells (passage 1) within 2–3weeks (Rojewski et al., Cytotherapy 2019, PMID 30926359). The protocols have been designed free of xenogeneic components. Proliferation of MSC is stimulated by a growth factor preparation derived of human platelet concentrates (human platelet lysate; hPL) which is rich source of the factors PDGF, TGF‐ß and bFGF which are essential for the growth of MSC (Fekete et al, Cytotherapy 2012, PMID 22296115).

BM‐MSC / ASC obtained from these protocols have been characterized in detail in pre‐clinical evaluations. Manufacturing licenses for MSC and ASC and a platelet‐derived growth factor concentrate have been obtained and they have been explored in several clinical trials for treatment of bone defects (ORTHO‐CT1: EudraCT Number: 2011‐005441‐13; ORTHO‐CT2: EudraCT Number: 2012‐002010‐39; MAXILLO1: EudraCT Number: 2012‐003139‐50). We will summarize results of completed clinical trials which confirmed feasibility and safety of autologous MSC /ASC treatment and provided evidence for efficacy (Gjerde et al, Stem Cell Res. 2018, PMID 30092840; Gomez‐Barrena et al, Biomaterials 2019, PMID 29598897) ‐ also providing the rationale for ongoing prospective trials comparing MSC treatment with standard of care (ADIPOA2: EudraCT Number: 2015‐002125‐19 and ORTHOUNION: 2015‐000431‐32).

Finally we will discuss the impact of donor type, use of bioreactors for ex‐vivo expansion and pre‐clinical evaluation of MSC in new indications, e.g. for acute intervention in severe trauma (Amann et al., Cytotherapy 2018, PMID 29223534; Amann et al., PloS One 2019; e0216862).

IN VITRO PRODUCTION OF MEGAKARYOCYTES

C Figueiredo, R Blasczyk

Institute for Transfusion Medicine, Hannover Medical School, Hannover, Germany

Introduction: In vitro produced Megakaryocytes (MKs) may serve as source to produce platelets (PLTs) ex vivo or in vivo. We have established a strategy to differentiate MKs from induced pluripotent stem cells (iPSCs) in bioreactors. This study aimed at the large‐scale production of MKs using microcarriers to increase the MK yield and to characterize their phenotype and functionality after irradiation as a method to decrease possible safety concerns associated to the iPSC origin.

Methods: iPSCs were cultured in an aggregate form in presence or absence of microcarriers using 50mL stirred flasks. Cells were differentiated into MKs using TPO, SCF and IL‐3 in APEL2 medium for a period of 22days. Non‐ irradiated or irradiated iPSC‐derived MKs were analysed for polyploidy, phenotype and proPLT production using flow cytometry and fluorescence microscopy. Also, PLT‐production was investigated in vivo. Non‐irradiated or irradiated MKs were transfused to NOD/SCID/IL‐2Rγc–/– mice and blood was analyzed for human PLTs.

Results: Differentiation of MKs in presence of microcarriers resulted in an 8‐fold increase of MKs per iPSC in comparison to only aggregates. This resulted in mean of total MK harvest of 18.7±6.8×10⁷ in microcarrier‐assisted bioreactors in comparison to 4.9±1.3×10⁷ MKs collected from bioreactors containing only aggregates. Interestingly, MKs produced in microcarrier‐assisted bioreactors showed higher proPLT formation capacity than MKs derived from only aggregates bioreactors. MK phenotype and DNA content was comparable between MKs derived from both types of bioreactors. Irradiation of MKs did not affect their phenotype and capability to form proPLTs or PLTs after transfusion into NOD/SCID/IL‐2Rγc^–/– mice.

Conclusion: Microcarriers showed to significantly increase the yield of iPSC‐derived MKs in stirred bioreactors to clinically relevant numbers. This may facilitate the use of iPSC‐derived MK for ex vivo production of PLTs, direct transfusion or for innovative MK‐based regenerative therapies.

WE ARE STARDUST – WHAT COMETS TELL US ABOUT OUR ORIGINS

K Altwegg

Space and Planetology, University of Bern, Bern, Switzerland

Although the Rosetta stone, found by the troops of Napoleon in Egypt near the city of Rosetta (Rashid) contains only a small amount of text in three languages it was key in deciphering Hieroglyphs. The Rosetta mission tried to achieve something similar: by looking at a tiny body its goal was to decipher the origin of the solar system, planets including Earth and life. After more than 12years the Rosetta spacecraft softly crash‐landed on comet Churyumov‐Gerasimenko on September 30, 2016. It has travelled billions of kilometers, just to study a small (4km diameter), black boulder named 67P/Churyumov‐Gerasimenko. The results of this mission now seem to fully justify the time and money spent in the last decades on this endeavor. Where are we from? Where are we going? Are we alone in the Universe? these are some of the big questions. In this talk I will show which answers we got from Rosetta and comet Chury. We follow the pathway of the material which makes up our solar system from a dark cloud to the solar nebula and finally to planets and life. I will show that indeed we are the result of stardust and that what happened here may happen elsewhere in the Universe.

THE SKY'S THE LIMIT – NEXT GENERATION ENGINEERED CELLULAR THERAPIES

L Jeker

Department of Biomedicine, Basel University Hospital and University of Basel, Basel, Switzerland

Cells, tissues and entire organs can collectively be seen as “living drugs”. Genetically unaltered cells are routinely used in clinical practice to treat diseases as diverse as anemia, bleeding disorders, leukemia and organ failure. Ground‐breaking advances in genetic and genome engineering technologies are propelling cell therapies to the frontline of Medical research and practice. The hematopoietic system is particularly amenable to genetic engineering because specific cell types can be purified based on the expression of specific surface proteins and the ability to culture and expand cells ex vivo.

The recent unprecedented clinical success of killer T cells reprogrammed by chimeric antigen receptors (CARs) to attack CD19 expressing tumor cells demonstrates the power of immunotherapy with genetically engineered immune cells. However, given the rapid development of novel genome engineering and synthetic biology tools we are likely only at the beginning of a new era of engineered cellular therapies. I will present recent progress in immune cell reprogramming, gene correction, safety aspects and remaining challenges such as manufacturing.

CELL‐FREE NUCLEIC ACIDS IN TRANSFUSION MEDICINE

S Waldvogel

Department of Medicine, University Hospital of Geneva, Geneva, Switzerland

Cell free nucleic acids (CFNA) circulate in the plasma of all individuals and are thought to be released by host and foreign cells into the circulation. After fractionation by centrifugation, CFNAs can be extracted from the supernatant of whole blood samples or manufactured blood products. These DNA or RNA sequences can be of human, bacterial, viral or fungal origin. Most of them are human double stranded DNAs. Research on CFNAs is increasing, thanks to technological advancements in molecular biology. Some of their results are already implemented in clinical practice in the areas of pre‐natal diagnosis, oncology and infectious diseases. The latter investigation focuses on the exploration of non‐human CFNAs, the field of metagenomics. High throughput sequencing associated with bioinformatics, the so‐called new generation sequencing (NGS), has sped up the investigations of non‐human CFNAs. This tool provides the opportunity to classify CFNAs into a human or non‐human category, and then to identify them. It is thus possible to explore simultaneously the whole landscape of bacterial, viral and fungal populations. Presently, NGS of human blood has already proven its feasibility and its value in identifying emerging viruses or investigating clinical cases of fever of unknown etiology. NGS of CFNAs is also particularly effective in analyzing the different genotypes of a virus in case of a co‐infection (e.g. Hepatitis C virus). Studying CFNAs with the new molecular technologies is therefore of great importance in transfusion medicine, especially regarding security and clinical transfusion reactions. First, transfusion transmitted infections are the most feared adverse complications. Second, febrile non‐hemolytic transfusion reactions are also the most frequently reported adverse events in hemovigilance systems and their physiological mechanism –if only one‐ remains not clearly elucidated. Investigating CFNAs could thus improve our understanding and strategy aiming at reducing those two clinical adverse events. Surveying comprehensively the composition of circulating infectious agents in a blood product by NGS technology could be very interesting for investigating a severe febrile transfusion reaction. Moreover, when the costs of analysis will be reduced, it might be possible to screen prospectively and regularly the whole metagenomics of asymptomatic blood donors, in addition to the classical epidemiological surveillance. For instance, in a study testing a NGS method on manufactured fresh frozen plasmas, an astrovirus (MLB2) has been identified. Finally, it is the responsibility of transfusion physicians implicated in the manufacturing of blood products to ensure that CFNAs within a blood product do not have a clinical impact on the innate immunity of the recipients. According to recent research in vitro, CFNAs purified from blood products can induce the transcription of inflammatory cytokines by mononuclear cells. As non‐human CFNA have an effect on Toll‐Like Receptors (TLR‐linked inflammatory pathways), it would be also relevant to insure that donor's CFNAs have no significant effect on the immune system of the recipient. In conclusion, CFNAs are very diverse molecules contaminating blood products. Technological progress makes now their investigation more available. Besides being useful markers of infection in asymptomatic donors, their impact on the recipients’ immunity should be further investigated.

IMPACT OF BLOOD DONATION ON EXERCISE TOLERANCE

JW Helge

Department of Biomedical Sciences, University of Copenhagen, Copenhagen, Denmark

An active life and regular training is part of a healthy life style and for many this includes participation in endurance exercise competition at different levels. Thus, it is highly relevant to know how a blood donation affects exercise performance and how close this can done to an endurance competition. Endurance exercise performance is determined by many factors, but three of the primary are maximal oxygen uptake, the relative load that an individual can sustain over time and finally the efficiency of movement in the given discipline.

Over the years, a number of studies have sampled blood volumes ranging from 450–1200ml and applied different methodological approaches to measure maximal oxygen uptake over a recovery period ranging between 3–28days. Overall, the general finding is a reduction in blood haemoglobin and an attenuated maximal oxygen uptake as well endurance performance after blood donation. In normal to well‐trained men maximal oxygen uptake and performance was normalized after two weeks in one study after a normal blood donation (450/470ml), but remained attenuated after four weeks in another study, despite the change in blood haemoglobin concentration was similar and the design and methodology also similar in the two studies.

In addition to maximal oxygen uptake the relative load that can be tolerated during exercise is probably also attenuated, through a decreased arterio‐venous oxygen extraction, but the available data is very limited. The first part of this talk will highlight the major findings and discuss some of the methodological issues that complicate interpretation and conclusions.

There are sex differences in circulating blood volume, haemoglobin concentration, haematocrit and hormone levels and thus it is entirely possible that there is a sex difference in the effect of blood donation on physical performance and the recovery after blood donation. In addition to the basic physiological sex differences, there is also a higher prevalence of iron deficiency in premenopausal women, physically active women and women donating blood. Therefore, we studied the influence of a standard 450mL blood donation on maximal oxygen uptake and endurance performance and the subsequent recovery in physically active women. We observed that in iron sufficient women blood haemoglobin concentration and maximal oxygen uptake were back to baseline 28days after blood donation, but endurance performance was normalized already after 14days. The second part of this talk will discuss the sex differences in the effect of blood donation on maximal oxygen uptake and endurance performance.

Overall, the available data suggest that, with a careful conservative approach, 4 – 5weeks are needed after a normal blood donation to be fully recovered to participate in endurance exercise competitions.

HOW DOES DEFERRAL IMPACT THE BLOOD DONOR?

TE Davison¹, B Masser², C Gemelli¹

¹Clinical Services and Research, Australian Red Cross Blood Service, Melbourne ²School of Psychology, The University of Queensland, Brisbane, Australia

More than one in ten attempts to donate blood result in a temporary deferral, due to concerns about the impact of the donation on the donor or recipient. There is well established evidence that temporary deferrals impact negatively on donors, with a large proportion of those deferred failing to return at the end of the deferral period. This presentation provides an overview of deferrals from the donor perspective, describing the likelihood of receiving a deferral for different donor subgroups. The impact of temporary deferrals on the future donation of donors, considering both short‐term and longer‐term donation patterns, will also be reviewed, outlining which donors are at highest risk of non‐return following a deferral, and what is known about the accumulative impact of multiple deferrals on donors. Several hypotheses have been proposed to account for the strong negative impact of deferrals on donor behaviour, and there is preliminary evidence of psychological factors, such as emotional reactions, predicting intention to return. Research is also beginning to emerge on the effectiveness of tailored interventions to mitigate the impact of deferrals on donor behaviour. The evidence for these preventative interventions, and for strategies to reactive donors once they have lapsed post‐deferral, will be reviewed. Recommendations for blood centres will be made, as well as suggestions for future research to address continuing gaps in knowledge.

THE GIFT OF LIFE – DOES IT APPLY TO DONATION FOR RESEARCH?

V Raivola

Finnish Red Cross Blood Service, Helsinki, Finland

In his influential study “The Gift Relationship” (1970), Richard Titmuss coined the idea of voluntary, non‐paid, blood donation being the gift of life for a fellow citizen. This metaphor has been powerful in mobilizing donors (Busby 2004). It conveys a direct relationship between blood donation and patients’ vitality, as well as a difference between gains and costs. As the gift of life, blood donation is seen to symbolize pure altruism and promoting solidarity between strangers. But can we apply the metaphor as successfully into donating blood for research? We asked a group of Finnish blood donors what they would think if the FRC Blood Service invited them to give a blood sample and personal information for research. The blood donors were usually willing to contribute to research for the public benefit, because they saw great potential in science to create solutions to help patients in the future. However, based on our interview data and previous research, we suggest that the analogy between gift of life and donation for research did not work all the way. The metaphor fails to address donors’ questions on new types of relationships, interests and risks related to the use of personal data for research. Left unanswered these could discourage donating for research. Hence, we argue that the gift of life metaphor is not applicable to donor recruitment at the research context. In this presentation we wish to look for a better metaphor for donation for research that blood services collecting research data could apply.

IMMUNOHAEMATOLOGICAL FEATURES OF PATIENTS WITH SICKLE CELL DISEASE

V Thonier

CNRGS, Institut national de la Transfusion sanguine, Paris, France

Sickle cell disease (SCD) is the most prevalent genetic disorder in France, where it affects mostly people of African descent or individuals from the French West Indies. Many other countries are also affected, and more recently, with increasing migration, SCD has become a worldwide public health issue.

Transfusion is still a key treatment for SCD patients; it is used either as curative or preventive therapy. As a result, they are much more exposed to transfusions than the general population. In 2017, according to the French hemovigilance system, in the general population the transfusion rate was 0.78%. Different cohorts of SCD patients report significantly higher transfusion rates, varying from 18% (in S/C or S/β⁺ patients) to 98% (in SS or S/β⁰ patients). Another difference, with the general population is how early in life SCD patients are transfused.

Consequently, SCD patients are much more exposed to alloimmunization and delayed hemolytic transfusion reactions (DHTRs), the latter being the most feared complication. In certain situations, defined as hyperhemolysis, autologous RBCs are also targeted and destroyed. This can put the patient in a life‐threating situation. Health care professionals should therefore be familiar with the main immunohaematological features of SCD patients and of DHTR.

Alloimmunization happens where there are discrepancies between the recipient's and donor's RBC phenotypes. The typical phenotype of SCD patients is D+C−E−c+e+, K−, Fy(a−b−), Jk(a+b−), S−s+. The distribution of the alleles encoding this phenotype is almost the reverse in Africa and Europe. Some so‐called low frequency antigens (LFA) should be considered as polymorphic antigens in the African population and these LFA are not present in most commercial panels. The situation is even more complicated when recipients lack high‐frequency antigens, the most common ones being Hr‐, Hr^B‐, Sec‐, Uneg, U+^var, Js(b‐), (Hy‐), and Jo(a‐). Finally, there is a high Rh diversity among people of African descent. Because they harbor variant alleles and/or partial RH antigens, they are at risk of developing alloantibodies. In this setting, screening for partial RH antigens makes sense. The figures illustrating this diversity vary with the approach used. One of them is to take into consideration RHD or RHCE*ce variant alleles. In several American studies, their prevalence was estimated to be 29–36% and 53–72%, respectively. Other teams take into consideration D, C and e partial antigens. Their prevalence was estimated to be 8.4–14%, 12.5–27.7%, 3.3–3.5%, respectively, and the alloimmunization rates were 17.6%, 14.3–30%, 7.1%, respectively.

As a result of these phenotype discrepancies, SCD patients are more likely to be alloimmunized. An overall immunization rate of 2–6% is commonly admitted in the general population. Depending on the unit selection policy and/or the study design, the immunization rate in SCD patients varies from 7% to 76%, the highest figures being established when an ABO/RH1‐only matching policy is implemented. In a meta‐analysis of 24 publications, the overall alloimmunization rates were around 20%.

Alloimmunization is thought to be enhanced by an inflammatory state, which is often present in SCD patients. They are more prone to develop a new alloantibody. Using a stochastic modeling of alloimmunization, they have a 61% increased risk of producing additional antibodies versus 30% in the general population.

Autoantibodies have been identified as a risk factor of alloimmunization. As a result, SCD patients often have complex mixtures of allo and autoantibodies. RH antibodies and those considered as irregular natural antibodies are present in a significant proportion.

Another characteristic of the antibodies in SCD patients is their evanescence; up to 70% of alloantibodies become undetectable within a few years of their initial development. Relatedly, about a third of DHTRs are reported to happen in patients with no previous history of immunization. In addition, a third of patients will not develop an antibody after a DHTR.

Identifying patients at risk of developing a DHTR is key to managing them properly.

PHENOTYPE AND ALLELE MATCHING: HOW FAR SHOULD WE GO?

C Folman¹, E van der Schoot², M de Haas^1,3,4

¹Immunohematology Diagnostics, Sanquin Diagnostic Services ²Experimental Immunohematology, Sanquin Research and Landsteiner Laboratory, Amsterdam University Medical Centre, Amsterdam ³Center for Clinical Research, Sanquin Research ⁴Department of Immunohematology and Blood Transfusion, Leiden University Medical Center, Leiden, Netherlands

Alloimmunization is a serious risk of red blood cell transfusion in patients with sickle cell disease (SCD) and can result in severe (delayed) haemolytic transfusion reactions, exacerbation of clinical symptoms and life‐threatening hyperhaemolysis. Once alloimmunized, the presence of alloantibodies in the patients’ blood further complicates pretransfusion testing and hampers the selection of compatible blood products. Numerous studies have shown that SCD patients have a relatively high risk of alloimmunization as compared to the ‘general’ population. This is not only explained by the large number of transfusions given but also by the increased exposure to foreign antigens as a result of differences in the antigen make‐up of the SCD recipients and the blood donor population. Other factors involved in the immune response such as age at first transfusion, inflammatory state, HLA typing are under investigation and are starting to unravel.

Because blood transfusion is still one of the main treatment modalities for SCD and some patients have a life‐long transfusion dependency it is important to minimize the alloimmunization rate. Theoretically, complete matching for all relevant blood group antigens would prevent alloimmunization. This however, is only possible when all donors are comprehensively. Matching strategies should be developed to minimize alloimmunization while balancing patients’ need and donor availability and is cost effective. To develop a (preventive) matching strategy some factors need to be established; 1) which antibody specificities are clinically relevant 2) which antigens are most immunogenic 3) what is the availability of specific antigen typings in the donor population 4) how should recipients (and donors) be typed, phenotypically and/or genotypically and to what extent. The latter is especially important in SCD patients since they are of African descent and the prevalence of genetic variations in this population is relatively high. RhD and RhCE variants are common and can remain undetected when serological typing is used but can be discovered with high resolution molecular typing. Patients with partial Rh phenotypes are at risk for alloimmunization. Apart from special Rh phenotypes in individuals from African descent, the Fy(a‐b‐) phenotype related to the GATA‐box mutation in the FYB allele and the U‐ or U^var phenotype resulting from genetic variations in the MNS alleles are also common.

Several studies have shown that in SCD patients antibodies directed against RhD, RhE, RhC and K are most frequently found when unmatched transfusions are given. Preventive matching for these antigens has proven successful in reducing alloimmunisation. Extended matching for all Rh antigens Fy(a), Jk(a) and Jk(b) can further decrease the alloimmunisation rate. Currently, different countries have preventive matching strategies in place for this vulnerable patient group. As genotyping is more and more available and within reach, optimal antigen typing approaches for patients and donors, combining serology and genetics are being developed. In this lecture several aspects of antigen typing approaches and preventive matching strategies that will most benefit SCD patients of will be discussed.

MANAGEMENT OF DELAYED HEMOLYTIC TRANSFUSION REACTIONS AND HYPERHEMOLYSIS SYNDROME

S Trompeter

No abstract available.

THE USE OF ARTIFICIAL INTELLIGENCE TO IMPROVE HEALTH CARE: WHAT'S IN IT FOR TRANSFUSION MEDICINE?

B Geerts

Anesthesiology, Amsterdam UMC, Amsterdam, Netherlands

Artificial intelligence has become a buzzword that will appear about anywhere in the media. We can forget that AI, or the subfield in this computer science field machine learning, has been around for over 50years. Improvements in computing power, abundance of data, progress in computer science, and the arrival of affordable cloud solutions have now brought it to our daily lives.

Also in health care news about AI has become omnipresent. And some landmark papers have come out on algorithms outperforming (teams of) physicians in the diagnosis of all kinds of skin disease, eye disease from retina scans, and detect cancer in CT scans. However, little of these solutions have actually shown up in our clinical practice yet.

In anaesthesia, we worked with the first algorithm to come to anaesthesia practice; Hypotension during surgery is associated strongly with poor outcome like myocardial ischemia, surgical complications, renal failure and even mortality. We worked on a machine‐learning trained algorithm that predicts hypotensive events using the arterial blood pressure curve up to 15–30min before the actual event.

To get FDA and CE approval, however, mere mathematic validation is required. This can be achieved on retrospective datasets. In reality, we need more before we can use these algorithms to support our decision‐making;

After internal (retrospective) and external (prospective but passive use) validation steps, clinical (i.e. RCT validation is needed. Moreover, we will need to assess the economic impact too. Ultimately this tool has now reached clinical practice and is starting to help us go from reactive to more proactive hemodynamic management.

Like this, we have started to work on machine‐learning tools to predict the incidence of specific types of patients coming into A&E and predicting infections after surgery.

We will discuss our approach, essentials to start with machine learning, practical learnings. We will also discuss a first project design to use machine learning in managing bleeding patients to get the best therapy advice for blood product use like plasma, fibrinogen et cetera. How can we start using this tool in unison with our existing tools to improve science and clinical practice in our respective (bio)medical fields?

BLOOD DONATION: INCENTIVES AND INDUCEMENTS – WHERE TO DRAW THE LINE?

P Flanagan

New Zealand Blood Service, Wellington, New Zealand

The ISBT Code of Ethics (the Code) identifies that blood donation should be voluntary and non‐remunerated (VNRD). The Council of Europe definition of VNRD contained within the Code states that ‘A donation is considered voluntary and non‐ remunerated if the person gives blood, of his/her own free will and receives no payment for it, either in the form of cash, or in kind which could be considered a substitute for money. This would include time off work other than that reasonably needed for the donation and travel. Small tokens, refreshments and reimbursements of direct travel costs are compatible with voluntary, non‐remunerated donation’. In practice however Blood Services need to ensure that sufficient suitable donors are available to meet the clinical needs of the patients and hospitals that they support. In an increasingly busy world Blood Services need to compete with other organisations promoting community health and well‐being in order to achieve this. Promotional and marketing activities are utilised to both attract new donors and to encourage repeat donation from regular donors. Care needs to be taken to ensure that these activities do not inadvertently breach the principles underpinning VNRD. The question can then be posed as to when does an activity cease to be an encouragement to altruistic donation and provides the potential donor with a benefit that acts as an inducement for them to donate. The Nuffield Council on Bioethics in its publication on ‘Human Bodies: Donation for medicine and research’ has developed an ‘intervention ladder’ to assist in answering this question. The ladder comprises six categories of interventions (rungs) designed to increase the likelihood that an individual will donate and ranks these in order of ethical complexity from provision of information (rung1) to provision of financial incentives (rung 6). In doing so they also identify that ‘how individuals respond to such inputs will clearly vary from person to person, and indeed inevitably there will be some degree of overlap in how people respond to neighbouring rungs’. This suggests that there is no easy or clear boundary to identify what might be acceptable and that some activities might be acceptable with one section of the population but not with others. This presentation will explore some of these boundary issues and assist the development of some principles to assist in managing the dilemmas.

PLATELET TRANSFUSION TRIGGERS IN NEONATES

E Lopriore

Neonatology, LUMC, Leiden, Netherlands

Thrombocytopenia is a very common hematological abnormality found in newborns, especially in preterm neonates. Two subgroups can be distinguished: early thrombocytopenia, occurring within the first 72h of life, and late thrombocytopenia, occurring after the first 72h of life. Early thrombocytopenia is associated with intrauterine growth restriction, whereas late thrombocytopenia is caused mainly by sepsis and necrotizing enterocolitis.

Platelettransfusions are the hallmark of the treatment of neonatal thrombocytopenia. Most of these transfusions are prophylactic, which means they are given in the absence of bleeding. However, the efficacy of these transfusions in preventing bleeding has never been proven. In addition, risks ofplatelettransfusion seem to be more pronounced in preterm neonates. Because of lack of data,platelettransfusion guidelines differ widely between countries.

In a recent randomized controlled trial (Planet‐2/Matisse study) among preterm infants with severe thrombocytopenia, we found that those randomly assigned to receiveplatelettransfusions at aplatelet‐count threshold of 50×10⁹/L had a significantly higher rate of death or major bleeding within 28days after randomization than those who receivedplatelettransfusions at aplatelet‐count threshold of 25×10⁹/L.

This presentationsummarizes the current understanding of etiology and management of neonatal thrombocytopenia.

DIAGNOSING TACO IN CHILDREN

F Gauvin

Pediatrics, CHU Sainte‐Justine, Montreal, Canada

Transfusion‐associated circulatory overload (TACO) is a severe transfusion adverse reaction that is associated with increased mortality and morbidity. The incidence of TACO in adults varies from 1% to 8%, but is probably underdiagnosed and underreported. The incidence in the pediatric population is undetermined.

TACO usually occurs in patients who receive a large volume of blood product over a short period of time. It is more common in patients with known risk factors such as cardiovascular disease, renal failure, and older or younger age (> 60years or<3years). Hospitalised patients and intensive care patients are also more at risk.

The typical presentation of TACO is respiratory distress (dyspnea, tachypnea) occurring within 12h of a blood transfusion. Associated signs and symptoms are hypoxia, hypertension, tachycardia, positive fluid balance, high central venous pressure, and acute or worsening pulmonary edema on chest X‐ray. Echocardiography and measurement of brain natriuretic peptide (BNP) or its N‐terminal prohormone (NT‐proBNP) is helpful for diagnosis.

Several definition criteria have been proposed for TACO, but none are adapted for children, particularly critically‐ill children who are more at risk. This is probably the main reason why TACO is even more underdiagnosed and underreported in the pediatric population. In a recent study, we compared the incidence of TACO in a pediatric intensive care unit using the International Society of Blood Transfusion (ISBT) criteria, with two different ways of defining abnormal values: 1) using normal pediatric values published in the Nelson Textbook of Pediatrics; and 2) using patients as their own controls and comparing pre‐ and post‐transfusion values with either a 10% or 20% difference threshold. We monitored for TACO up to 24h post‐transfusion. A total of 136 patients were included. TACO incidence varied from 1.5% to 76%, depending on the definition used. With such wide variability, we conclude that a more operational definition of TACO is needed in pediatrics, particularly for critically‐ill children.

Differential diagnosis from other dyspnea‐associated transfusion adverse reactions (e.g. transfusion‐associated lung injury, anaphylaxis) is important because treatment differs, as do guidelines to the blood bank. Treatment for TACO is similar to that of any other cardiogenic pulmonary edema: oxygen, diuresis, ventilatory support. Prevention is possible by avoiding unnecessary transfusions, transfusing only the necessary amount of blood product, avoiding rapid transfusions, and using diuretics.

APPLYING PATIENT BLOOD MANAGEMENT (PBM) PRINCIPLES IN PAEDIATRIC PATIENTS

G Crighton¹, H New², S Stanworth³

¹Haematology, Royal Children's Hospital, Melbourne, Australia ²Centre of Haematology, Imperial College, London ³Clinical Haematology, John Radcliffe Hospital, NHS Blood and Transplant Oxford, Oxford, United Kingdom

Globally, children less than five years of age have the highest prevalence of, and the most severe degrees of anaemia. Iron deficiency is the main contributing cause of anaemia in children and children are the most vulnerable population group to the long‐term detrimental effects of anaemia and iron deficiency.

Patient blood management (PBM) refers to an evidence‐based bundle of care that aims to improve patient outcomes through optimal use of transfusion therapy, assessment and management of anaemia, optimising haemostasis and the use of blood conservation strategies. Patient blood management is now internationally endorsed as a standard of care and PBM interventions are well established in adults, yet implementation in paediatric settings trails behind. This is in part, because paediatricians are unfamiliar with the term PBM, and also that the definitions, guidelines and implementation strategies developed for adults are not always applicable to neonates, infants and children. There is also a paucity of literature that specifically examines paediatric blood management. This article discusses a number of key elements important in paediatric PBM and how the principles of PBM can be applied to paediatric settings.

A RISK‐BASED DECISION‐MAKING FRAMEWORK FOR BLOOD SAFETY: WHAT'S THE CASE FOR ZIKA?

E Bloch

Pathology, Johns Hopkins University School of Medicine, Baltimore, United States

Risk‐based decision making (RBDM) is a decision‐making approach that weighs the complex interplay of all the factors that influence policy pertaining to a given health threat. RBDM strives to minimize risk, optimize outcomes and prioritize resources for maximum good. In the context of blood transfusion safety, RBDM stems from a growing awareness that the current paradigm of attaining “safety at any cost” is unsustainable and unbalanced. Such has been unduly shaped by historical failures, mostly notably surrounding the response to HIV in the 1980s. By contrast, RBDM strives to evaluate all input objectively. Global assessment spans scrutiny of the best available scientific data at the time of decision‐making, broad engagement of key stakeholders, and careful consideration toward the ethics, social values, politics, economics and historical precedent both for and against intervention. RBDM emphasizes proportionality. It should also allow for review of decisions in light of new data, with the scope to amend accordingly, acknowledging that the merits of a given policy can change based on new findings. Zika virus, a formerly obscure mosquito‐borne flavivirus, emerged rapidly in 2015, spurring an international public health emergency. The response in the United States (US), offers an illustrative case study of the challenges surrounding intervention for transfusion‐associated infections. The demonstration of severe clinical complications in selected patient subsets (i.e. fetuses), large numbers of travel acquired cases and rare cases of autochthonous transmission, evidence of viremia in blood donors and plausible transfusion transmissibility spurred the mandatory adoption of individual donor nucleic acid testing (ID‐NAT), a policy decision that remains contentious. The Zika pandemic has waned with low numbers of cases reported in the continental US in 2018, all of which were acquired through travel. Four years after the emergence of Zika in the Americas, no clinical cases have been attributed to blood transfusion. The modeled cost‐utility of ID‐NAT in the US and US territories (e.g. Puerto Rico) is $341 million per Quality Adjusted Life Year. The raw cost of testing is $US137 million per year. Blood collection agencies are under economic strain, deterring investment in strategies to contend with infectious risks that —unlike Zika—have demonstrated risk to the blood supply (e.g. bacterial contamination, Babesia). Nonetheless, while the decision to screen for Zika in US blood donors could be viewed unfavorably, in many ways it has been true to RBDM. Such highlights the many challenges of RBDM, whereby decisions need to balance historic precedent, disparate stakeholders, public perception, national culture and politics, which have significant impact on decision‐making. The challenge rests with reconciling these factors to restore proportionality and scope for revision.

SCREENING STRATEGIES TO MINIMIZE THE RISK OF TT‐HEV INFECTIONS

T Vollmer, C Knabbe, J Dreier

Institut für Laboratoriums‐ und Transfusionsmedizin, Herz‐ und Diabeteszentrum Nordrhein‐Westfalen, Universitätsklinik der Ruhr‐Universität Bochum, Bad Oeynhausen, Germany

Background: The risk and importance of transfusion‐transmitted hepatitis E virus (TT‐HEV) infections by contaminated blood is currently a controversial discussed topic in transfusion medicine. In particular, the infectious dose is a not finally determined quantity. The different countries have chosen different approaches to deal with this pathogen. One central question is the need of individual NAT screening (ID) versus minipool NAT screening (MP) approaches to identify all relevant viremias in blood donors.

Aims: Comparison and evaluation of the available screening strategies in relation to the infectious dose to minimize the risk of TT‐HEV infections.

Methods: We systematically reviewed the presently known cases of TT‐HEV infections and available routine NAT‐screening assays. Furthermore, blood donation screening strategies for HEV EHEV in effect in the European Union were compared. We also describe our own experiences of HEV screening utilising an ID‐NAT‐based donor screening algorithm compared to MP‐NAT in pools of 96 samples. From November 2017 to January 2018, a total of 10,141 blood donations were screened for the presence of HEV RNA using a MP‐NAT (in house, RealStar HEV RT‐PCR Kit) and an ID‐NAT (cobas 6800 platform).

Results: The review of the literature revealed a significant variation regarding the infectious dose causing hepatitis E. In the systematic case review, all components with a viral load (VL) greater than 5.00E+04IU caused infection (definitive infectious dose (DIFD). The lowest infectious dose resulting in TT‐HEV infection observed in general was 7.05E+03IU (minimal infectious dose (MIFD). The infectious dose of the different blood products is mainly influenced by the remaining plasma content.

Our data comparing the two different HEV screening algorithms revealed eight HEV RNA positive donations using a MP‐NAT (incidence 1:1,268), whereas 17 HEV RNA positive donations were identified by ID‐NAT (incidence 1:597); all ID‐NAT only positive donations had VL<25IU/ml.

Summary/Conclusions: Taken into account the current knowledge on the required MIFD, the DIFD, and the analytical sensitivities of the screening methods, we extrapolated the detection probability of HEV‐RNA positive blood donors using different test strategies (NAT assay, ID vs. minipool with different pool sizes). We also considered the amount of plasma in the different blood products and calculated the infectious doses needed to be detected. Only ID testing would be sufficient to detect the minimum VL in the donor to avoid TT‐HEV infections based on the currently known MIFD, but a highly sensitive MP‐NAT should be adequate as a routine screening assay to identify high viremic donors and avoid TT‐HEV infections based on the DIFD. We have also determined that the incidence of HEV infection was approximately 50 % higher if ID‐NAT was used. However, VL were below 25IU/ml and will most likely not result in TT‐HEV infection taken into account the currently known MIFD or DIFD. The clinical relevance of and need of identification of these low level HEV positive donors still require further investigation.

UPDATE ON PATHOGEN INACTIVATION TECHNOLOGY

M Lozano, J Cid

Hemotherapy‐Hemostasis, University Clinic Hospital, Barcelona, Spain

In the last 25years several pathogen inactivation (PI) technologies have been developed to be applied to blood components. Technologies for inactivating pathogens in plasma and platelets are available in the European Union, and some others are currently under development. The first PI technology introduced in the market was for plasma, and was based on the addition of methylene blue and the illumination with light (Theraflex MB‐Plasma, MacoPharma and Grífols). For platelets and plasma two technologies are licensed, one is based on the addition of amotosalen and the illumination with ultraviolet light (UV) (Intercept^®, Cerus) and the other one combines the addition of riboflavin (vitamin B2) and the illumination with UV light (Mirasol^®, Terumo BCT). Currently another technology for platelet inactivation, based on the illumination with UVC light and strong agitation is under development (Theraflex, MacoPharma). For red blood cells one technology based on the addition of one molecule (amustaline, Cerus) is being developed. The mechanism of action, and the spectrum and level of inactivation of pathogens varies among the different technologies. In addition, the number of studies with clinically relevant endpoints and the number of patients included in the studies is not homogeneous. There is published evidence for most of them that show that the treated blood components are safe and efficacious for the patients although, for treated platelet concentrates some decrease in the posttransfusion recovery and survival of the transfused platelets occur, with differences between the different technologies. However, cumulative experience on the use in routine, for some of the technologies for almost 20years, support the concept of the safety and efficacy of the blood components treated with pathogen inactivation technologies without a significant increase in utilization. The use of pathogen inactivation for blood components is not widespread. Differences in epidemiology between countries, infectious risk perception, concerns about potential adverse effects associated with its use and economical considerations might explain the differences observed in its implementation.

P1PK: A BLOOD GROUP SYSTEM WITH AN IDENTITY CRISIS

Å Hellberg

Division of Laboratory Medicine, Office of Medical Services, Clinical Immunology and transfusion medicine, Lund, Sweden

The history of the P1 and P^k antigens is complicated and sometimes confusing because of several changes to the nomenclature. The association between the antigens and their genetic home has raised many questions as well as the longstanding enigma regarding the molecular mechanism underlying the common P1 and P2 phenotypes.

The system (ISBT no. 003) currently includes three different antigens, P1, P^k and NOR. The P1 antigen was discovered already in 1927 by Landsteiner and Levine while P^k and NOR were described in 1951 and 1982, respectively. As for the ABO system, naturally‐occurring antibodies of IgM and/or IgG classes can be formed against the missing P1/P^k carbohydrate structures. Anti‐P1 is usually a weak and cold‐reactive antibody very rarely implicated in hemolytic transfusion reaction (HTR) or hemolytic disease of the fetus and newborn (HDFN). However, some antibodies against P1 have been reported to react at 37°C, bind complement and cause both immediate and delayed HTRs. The P^k antibodies can cause HTR and anti‐NOR is regarded as a polyagglutinin with unknown clinical significance. A higher frequency of miscarriage is seen in women with the rare phenotypes p and P1^k /P2^k. The RBC of the fetus as well as of the newborn express low amounts of P1, P and P^k antigens but the placenta shows high expression and is consequently a possible target of the antibodies and the cause of the miscarriages. The P^k and P1 antigens have wide tissue distributions and can act as host receptors for various pathogens and toxins. Furthermore, altered expression of P^k antigen has been described in several cancer forms.

A longstanding question has been why individuals with p phenotype not only lack P^k and P expression but also P1. Recently it was clarified that the same A4GALT‐encoded galactosyltransferase synthesizes both the P1, P^k and NOR antigens and in addition the P1 and P2 phenotypes was confirmed to be caused by transcriptional regulation. Transcription factors bind selectively to the P¹ allele in the 5’‐regulatory region of A4GALT, which enhance transcription of the gene.

It has been debated whether the P^k and P1 antigens exist on glycoproteins in the human RBC membrane or if glycolipids are the only membrane components carrying these epitopes. A recent publication shows that the P1 antigen can be detected on human RBC glycoproteins and thus glycosphingolipids can no longer be considered as the sole carriers of the antigens.

The blood group system which started out with one antigen, P1, has now gained two more members namely P^k and NOR. Step by step the biochemical and genetic basis underlying the antigens expressed in this system has been revealed but still many questions remain to be solved.

WHY ARE THE GATA AND KLF REGULATOR GENES INCLUDED IN THE BLOOD GROUP TABLES?

N Nogues

Immunohematology, Banc de Sang i Teixits, Barcelona, Spain

Neither GATA1 nor KLF1 represent a blood group system but mutations in the genes encoding these transcription factors (TFs) have been shown to result in simultaneous altered expression of blood group antigens in certain rare blood group phenotypes. In particular, mutations in the KLF1 gene are responsible for the dominantly inherited In(Lu) phenotype, commonly referred to as Lu(a‐b‐) because of the gross reduction in Lutheran antigens expression. Red cells from In(Lu) individuals, though, have also weakened expression of other blood group antigens, like the high‐incidence antigen AnWj, the antigens of the Indian blood group system (CD44) and P1, among others. Since the first description of KLF1 variants associated with the In(Lu) phenotype, many other variants of this gene have been reported with an impact on blood group antigen expression and they are listed on the KLF1 Table of Blood Group Alleles. Other than KLF1, a mutated GATA1 gene has also been found associated with the X‐linked form of the Lutheran‐mod phenotype and has likewise been registered in the GATA1 Allele Table. Besides the effect of TF variants on blood group antigen expression, there are transcription factor‐binding site polymorphisms in regulatory regions of blood group genes, which also have an impact on the expression of the encoded antigens in red cells. The first example of such type of polymorphisms was described in 1995, when the disruption of a GATA motif in the ACKR1 gene promoter was found to abolish erythroid gene expression in Fy(a‐b‐) individuals of African descent. The impact of mutations affecting GATA1 binding sites has also been described in some ABO subgroups, like the Am and Bm phenotypes. A regulatory element with GATA binding sites in the first intron of the ABO gene has been found to be altered in individuals with these phenotypes, either by deletion or by a point mutation disrupting the GATA motif.

Recent findings have also revealed that Xga expression on red cells is dependent on GATA1 binding to a control element located 3.7kb upstream of the XG gene. A single nucleotide polymorphism (SNP) within this region was shown to correlate very well with the expected distribution of the Xga negative phenotype in different populations. Further work has demonstrated that this G>C SNP disrupts a GATA1 binding site and consequently abolishes erythrocyte Xg expression. Overall, these investigations have allowed to elucidate the underlying genetic basis for Xga expression and have made Xga genotyping possible. Similar to Xga, the P1 antigen has been known for a long time to be determined by the A4GALT gene but the molecular basis underlying the common P1/P2 phenotypes has remained elusive till recently. Several cis‐regulatory SNPs had been identified in non‐coding sequences around exon 2a, which showed a very good correlation with P1 antigen expression. Interestingly, potential binding sites for several hematopoietic TFs were identified in the same region. Finally, recent investigations have demonstrated the role of the RUNX1 TF in the expression of P1 antigen, by selective binding to a regulatory site present in P1 but not in P2 alleles. To summarize, variation in blood group antigen expression may result from mutations or polymorphisms in the regulatory region of blood group genes. Recent reports have unravelled the molecular mechanisms underlying the expression of P1 and Xga blood group antigens, which involves TF binding to allele‐specific regulatory elements. Similar mechanisms may also regulate antigen expression in other blood group systems.

CELL‐FREE FETAL DNA FOR FETAL BLOOD GROUP GENOTYPING: NON‐INVASIVE PRENATAL TESTING

FB Clausen

Clinical Immunology, Copenhagen University Hospital, Copenhagen, Denmark

Since the discovery of cell‐free fetal DNA (cffDNA) in pregnant women's blood, the development of noninvasive prenatal testing (NIPT) has provided new diagnostic applications in prenatal care. In transfusion medicine and clinical immunology, cffDNA is extracted from maternal plasma to predict fetal blood groups with the purpose of 1) guiding targeted Rh prophylaxis in non‐immunized RhD negative women and 2) assessing the risk of hemolytic disease of the fetus and newborn (HDFN) in immunized women. I will give an overview of noninvasive prenatal testing of fetal blood groups. Based on the literature, I will summarize the current experience with noninvasive prenatal testing of fetal RHD and other blood groups. For RhD negative pregnant women, routine clinical testing is available in several countries world‐wide to assess the risk of HDFN in D immunized women, and routine testing to guide Rh prophylaxis is now implemented as nationwide service in 6–7 European countries. Noninvasive prenatal testing for fetal RHD is highly accurate with sensitivities of 99.9%, as reported from clinical programs. In general, the sensitivity is challenged be low quantities of cffDNA, especially in early pregnancy. The specificity is challenged by the polymorphic Rh blood group system, where careful attention is needed to navigate among the many RHD variants. RHD variants may complicate cffDNA analysis and interpretation of results, especially in populations with mixed ethnicities. Despite these challenges, fetal RHD testing is very feasible when implemented with careful attention to these issues. For blood groups that are determined by SNPs, such as KEL or Rhc, the main challenge has been interference from the maternal DNA when analyzing the fetal DNA which has resulted in low accuracy and lower sensitivity, when using qPCR. With the application of more novel techniques such as next generation sequencing and droplet digital PCR, accurate noninvasive prediction of these fetal blood groups has been demonstrated. The success of predicting fetal RhD and its successful clinical implementation into national programs should encourage wide‐spread use of cell‐free DNA based analysis. Future work on noninvasive prenatal testing of fetal blood groups determined by SNPs may consolidate the application for cell‐free DNA testing for such targets, including human platelet antigens. At ISBT, the newly formed cfDNA subgroup of the Red Cell Immunogenetics and Blood Group Terminology Working Party will work to facilitate clinical applications, implementation and evaluation of cell‐free DNA testing.

IT IN BLOOD BANKS, PRESENT AND FUTURE. A NORDIC PERSPECTIVE

H Jensen

Aarhus University Hospital, Dept. of Clinical Immunology, Aarhus N, Denmark

Blood banks in most of the Nordic countries all share a vein‐to‐vein approach which in short means that the collection of blood, the preparation of blood components, testing/release and storing is served by a single actor. On top of that recipient and donor blood grouping, crossmatching, delivery, registration of transfusion and of any complications is usually handled in a single blood banking information system (BBIS). This means that blood banks in the Nordics are traditionally operated by a single vendor.

The needs for process control in a single vendor BBIS, the present solutions, unsolved challenges and untapped possibilities of streamlining processes have been scrutinized with the intention to describe separate processes and to acquire best of breed or best of suite IT‐systems. The aim for integrations, rather than building an integrated IT‐system, to support the need for a vein‐to‐vein process is a precondition in the Nordic countries.

With multiple IT‐systems supporting isolated processes, we intend to facilitate development in these and furthermore increase the flexibility in the whole process.

We set out to reveal any existing knowledge in the literature on IT vendor strategies for blood banks, but we didn't succeed in identifying any relevant literature. However, a systematic literature study on vendor strategies when choosing Health IT was based on the PRISMA method, and identified 837 studies, but only 10 was eligible for full text review and 5 met the inclusion criteria.

Even this broader literature study reveals very little evidence. Two studies find single vendor strategies poor and conclude “best of suite” solutions to be optimal. One study was not able to correlate vendor strategies to the investigated productivity, but concludes that best of suite and best of breed strategies requires larger organizational changes than a single vendor strategy.

In summary, the existing research is contradictory.

This paper adds basic knowledge for breaking down the process control of blood banking in smaller processes. This adds the possibility for identifying best of breed or best of suite vendors, instead of relying on single vendor IT solutions. Furthermore it is a call for more research in the field of vendor selection strategies which this study didn't succeed in identifying.

APPLYING DRONES TO SUPPLY BLOOD TO REMOTE AREAS: RWANDA'S EXPERIENCE

S Gatare

Biomedical Services, Rwanda‐Ministry of Health‐National Blood Services, Kigali, Rwanda

Background: In Rwanda, blood transfusion services started in 1976. During the 1994 genocide against the Tutsis almost all the socioeconomic fabric of Rwanda was destroyed as well as its health infrastructure. The healthcare system was suffering in its aftermath, and there were health inequalities between urban and rural areas, including access to blood for transfusion. From 1995, the government started to rebuild all courses of life including the health system and the blood service in particular. The National Center for Blood Transfusion (NCBT) was then mandated to provide safe, effective and adequate blood and blood components to all patients in need. This was pivotal in achieving health related MDGs 4, 5 and 6.

Today, Rwanda has an ambitious vision to put all 12 million citizens within 30minutes of any essential medical product. While every second matters in emergency management, the use of Drones was the perfect solution to many of the last mile challenges that have been traditionally difficult to overcome. It is impossible to forecast accurately down to the need of a single patient. The Government has provided an easy solution by centralizing supply and providing on‐demand, emergency medical deliveries by drone.

Doctors are now empowered to provide the quality care with all the supplies on hand, patients can now be treated close to home, and we eliminate waste from potential overstocking when health workers know that they have a quick and reliable source of supply.

Description: In 2016, the Government of Rwanda started to operate the world's first national drone delivery program for blood and other lifesaving medical products. These drones can carry two to six units of blood at a time and deliver in 15 ‐ 45minutes depending on a hospital's location. The average duration was between 2 ‐ 4hours round trip with the vehicle system, before the use of Drones. Drones currently deliver blood to 25 health facilities throughout the country and are set to reach 90% of transfusing health facilities outside Kigali by the end of the year.

Within the first year, Healthcare workers saved an average of 3.1hours per delivery and a total of 10,115hours of lost time on road pick up they could instead dedicate to patient care. By March 2019, over 11,000 deliveries have been made, with 30% of those being emergency deliveries. A total of more than 19,500 blood units have been delivered. In February 2018, Zipline obtained the highest rating from the health facilities being served in a performance evaluation conducted by the National Center for Blood Transfusion.

When a doctor or medical staffer needs blood, they place an order through the Hemovigilance order portal. They are then sent a confirmation message saying a drone is on its way. The drone flies to the health facility at up to 100km/h. When it is within five minutes of the destination, the medical staffer receives a notification. The drone then drops the package, attached to a parachute, into a special drop zone.

Conclusion: Supply is not a developing country problem, it is a global issue. Rwanda was just the first one to recognize the potential of this technology and decided to do something about it first and fast, to ensure access to universal access to all blood products.

NOVEL CHALLENGES IN BLOOD INVENTORY MANAGEMENT AND BEDSIDE TRANSFUSION

L Lodge

Scottish National Blood Transfusion Service, Edinburgh, United Kingdom

Supporting the provision of viable transfusion services in remote/rural areas is more than just a geographical challenge. Limited qualified blood bank resource; small throughput volumes; increased regulation are only three of the additional factors combining to threaten safe and sustainable transfusion service delivery. Inventory management and out of hours service provision were identified as essential areas where it was thought that technology, in the form of a remotely controlled blood fridge, could provide a key element of the overall solution. A Radio Frequency Identification (RFID) fridge racking system was installed in a standard blood bank fridge and connected to the Laboratory Information Management System (LIMS) common to both the remote and central blood bank. The central blood bank was enabled to test patient samples from the remote laboratory, identify components located in the remote fridge suitable for the patient and allow correct component issue, even when qualified staff were unavailable in the remote laboratory. Testing has concluded that installation of remote fridge management can play a major role in helping to maintain a remote inventory permitting patient compatible components to be issued. By sustaining transfusion services in remote communities we can avoid transportation of patients who require transfusion support to locations miles from home.

WORK‐UP IN CASE OF GRANULOCYTE ANTIBODIES

BK Flesch

Laboratory of Immunogenetics / HLA, DRK Blutspendedienst West, Bad Kreuznach, Germany

Background: Granulocyte allo‐ and autoantibodies have been implicated in a variety of clinical conditions such as primary and secondary autoimmune neutropenia, alloimmune neutropenia, and febrile and severe pulmonary transfusion reactions like transfusion related acute lung injury (TRALI).

Aims: To provide a work‐up in case of granulocyte antibodies.

Methods: An overview of the clinical conditions, methods for antibody testing, typing and antibody specificities related to the different diseases will be provided.

Results: Autoimmune neutropenia occurs as primary form in infants younger than three years or secondary to infections, malignancies or other autoimmune diseases mainly in adults and is associated in most cases with antibodies to HNA‐1a. Antibodies to HNA‐1b, FcgRIIIb and HNA‐2 have been reported, too. Neonatal alloimmune neutropenia results from maternal antibodies transferred transplacentally to the fetus and is caused by all known HNA‐antibody specificities, i.e. HNA‐1a, ‐1b, ‐1c, ‐1d, HNA‐2, HNA‐3a, ‐3b, HNA‐4a, ‐4b and HNA‐5a specificity. HNA and HLA antibodies can induce mild febrile transfusion reactions and TRALI. Since the introduction of the male only plasma strategy, in many countries the TRALI incidence decreased but it is still one of the most common causes of severe transfusion reactions. Especially HNA‐3a antibody containing plasma from female donors is responsible for severe or even fatal reactions. But also HNA‐1a, ‐1b, HNA‐2 and HLA class I and class II antibodies were reported. The latter activate monocytes to secrete soluble factors that act on the primed neutrophils in the narrow lung capillaries.

Laboratory testing: Laboratory work‐up requires the knowledge of the patient's clinical condition and the methods that are appropriate to detect the relevant antibodies. The classical granulocyte agglutination test (GAT) in combination with the granulocyte indirect immunofluorescence test (GIFT) can detect nearly all relevant antibodies. HNA‐1, ‐2, ‐4, ‐5 and HLA class I antibodies are clearly detectable in the GIFT while HNA‐3 antibodies strongly agglutinate neutrophils in the GAT. The monoclonal antibody‐specific immobilization of granulocyte antigens (MAIGA) test detects all HNA‐antibodies except for HNA‐3 with high glycoprotein specificity and sensitivity but is time consuming and requires highly skilled personnel. For TRALI diagnostics laboratory testing is completed by methods like the indirect lymphocyte immunofluorescence test (LIFT) or ELISA for HLA class I antibodies and HLA class II specific ELISAs. Since several years fluorescent bead based assays (Luminex) enable faster and more automated HNA antibody detection but to date not all specificities, especially HNA‐3, can be reliably detected so that still the classical GIFT and GAT have to complete the methodological spectrum. Serological typing today is mostly reduced to the determination of HNA‐2 in the GIFT because the molecular reason for the HNA‐2‐null phenotype is not completely understood. Establishing only one PCR‐ASP reaction for the main CD177*787A>T polymorphism would comprise the risk to miss other molecular causes. However, for all other HNA allelotyping by PCR methods is the first choice.

Summary/Conclusions: Granulocyte serology still today is widely based on a variety of manual methods and will be reserved to specialized laboratories as it requires experienced laboratory staff and profound knowledge of granulocyte immunobiology.

OPTIMAL ALGORITHMS FOR SEROLOGICAL AND MOLECULAR TYPING IN PLATELET ALLOANTIBODY INVESTIGATIONS

M Ahlen

Norwegian National Unit of Platelet Immunology, Laboratory Medicine, University Hospital of North Norway, Tromso, Norway

Maternal alloantibodies against antigens on human platelets can cause severe thrombocytopenia and bleeding in fetus or newborn, identified as Fetal/Neonatal Alloimmune thrombocytopenia (FNAIT). Although most cases the thrombocytopenia is self‐resolving within the two first weeks of life, some infants present bleeding symptoms and thus require platelet transfusion. A set of laboratory analyses are required to confirm the FNAIT diagnosis. In addition to guiding compatible platelets to the affected newborn, the correct diagnosis will be valuable to assess the risk of FNAIT in subsequent pregnancies.

In addition, platelet alloantibodies may also complicate platelet transfusions by immune‐mediated platelet refractoriness, and require proper identification of the patient's antibody specificities prior to selection of donor platelets. The algorithm for laboratory investigations include both serological and molecular assays, and depend on the objective and timing: whether there is an urgent need for platelet transfusion, follow‐up of a pregnancy with known risk, or to do full‐scale laboratory testing to confirm diagnosis. Molecular genotyping should include all HPA systems relevant for the local population (in Caucasians HPA‐1, ‐2, ‐3, ‐5 and ‐15), preferably with optional extended panels for systems for low frequency populations due to immigration/mobility and for less frequently seen alloantibodies (HPA‐4, ‐6 to ‐11 are most commonly included). Serological testing of antibody binding to platelets is often initially tested by flow cytometry analysis (direct test and/or cross‐match). However, the detection of platelet‐specific antibodies is often complicated by the presence of anti‐HLA class I antibodies and thus require sensitive platelet glycoprotein‐specific assays. Serological testing for platelet‐specific antibodies includes as a minimum panels of antigens on GPIIb/IIIa, GPIb/IX, GPIa/IIa and CD109 and preferably additional targets for populations with Asian/African origin. Several methods are available; i.e. bead‐based assays and ELISA based methods. However, most reference laboratories perform variants of the monoclonal antibody immobilization of platelet antigens assay (MAIPA), as reported by the 19th International Platelet Immunology Workshop of ISBT (2018). The investigations also include measurement of the anti‐HPA‐1a by quantitative MAIPA if present, as this is reported to potentially predict the severity of FNAIT.For pregnancies with known risk of FNAIT, there are methods available to perform non‐invasive prenatal typing from maternal plasma. The most feasible and so far appropriate for routine testing is fetal HPA‐1 typing with quantitative PCRor by melting curve analysis. Other sophisticated, yet resource‐demanding techniques have also recently been reported ‐ importantly also for typing of other HPA‐systems.

MOLECULAR BASIS OF HNA‐2 EXPRESSION

B Bayat

Justus Liebig University, Institute for clinical immunology and transfusion medicine, Giessen, Germany

Human neutrophil antigen 2 (HNA‐2) is a neutrophil‐specific antigen located on GPI‐anchored glycoprotein CD177 (also known as NB1). HNA‐2 is absent on the neutrophil surface of 3–5% of the healthy individuals that divided the population to HNA‐2 positive and HNA‐2 null individuals. Exposure of HNA‐2 null individuals to HNA‐2 positive neutrophils during pregnancy, after transfusion or transplantation, induces immunization against HNA‐2 and consequently the production of iso‐antibodies. The HNA‐2 iso‐antibodies are involved in the mechanism of neonatal alloimmune neutropenia (NAIN), transfusion‐related acute lung injury (TRALI) and graft failure following bone marrow transplantation.

Presence of CD177 on a neutrophil surface of HNA‐2 positive individuals follows a bimodal expression that categorizes the circulating neutrophils to HNA‐2 positive and negative subsets. The CD177 gene contains 9 exons encoding a protein of 437 amino acids. The lack of HNA‐2 (in HNA‐2 null individuals) is associated with the presence of a missense mutation, CD177*c.787A>T in exon 7 of CD177 gene inducing a premature stop codon in codon 263. This mutation alone or in combination with CD177*c.997delG has been introduced as the main reason for the absence of CD177 in HNA‐2 null individuals.

A pseudogene (CD177P1) highly homologous to exons 4–9 of CD177 gene is located downstream of the CD177 gene. Conversion of exon 7 of CD177P1 into CD177 gene is responsible for the generation of CD177*c.787A>T missense mutation.

In addition, the heterozygosity or homozygosity of CD177*c.787A>T is accounted for regulation of HNA‐2 negative and positive neutrophils subpopulations. Genotyping has revealed the HNA‐2 null individuals, heterozygous for CD177*c.787A>T mutation without CD177*c.997delG, indicating the presence of a complementary mechanism regulating CD177 expression. Newly in HNA‐2 null individuals and individuals with atypical CD177 expression a CD177*1291G>A polymorphism in combination with CD177*787A>T is described.

Altogether these data indicate a complex compound mechanism(s) for regulation of CD177 expression on the neutrophil surface.

This presentation will summarize recent findings on CD177 expression and highlights the potential genotyping methods for genetic assessment CD177 expression of donors and patients.

GETTING WISE WITH ABO‐INCOMPATIBLE LIVE RELATED RENAL TRANSPLANT: A TERTIARY CENTER EXPERIENCE

S Agrawal, M Chowdhry, S Gajulapalli, M Avel

Transfusion Medicine, Indraprastha Apollo Hospital, New Delhi, India

Background: There is substantial growth in ABO‐incompatible (ABOi) renal transplants across the world. Desensitisation by a combination of apheresis and immunosuppression protocols has led to an increased donor pool and reduced the number of patients waiting for a kidney transplant.

Aims: To ascertain the efficacy of ABO titre reduction by two different techniques of apheresis and analyse the parameters influencing the graft outcome.

Methods: All consecutive ABOi renal transplants performed at our centre from February 2012 to January 2019 were retrospectively reviewed. Those with a follow‐up data of at least one year were included for analysis. Immunosuppression included rituximab (375mg/m²) followed by steroids, mycophenolate mofetil and tacrolimus.

Conventional therapeutic plasma exchange (cTPE) for 1–1.5 plasma volume was performed pre and post‐transplant (based on graft function/rising titres) on Haemonetics MCS + cell separator (Braintree, MA, USA) using 5% albumin and 2 units of AB group FFP (Fresh Frozen plasma). Pre‐transplant Immunoadsorption (IA) was done by processing 8 plasma volumes in one session using antigen‐specific immunoadsorbent column (Glycosorb ABO, Glycorex Transplantation, Lund, Sweden) on Terumo BCT's Spectra Optia^®. A pre‐transplant ABO titre (IgM+IgG) of ≤8 was targeted. The patient and graft survival in 1year was evaluated.

Results: Of the 104 transplants performed, 82 patients with a follow‐up of 1year were included for analysis. All combinations of ABO incompatibilities were accepted. Mean age was 38.45±14.79years. There were 67 males and 15 females. Seventy‐one patients underwent cTPE only and the remaining 11 underwent IA ± cTPE for desensitisation.

The median baseline titre was 32 (16–1024). The median titre at transplant and at discharge was 2 (1–8) for all patients. Mean serum creatinine at discharge was 1.3±0.97mg/dl. The mean number of cTPE performed were 3.6±2.3 per patient in the pre‐transplant period for those undergoing only cTPE. Mean number of IA and cTPE performed in patients of IA ± cTPE is 1.7±0.47 and 2.4±2.6 respectively. The mean reduction in serial dilution of titres was 4.9±2.0 in these patients which were statistically higher than those who underwent cTPE only (3.4±2.1, P=0.04). The number of procedures required to reach the target titre was not significantly different for anti‐A and anti‐B (P=0.5). The mean post‐transplant cTPE performed was 1.3±2.3. There was graft dysfunction in 8 patients (9.75%) of which 3 (37.5%) were salvaged and 1 left against the medical advice. Graft survival was influenced by the titre at transplant (P=0.016). The overall patient survival in 1yearwas 91.46% and death‐censored graft survival was 92.1%.

Summary/Conclusions: The net reduction in ABO titre after apheresis is better with the use of IA. The titre at transplant affects the graft outcome irrespective of the method used to achieve this target. Also, the outcome is independent of the titre being monitored (anti‐ A or anti‐B).

The use of plasma exchange and immunosuppression make the outcome of ABOi renal transplant acceptable and comparable with the western counterparts.

DEVELOPMENT OF DETAILED TRANSFUSION EXPOSURE INFORMATION FROM PATIENT ELECTRONIC HEALTH RECORDS USING NATURAL LANGUAGE PROCESSING

BI Whitaker¹, A Williams¹, J Duke², R Boyd², K Natarajan³, S Anderson¹

¹Office of Biostatistics and Epidemiology/CBER, U.S. Food & Drug Administration, Silver Spring, MD ²Health Analytics and Informatics, Georgia Tech Research Institute (GTRI), Atlanta, GA ³Columbia University, New York, NY, United States

Background: The extraction and use of individual ISBT‐128 labeling codes from the electronic health record (EHR)provide valuable identification of blood component patient exposures within an institution. However, to precisely relate blood component exposures to subsequent adverse events, there is often a need for precise transfusion timing and vital signs information.

Aims: Observation of transfusions by nursing staff is standard practice in the US. Transfusion notes typically contain blood product information, start and end times of the transfusion, a flowsheet of vital signs, and documentation of any transfusion reactions. We applied natural language processing (NLP) in order to capture this important supplemental information.

Methods: Nursing notes containing semi‐unstructured data from New York Presbyterian Hospital were utilized in a three‐stage process to extract pertinent information regarding transfusion events. Columbia University Medical Center (CUMC) leveraged a Python‐based NLP platform, ClarityNLP, developed by Georgia Tech Research Institute (GTRI). Before deploying the tool within CUMC's environment, a representative set of 15 de‐identified transfusion nursing notes were sent to GTRI to use as training cases for the NLP extraction tool. In addition to extraction of the existing information, the NLP tool created derivative information; for example, a derived field was created for the total transfusion time in minutes and also for the elapsed times from the start of the transfusion to each set of vitals taken. After GTRI trained the NLP tool, it was installed within the CUMC environment for use. A separate set of transfusion nursing notes was used to test ClarityNLP within the CUMC environment. The parser was able to accurately extract all intended information from the semi‐unstructured notes and output it in a structured format. It accurately populated fields such as transfusionStart, transfusionEnd, elapsedMinutes, reaction, bloodProductOrdered, datetime, timeDeltaMinutes as well as baseline vitals and subsequent measurements taken during the transfusion.

Results: As a final performance evaluation, ClarityNLP was run on roughly thirty‐four thousand transfusion notes. The output for this run included information on blood products ordered, transfusion start and end times, whether patients experienced a reaction to the transfusion as well as vitals measurements taken before and during the transfusion, including temperature, oxygen level, blood pressure and heart rate. For the validation process, 100 transfusion nursing notes were sampled and reviewers assessed the accuracy of the information regarding 1) blood product ordered, 2) whether the patient experienced a reaction, and 3) the start and end times of the transfusion. For each of these fields across all 100 sampled notes, the ClarityNLP tool reproduced these data points with 100 percent accuracy. In addition, the tool supplied transfusion end times for numerous structured records that were missing this key data point.

Summary/Conclusions: ClarityNLP can very efficiently digest a large number of transfusion nursing notes simultaneously and also does an excellent job of extracting the main characteristics of a transfusion, which can be used in partnership with structured data to produce a more accurate and more complete picture of patient transfusions.

RHD IMMUNOGLOBULIN: ARE WE GETTING IT RIGHT IN OBSTETRICS?

L Bielby¹, B Glazebrook¹, C Akers¹, P Beard¹, K Bastin¹, J Daly²

¹Blood Matters, Department of Health & Human Services, Australian Red Cross Blood Service, Melbourne ²Pathology Services, Australian Red Cross Blood Service, Brisbane, Australia

Background: The Blood Matters Serious Transfusion Incident Reporting (STIR) program (Australia) has collected RhD Immunoglobulin (RhDIg) incidents since 2015. STIR receives reports from four jurisdictions with 93 health services registered.

Although the number of reports was small, troubling trends were emerging. To better understand RhDIg use, an audit comparing use to Australian RhDIg guidelines (2015) was undertaken by Blood Matters.

Aims: To review:

• Health services’ policies and procedures on RhDIg use

• Compliance with the currentguidelines for the use of RhDIg inobstetrics in Australia

Methods: Health services with level 2 or higher maternity services (providing antenatal, intrapartum and postnatal care) were invited to participate in a two‐part audit (n=79).

Part 1: Policy for the use of RhDIg in obstetrics.

Part 2: Clinical practice audit of 30 RhD negative women who delivered during the period July 2017 to June 2018.

Results: Part 1: Policy was reported on by 48 (61%) sites. Of these 43 (90%) had a policy for use of RhDIg, with all specifying RhDIg should be administered at 28 and 34weeks, with sensitising events and at delivery for all at risk women. Documented consent was required at 44 (92%) sites, with 23 specifying one consent covered all administrations (antenatal, delivery and sensitising events), 17 requiring consent at each administration, and 4 a separate consent for routine administration versus sensitising events. No specific educational requirements for staff ordering or administering RhDIg were required at 15 (31%) sites.

Despite 11 (23%) health services reporting 36 administration errors in 12months, STIR received only 9 reports during this time.

Part 2: There were 939 individual practice audits returned from 43 (54%) sites. Of these 19 women either refused or did not require antenatal prophylaxis, leaving 920 women in the audit. There was no documented prophylaxis or refusal in 42 (4.6%) women, and 45 (4.9%) received only one of the two recommended doses. Antenatal dosing was not administered at the right dose and time for either or both doses in 248 (27%) women.

Sensitising events were reported in 104 women, with 26 (25%) not given the recommended dose of RhDIg.

At delivery, 604 women had an RhD positive infant, with 593 (98%) receiving postnatal prophylaxis, leaving 11 women at risk, and 514 (85%) had quantification of fetomaternal haemorrhage. Of the 265 women in the audit who delivered RhD negative infants, 12 (5%) received RhDIg. (The remaining 51 infants had unknown RhD status).

Overall, 261 (28%) women audited either missed at least one dose, received a lower than recommended dose or had incorrect timing of administration, putting these women at risk of antibody development. Three women did not receive any RhDIg during their pregnancy with no indication of refusal.

Poor documentation was a limitation in this audit with implications for patient care and traceability.

Summary/Conclusions: RhDIg is administered well against guidelines postnatally (98%). Antenatal dosing for routine prophylaxis and sensitising events was suboptimal in 28% of women, increasing the risk of immune sensitisation and haemolytic disease of the newborn in future pregnancies.

THE EFFECT OF INTRAUTERINE TRANSFUSION ON FETAL RETICULOCYTE COUNTS AND OUTCOME IN ALLOIMMUNE HEMOLYTIC DISEASE OF THE FETUS AND NEWBORN

IM Ree^1,2, E Lopriore², C Zwiers³, M Janssen², M de Haas^4,5,6

¹Clinical transfusion research, Sanquin ²Neonatology ³Obstetrics, Leiden University Medical Center ⁴Center for Clinical transfusion research, Sanquin ⁵Immunohematology and blood transfusion, Leiden University Medical Center, Leiden ⁶Immunohematology diagnostics, Sanquin, Amsterdam, Netherlands

Background: Red cell alloimmunization can lead to hemolytic disease of the fetus and newborn (HDFN). Fetuses with severe HDFN may require treatment with intrauterine red blood cell transfusions (IUTs) to treat fetal anemia. After birth, these infants often require postnatal red blood cell transfusions due to persisting anemia. Preliminary research shows that the absolute reticulocyte count at birth in infants with IUT is significantly lower compared to infants without IUT treatment, suggesting that treatment with IUT may suppress erythropoiesis and lead to increased postnatal transfusion dependency. The effect of the number of IUTs on fetal reticulocyte counts and neonatal outcomes is unclear.

Aims: Our aim was to quantify the effect of one or multiple IUTs on the fetal reticulocyte count in a population of fetuses with severe HDFN, referred to our institute between January 2005 and December 2018.

Methods: Observational study of all consecutive 190 infants with HDFN caused by D antibodies treated with one or multiple IUT(s) and born at the Leiden University Medical Center (LUMC) between January 2005 and December 2018. Data was extracted from maternal and neonatal medical files, including laboratory samples before and after IUT procedures and neonatal outcomes. The primary outcome was adjusted by using a generalised estimating equation (GEE) model to account for the possible interrelations between the outcomes of multiple IUTs in the same fetus including gestational age at time of IUT.

Results: [preliminary] The median number of IUT per fetus was 2 (interquartile range [IQR] 2–4), the first IUT was performed at a median gestational age of 29.0weeks (IQR 24.6–32.1) weeks. Per IUT, the median hemoglobin level before IUT procedures increased in accordance with gestational age from a median a 6.9g/dl (IQR 5.3–8.5) to 8.5g/dl (IQR 7.4–9.7) after 5 IUTs (P<0.001). The median reticulocyte count showed a significant decrease over the course of multiple IUTs, falling 72% from 296*10⁹/L (IQR 234–388, or 177‰ [IQR 121–242]) before the first IUT to 83*10⁹/L (IQR 5–175, or 34‰ [IQR 2–72]) before the second IUT, with an additional 26% decline to 7*10⁹/L (IQR 4–136, or 3‰ [IQR 2–56]) before the third IUT and subsequent IUTs (P<0.001). The number of IUTs was related to the median reticulocyte level at birth, decreasing significantly from 140*10⁹ (IQR 13–237, or 58‰ [IQR 45–83]) after 1 IUT to 8*10⁹ (IQR 3–18, or 2‰ [IQR 1–5]) after 5 IUTs (P=0.002). Although not statistically significant, the number of IUTs may also be related to the proportion of infants requiring postnatal transfusions, varying from 88% to 100% (P=0.596), and the number of postnatal transfusions per infant, varying from a median of 2 (IQR 2–3) to 3 (IQR 2–4) after 5 IUTs (P=0.327).

Summary/Conclusions: In severe HDFN, treatment with IUT cause a decline in fetal and neonatal reticulocyte counts. Already after two IUTs, only 2% of the initial fetal reticulocyte count remain and apparently recovery is slow, since higher proportions of infants develop persisting hyporegenerative anemia requiring multiple postnatal transfusions.

IDARUCIZUMAB FOR DABIGATRAN REVERSAL – A SINGLE‐CENTER ANALYSIS OF TWO YEARS’ EXPERIENCE

A Sarmento¹, BL Pinto², D Cibele², L Gonçalves², F Araújo², C Koch²

¹Centro Hospitalar do Baixo Vouga, E.P.E., Aveiro ²Department of Transfusion Medicine and Blood Bank, Center of Thrombosis and Haemostasis, Centro Hospitalar e Universitário de São João, E.P.E., Porto, Portugal

Background: Idarucizumab is a humanized monoclonal antibody fragment developed to reverse the anticoagulant effect of Dabigatran. It has been used with increasing frequency since its approval in late 2015. Previous studies reporting results of clinical trials have proven its efficacy and safety and real‐life experience data related to its usage has been growing since then.

Aims: To assess Idarucizumab usage in our center, Centro Hospitalar Universitário de São João, in a 2‐year period, regarding its effect on coagulation assays and clinical outcomes, in patients who received idarucizumab.

Methods: We performed a retrospective analysis of idarucizumab usage in 33 patients, between January 2017 and December 2018, who presented serious active bleeding (group A, 20 patients) and patients who required urgent procedures (group B, 13 patients).

Results: After reversal therapy with idarucizumab, all patients presented a normalized unbound dabigatran concentration inferior to 30ng/ml, despite being treated with either the standard 5g dose of intravenous idarucizumab (45% of all patients) or one single 2.5g dose.

Transfusional support, with packed red blood cells, was needed in 33% of all patients; 6% received pooled platelet transfusion; 6% received fresh frozen plasma; 6% received either activated or non‐activated prothrombin complex concentrate .

Patients in group A, excluding patients with intracranial bleeding, presented impaired renal function, with a mean serum creatinine concentration of 1.9mg/dl. Patients in group B presented worse renal function (mean serum creatinine concentration of 3.27mg/dl).

None of the patients presented thrombotic events in the 60days following the administration of the reversal agent.

Overall 30‐day mortality rate was 24% among all patients. None of the deaths were due to continued bleeding or thrombotic events.

Summary/Conclusions: Idarucizumab completely reversed dabigatran anticoagulant effect without causing a prothrombotic state in patients in the following 60days.

The decision to give one single 2.5g dose of idarucizumab to some patients was due to limited idarucizumab availability. However, our results show that it may be reasonable to individualize dosing to selected patients with lower dabigatran concentrations and normal renal function.

PATHOGEN REDUCTION OF FROZEN PLATELETS

D Kutac^1,2, M Bohonek^1,3, L Landova¹, B Kostrouchova¹, E Staskova¹, M Blahutova¹, I Malikova⁴

¹Department of Hematology and Blood Transfusion, Central Military Hospital Prague, Prague ²Faculty of Military Health Sciences, University of Defence Hradec Kralove, Hradec Kralove ³Faculty of Biomedical Engineering, Czech Technical University in Prague, Czech Republic ⁴First Faculty of Medicine Charles University of Prague and the General University Hospital Prague, Prague, Czech Republic

Background: The shortage of platelets can be mitigated by building an inventory of cryopreserved platelets (CP). This strategy has been, successfully used in military as well as in civilian medicine in support of a massive transfusion protocol. With emerging infectious threats, pathogen reduction is promoted in civilian and military transfusion medicine. Many studies for evaluation of quality parameters of frozen platelets (PLTs) are in progress or have been published. However few evaluated the impact of modern pathogen reduction technology on frozen platelets.

Aims: Methods of pathogen reduction technology (PRT) are successfully used for treating plasma, fresh platelets and fresh whole blood. This study explores the impact of pathogen reduction technology on subsequent freezing of platelets. The goal of this study was a comparison of in vitro quality parameters between PRT treated and untreated platelets.

Methods: A comparative study of CP, PRT treated (T‐CP) and untreated (C‐CP), was performed. Both arms of the study were collected as apheresis PLTs type O on a Haemonetics MCS+ machine (Haemonetics corp., USA) and processed under standard procedures. Type‐O CPs were processed with 6% DMSO, frozen at ‐80°C and reconstituted in thawed AB plasma. 16 units were treated with PRT (Mirasol, Terumo BCT, USA) before freezing, and 15 CPs in the control group were left untreated. After reconstitution of CPs the following laboratory tests were performed: blood cells count, PLT, MPV, PLT recovery, glucose, lactate, pH, pO2, pCO2, HCO3, TEG, TGT, CD41, CD42b, Annexin V, CCL5, CD62P, Kunicki score and presence of aggregates>2mm.

Results: No significant differences between both groups were found for the following parameters: number of PLT / unit, recovery, MPV, pH, TGT – Peak, CD41 and Annexin V. T‐CPs showed significantly higher TEG coagulation index, TGT – tLag with faster tPeak, velocity index and ETP as well as metabolic markers were increased, evidenced by lower pO2, pCO2, glucose, lactate and higher HCO3. Expression of CCL5 and sCD62P were elevated in T‐CPs when compared to C‐CPs. Morphological changes evaluated by Kunicki score was higher in T‐CPs. All units of T‐CP contained visible aggregates, with a tendency to increase over time, while no aggregates were seen in C‐CP.

Summary/Conclusions: T‐CPs showed better in vitro hemostatic activity with less morphological changes than their untreated counterparts. T‐CPs seem to be more metabolic activated than C‐CPs, however platelet functionality did not seem impaired. On the other hand, the occurrence of aggregates in T‐CPs probably needs further research or adaptation of preparation procedures. In the meantime T‐CTs should be administered early after thawing with an active hemovigilance in place.

PHENO‐ AND GENOTYPE IN BLOOD GROUP TYPING – TWO SIDES OF THE COIN – PART I

CM Westhoff

Immunohematology and Genomics, New York Blood Center, New York, United States

Antibody‐based typing, with a positive result reflected in agglutination of the red cells (RBCs), has served the profession for nearly a century enabling safe and effective transfusion therapy. The power of RBC typing by serologic methods lies in the availability of standardized antibody reagents which target many of the specificities of significance for transfusion, and the ability to directly detect antigen expression on the RBCs. Hemagglutination has historically been relatively inexpensive, particularly for ABO and RhD as the most important blood groups in most populations. Serologic RBC typing is reliable, requires no sophisticated equipment, is generally straightforward to perform, and is fast requiring less than 1h to results. Hence, antibody‐based testing has been considered the “gold standard” for blood group typing.

With the age of genomics, DNA‐based genotyping is increasingly being used as an alternative to antibody‐based methods. Most antigens are associated with single nucleotide changes (SNPs) in the respective genes. Genotyping has been validated by comparison with antibody‐based typing and has been shown to be highly correlated. The power of genotyping of RBCs lies in the ability to test for antigens for which there are no serologic reagents, and to type numerous antigens in a single assay using automated DNA‐arrays. This increases accuracy and weak antigen expression can be revealed. Fresh RBC samples are not required for DNA extraction, and there is no interference from transfused RBCs or IgG bound to the patient's RBCs. DNA‐based typing is economical in that it provides much more information, but testing requires special equipment, training, and 24‐h turn around.

What then is the best approach to use? Will serologic typing be replaced by DNA‐based typing? Indeed, genotyping will increasingly be used in the practice of transfusion medicine, especially with the growth of whole genome sequencing (WGS). However, because serologic typing for ABO and RhD is fast and accurate and often relied on for sample identification, genotyping will not be the sole means for routine typing. Genomic sequencing approaches will certainly reveal unrecognized changes and genetic variability in RBC membrane proteins, but not all variation will be immunogenic. A genetic polymorphism must be associated with antibody production to be considered a blood group antigen. The importance of an antibody, and antibody reactivity, then will continue to be the central defining principal in transfusion. As two sides of a coin, both are key to safe and effective transfusion therapy.

PHENO AND GENOTYPE IN BLOOD GROUP TYPING – TWO SIDES OF THE COIN – PART II

C Gassner

Independent Researcher, Zurich, Switzerland

Since the mid‐1980s, research in the molecular basis of structural and functional aspects of proteins carrying and producing the antigens has led to an upgraded and modern understanding of blood group variation. Most commonly, single nucleotide polymorphism (SNP) based basic molecular typing techniques were utilized to test new findings on a smaller subset of samples, and resulted in concordant sero‐ and genotyping results in general. However, quite commonly a small number of all samples delivered discrepant results, triggering consecutive rounds of analysis and resolution, finally resulting in a better knowledge with respect to the underlying blood group variation. Such rounds of repetitions represented the synergistic incremental process of learning, learning for serologists and molecular biologists. Inheritance of public, presence of high frequency, or low frequency, or partial antigens and notice of weakly expressed, or almost undetectable antigens marked the path of incremental learning and may best be exemplified by discoveries within the blood group systems ABO, RhD and Kell.

Naming for pheno‐ and genotypes coevolved alongside the permanent discovery of new antigens. At present, antigens and their antithetic counterparts (if tested and if existent), are more commonly reported independently, as exemplarily shown for the following Kell phenotype consisting of three antithetic antigen‐couples: kk, Kp(a+b+), Js(a+b+). Alternatively, the same phenotype could be stated as: KEL:‐1,2,3,4,(5),6,7. Genotypes, on the other hand, rather mirror the actual biological background, e.g. display the two parental alleles (or “haplotypes”) present in an individuals’ sample. Genotype of the above mentioned example would read: KEL*02.03/02.06 (italicized). In an idealized diagnostic environment and for most blood group systems encoded by proteins, every single blood group allele would be defined by its full genomic sequence, derived from one parental chromosome (including “some” 5’‐ and 3’‐untranslated regions). Thereby, every such “ideal allele” would fully declare presence or absence of all its public, low‐ and high‐frequency antigens and possess its “ideal name”. By trend, biallelic SNPs and their immediate relation to antithetic antigen couples might have distracted from the originally intended meaning of “blood group alleles”, more recently. Finally, genotypes only dependent on (ideal) allele names, and considering Mendelian inheritance patterns (dominant, recessive), would allow for fully comprehensive phenotype predictions.

More recently, blood group serology, e.g. the “second side of the coin”, seems to gain momentum. Since the advent of whole genome sequencing and access to many more than 1000 human genomes, it seems that dozens of new blood group alleles are discovered, almost on a daily basis. Beside the challenge of naming this multitude of alleles, respective discoveries are frequently made in samples lacking any phenotypic blood group pre‐values. Clear procedures will be needed to address naming and analyzing the phenotypes resulting from previously unknown alleles. As a consequence, questions asked 30years ago have changed: Today, molecular biologists looking at hundreds of newly discovered blood group alleles find themselves not being asked by serologists any more: “Can you confirm my serology?”, but instead, pose their question to the experts for the blood group phenotype: “Can you confirm my genotype?”

RHEFERENCE DATABASE: THE COMPREHENSIVE DATABASE FOR RHD VARIANTS

A Floch^1,2, S Teletchea³, F Pirenne^1,2, C Tournamille^1,2, A de Brevern^2,4

¹INSERM (French National Institute for Health and Medical Research) ‐ U955 Team 2 ‐ IMRB ‐ UPEC, French Blood Agency, Creteil ²Labex GR‐Ex, Paris ³CNRS UMR 6286, UFIP ‐ Nantes University, Nantes ⁴INSERM UMR_S 1134, Paris University, Reunion Island University, INTS (Institut National de Transfusion Sanguine), Paris, France

Background: Since the discovery of RH blood group system molecular basis in the 1990s, hundreds of articles have been published, describing novel RHD alleles or adding to the knowledge of known alleles. These works present heterogeneous data: molecular, serological etc. The existing allele lists (ISBT allele tables, Rhesus base website) are valuable but have a limited number of features.

Aims: To develop a comprehensive, elaborate database for RHD alleles allowing complex queries and linking all relevant data.

Methods: We have developed a modern database, named RHeference, based on BioDjango, an open Python framework for bioinformatics. It contains all published information regarding in‐frame RHD alleles, thanks to a comprehensive survey of the scientific literature.

RHeference contains the nucleotide and amino acid sequence for each published Rh1 (D) variant, all nomenclatures, RH antigen expression, ISBT classification (weak D, partial D, Del), data about anti‐Rh1 antibody formation… The sources and citations for all information are included in the allele index card.

Results: The novelty and major strength of RHeference is to allow easy navigation through pre‐calculated queries: starting from molecular data (presence or absence of mutations at specific positions), name, antigen expression, citation… The data can be browsed in a horizontal manner by means of dedicated links.

Summary/Conclusions: With several hundred citations and 443 RHD alleles (to date), RHeference provides an overview of current knowledge regarding the RHD gene. It is an invaluable tool thanks to its features and endless query options. The database will continue to grow as our knowledge expands and new articles are published. New features, especially regarding 3D structure, will be included in the future.

AUTOMATED TYPING OF COMPLEX RH GENOTYPES FROM WHOLE GENOME SEQUENCES

W Lane^1,2, J Halls^1,2,3, S Vege⁴, D Simmons^1,2, B Bujiriri¹, H Mah¹, P Kumar⁵, M Lebo^1,2,5,6, C Westhoff⁴

¹Pathology, Brigham and Women's Hospital ²Harvard Medical School ³Pathology, Beth Israel Deaconess Medical Center, Boston ⁴New York Blood Center, New York ⁵Partners Personalized Medicine ⁶Laboratory for Molecular Medicine, Boston, United States

Background: Rh is one of the most diverse and complex blood group systems. Although serologic methods satisfy most routine Rh typing needs, serology is unable to fully resolve all clinical important phenotypes. Weak and partial phenotypes often require DNA based methods such as SNP and/or Sanger sequencing. Recently, the use of next generation sequencing (NGS) has been reported for blood group typing including complex Rh phenotypes, but analysis is not straightforward and often requires interpretation by subject matter experts.

Aims: We recently developed automated software (bloodTyper) capable of determining RBC antigens from whole genome sequencing (WGS), including validation for common Rh antigens. Here we sought to design and validate bloodTyper for interpretation of samples with complex Rh variation.

Methods: Twenty samples with SNP and Sanger sequencing based RHD genotyping and two with RHCE genotyping were selected to represent a diverse set of complex RHD and RHCE alleles, including hybrids. Archived DNA underwent WGS followed by blinded evaluation using bloodTyper interpretive software. Structural variations (SV) were determined using a combination of three independent detection methods including: sequence read depth of coverage, split reads, and paired reads.

Results: In this targeted dataset, bloodTyper was able to correctly identify D negative (RHD deletion and RHD*Psi), weak D (RHD*weak D Types 1, 2, and 3), partial D (RHD*weak partial D 4.0, *DAR, *DOL, *DIIIa, *DIVa, *DAU0, *DAU3, *DVI types 1 and 2), compound heterozygotes [RHD*weak partial D 4.0/*Psi, *weak partial D 4.0/*DIIIa‐CE(4‐7)‐D, *DIIIa/*DIIIa‐CE(4‐7)‐D, *D/*DIIIa‐CE(4‐7)‐D, *DAU5/*DIIIa‐CE(4‐7)‐D, and *DAU5/*486+1a (Del)], and partial RhCE (RHCE*CeRN/*ce, *ceHAR/*cE). Sequence read depth SV methods accurately determined RHD zygosity and detected the presence of RHD hybrids [RHD*DIIIa‐CE(4‐7)‐D, *DVI type 1, and *DVI type 2]. RH*C could be detected by sequence read depth analysis looking for RHD exon 2 gene conversion into RHCE, as previously reported; however, we also show that the C antigen can be detected by split read SV methods containing RHD intron 2 sequences followed by the unique 109bp insertion characteristic of RH*C. RHD hybrids and RH*C could also be confirmed by detecting paired reads spanning the two intronic breakpoints with one read mapping to RHD and the other to RHCE. RHCE*CeRN and *ceHAR were detected using a paired read SV approach with one misaligned read mapping to the rearranged RHD exon and the other to the RHCE intronic region spanning the breakpoint.

Summary/Conclusions: Automated interpretation of WGS data allowed for accurate RHD genotyping including zygosity using a combination of SV approaches. This approach also successfully interpreted RHD and RHCE complex compound heterozygotes. Such automated analysis would be scalable and could become a routine part of large genomic sequencing projects, allowing for the determination of complex Rh genotypes in an unprecedented number of donors and recipients and help during difficult alloantibody workups.

METAGENOMICS AND BLOOD‐DERIVED PRODUCTS

S Waldvogel‐Abramowki, S Taleb, O Preynat‐Seauve

Geneva University Hospitals, Geneva, Switzerland

Transfusion‐transmitted infections remain a permanent threat in medicine. It keeps the burden of the past, marked by serious infections transmitted by transfusion, and is constantly threatened by emerging viruses. The global rise of immunosuppression among patients undergoing frequent transfusions exacerbates this problem. Over the past decade, criteria for donor selection have become increasingly more stringent. Although routine Nucleic Acid Testing (NAT) for viral‐specific detection has become more sensitive, these safety measures are only valuable for a limited number of select viruses. The scientific approach to this is however changing, with the goal of trying to identify infectious agents in donor units as early as possible to mitigate the risk of a clinically relevant infection. To this end, and in addition to an epidemiological surveillance of the general population, researchers are adopting new methods to discover emerging infectious agents, while simultaneously screening for an extended number of viruses in donors. Next Generation Sequencing (NGS) offers the opportunity to explore the entire viral landscape in blood donors, the so‐called metagenomics, to investigate severe transfusion reactions of unknown etiology. In the not too distant future, one could imagine this platform being used for routine testing of donated blood products. This presentation summarizes the current knowledge regarding this area, with a special focus on the viral landscape described in red blood cells, platelets and plasma for transfusion.

MOLECULAR‐ AND IMMUNO‐ MAGNETIC AGGLUTINATION ASSAYS ON A SINGLE ANALYTICAL PLATFORM: APPLICATION TO DIAGNOSIS OF ARBOVIRAL INFECTIONS

F Leon¹, E Pinchon¹, N Temurok², Jean‐F Cantaloube¹, P Gallian³, V Foulongne¹, M Clos², P Van de Perre¹, A Daynes², J‐P Molès¹, C Fournier‐Wirth¹

¹UMR PCCI Pathogenesis and Control of Chronic Infections (EFS ‐ Inserm ‐ Université de Montpellier), Etablissement Français du Sang ²Innovation and Technology Department, HORIBA Medical, Montpellier ³Etablissement Français du Sang Alpes‐Méditerranée, Marseille ⁴UMR PCCI Pathogenesis and Control of Chronic Infections (EFS ‐ Inserm ‐ Université de Montpellier), CHU de Montpellier ⁵UMR PCCI Pathogenesis and Control of Chronic Infections (EFS ‐ Inserm ‐ Université de Montpellier), Inserm, Montpellier, France

Background: The screening of blood donors and returning travelers from active transmission areas have highlighted the importance of diagnosis of acute arboviral infections. In the context of co‐infections and similar clinical signs in endemic zones, the differential diagnosis of arboviruses is essential to discriminate the causative agent. The detection of viral nucleic acids in serum or plasma provides a definitive diagnosis, however, in most instances, viremia is transient within less than two weeks after the onset of clinical illness. In addition, the cross reactivity due to the high degree of structural and sequence homology between ZIKV and other flaviviruses is a significant concern. The combination of molecular (identification of viral genomes) and immunological assays (detection of the immune response) is a key challenge to follow the natural history of these infections and to improve the patient management and the epidemiological surveillance.

Aims: In this context, we have developed an innovative platform based on agglutination of superparamagnetic nanoparticles (NPMag) covalently grafted either with nucleic or proteic probes to face the continuing emergence of arboviruses.

Methods: Dengue (DENV) and ZIKA (ZIKV) viruses are selected as models in this study. A pan‐flavivirus RT‐PCR is used for the molecular assay to amplify the viral genomes. Then, biotinylated viral amplicons are captured specifically on complementary original polythiolated probes coated on NPMag. For the immunological assay, NPMag are grafted with viral NS1 proteins to capture anti‐DENV or anti‐ZIKV antibodies potentially present in the plasma samples. Both tests are performed in disposable cuvettes in a homogeneous format. A magnetic field generated by an electromagnet is applied to the reaction medium to align the NPMag into chains to enhance the capture of the targets between two NPMag. Aggregates formed are detected when the field is turned off. The optical density is measured in real‐time at 650nm during several cycles of magnetization / relaxation.

Results: In this study, molecular analytical performances were evaluated on human samples from blood donors with no history of infections as negative controls, on viral standards and on clinical samples. Using viral references standards, we have observed sensitivities of 10 ‐ 10² TCID50/mL for ZIKV and DENV (serotypes 1/2/3/4) after a detection phase of around 5min. The first results obtained on 16 ZIKV (+) clinical samples previously tested by commercial real‐time PCR (Ct<36, Altona) showed an 88 % correlation between the two detection methods. No false positive results or cross reactions were observed. Concerning immunological assays, commercial human plasma from donors tested positive for DENV or ZIKV antibodies were detected positive with our innovative approach in less than 10min (sampling + detection) instead of 2h with classical ELISAs. Further assays on clinical samples are planned to confirm these preliminary results.

Summary/Conclusions: This innovative strategy combining molecular and immunoassays on the same analytical platform offers new opportunities for rapid blood testing to improve the surveillance and the prevention of arboviral infections.

ZIKA, CHIKUNGUNYA, AND DENGUE VIRUS INCIDENCE IN BLOOD DONORS IN BRAZIL 2016–2018

B Custer¹, T Goncalez¹, D Brambilla², K Gao³, L Amorim⁴, A Carneiro Proietti⁵, A MendroneJr⁶, P Loureiro⁷, L Capuani⁸, C Alencar⁹, M Stone¹, C McClure², J Linnen³, M Busch¹, E Sabino⁹, B for the REDS‐III International Component¹⁰

¹Vitalant Research Institute, San Francisco ²RTI, International, Rockville, MD ³Grifols, San Diego, CA, United States ⁴Hemorio, Rio de Janeiro, RJ ⁵Hemominas, Belo Horizonte, MG ⁶Fundacao Pro‐Sangue, Sao Paulo, SP ⁷Hemope, Recife, PE ⁸FMUSP ⁹USP, Sao Paulo, SP, Brazil ¹⁰National Heart, Lung, and Blood Institute, Bethesda, MD, United States

Background: Except for surveillance based on clinical case diagnosis, data on the incidence of Zika (ZIKV), Chikungunya (CHIKV) and Dengue (DENV) arboviruses in the population are not available in Brazil.

Aims: The objective of this study was to assess the contemporaneous incidence of these agents in donors at 4 large geographically dispersed blood centers located in the southeast and northeast of Brazil.

Methods: In the Brazil public blood bank system, NAT screening for HIV, HCV and HBV is performed on minipools (MP from 6 donations – MP6). The residual volume of MP6 plasma, 0.35 – 0.45mL, is routinely discarded. Beginning in April 2016 to the present 4 blood centers saved ˜67 MP6 per week for retrospective testing using a qualitative research Transcription Mediated Amplification (TMA) triplex assay for ZIKV, CHIKV, and DENV developed by Grifols/Hologic. MPs were shipped to the USA and batch tested at Grifols. Testing was conducted with the MP6 diluted with added plasma from USA donors to achieve a testing volume equivalent to pool sizes of 18 or by pooling 3 MP6 samples into MP18. Pools were classified as negative or positive based on singleton testing. To estimate the percent viremic donors per month each denominator was adjusted to account for the number of donations included in each pool each month and 95% confidence intervals (CI) were calculated using the method developed by Biggerstaff (data not shown).

Results: This abstract reports on testing conducted between April 2016 and January 2018 comprised of donations from 106,014 donors. The 3 arboviruses were detected in donors in different geographic locations in Brazil during the late rainy season in 2016 (April – June). DENV viremia was found in Sao Paulo, Belo Horizonte, and Rio de Janeiro in 2016 with peak estimated monthly viremia of 416 per 100,000 donors in Belo Horizonte. CHIKV viremia was found in Sao Paulo and Recife in 2016 with peak monthly viremia of 426 per 100,000 donors in Recife. CHIKV was also detected in Rio de Janeiro in 2017 and in Recife in January 2018. ZIKV viremia was found in Belo Horizonte and Rio de Janeiro in 2016 and 2107 with peak monthly viremia of 638 per 100,000 donors in both locations. ZIKV viremia was detected in MP samples for donors tested from Rio de Janeiro in 12 out of 19months, including in January 2018 at 116 per 100,000 donors, and also detected outside of the seasonal rainy period in both 2016 and 2017.

Summary/Conclusions: The three arboviruses were circulating in Brazil at the same time in 2016. The testing for ZIKV in this study occurred after the major outbreak in the northeast part of Brazil had waned and this likely explains why we did not detect ZIKV in Recife during the study period. The ZIKV detection pattern in Rio de Janeiro suggests that transmission routes other than by mosquito, such as sexual transmission, may be maintaining the virus in circulation for longer periods of time.

VECTOR‐BORNE DISEASES AND BLOOD TRANSFUSION SAFETY: A EUROPEAN PERSPECTIVE

M Janssen, R Lieshout, R Garzon Jimenez

Donor Medicine Research, Sanquin, Amsterdam, Netherlands

Background: Arboviral infectious diseases have been identified as a threat to human health. In 2002 they became a concern for blood transfusion safety as well. Despite the implementation of safety measures to prevent transmission by blood transfusion, large variation exists regarding the perception of the necessity for such measures. This variation may be caused by a lack of understanding of the risk of transmission by blood transfusion.

Aims: To estimate the risk of transfusion‐transmitted arboviruses in travelling donors to affected areas within Europe and identify the risk perception in power stakeholders.

Methods: An integrated EUFRAT based model was developed to calculate the transfusion‐transmission risk for Europe of West Nile virus (WNV), chikungunya and Zika virus outbreaks. Data on blood supply characteristics from countries within Europe was obtained from the 2013 to 2015 reports from the Council of Europe. Additionally, inter‐European travel information from 2013 to 2016 was obtained from EUROSTAT. A European average number of transmissions by blood transfusion per infectious case observed was estimated for arboviral outbreaks.

In collaboration with the European Blood Alliance, 13 participants from 12 different organisations involved in blood safety decision‐making were invited for a semi‐structured interview. Selection of participants was based on either stakeholder‐group membership, activities developed, or country of origin (endemicity of viruses). Primary (within groups) and secondary (between groups) data comparisons were performed to identify similarities and differences in risk perception and factors influencing these perceptions.

Results: The European WNV travelling donor's transmission rate (R0) was estimated at 7.1E‐03 infected individuals per neuro‐invasive case observed. The required number of infections to obtain one infected blood product in Europe with chikungunya and WNV was respectively 6238 and 140.

The qualitative study showed that the perception of transfusion transmission risk of arboviruses is low for WNV and very low for chikungunya and Zika virus. This perception is predominantly influenced by 5 factors: (1) the presence of a competent vector and viral circulation; (2) evidence of being a transfusion transmissible disease; (3) health complications associated with the disease; (4) availability of measures for prevention and control; and (5) influence of‐ and interaction between players involved in the process.

Summary/Conclusions: Despite the low perception of blood transfusion risk among the stakeholders, the lack of active participation of some of them in the decision‐making process, leads to differences in the implementation of blood safety directives and recommendations. The use of scientific evidence in the decision‐making processes becomes more important as this may lead to more transparency and uniformity in decisions made. The model developed can be easily applied to determine the risk for blood transfusion throughout Europe due to travelling‐donors, but also provides decision‐makers with further evidence for a rational approach for the management of blood safety.

APPLICATION OF A ZIKA VIRUS SEROLOGICAL ASSAY TO EVALUATE INFECTION RATES IN DONORS FOLLOWING THE 2016 EPIDEMIC IN PUERTO RICO

G Simmons^1,2, M Stone^1,2, C Cheng¹, P Williamson³, M Busch^1,2

¹Vitalant Research Institute ²Laboratory Medicine, University of California, San Francisco, San Francisco ³Creative Testing Solutions, Tempe, United States

Background: Zika virus (ZIKV) caused a dramatic epidemic in Puerto Rico (PR) during 2016, with up to ˜2% of blood donors reactive for ZIKV RNA in ID‐NAT testing at the peak in June 2016.

Aims: Perform a serosurvey for anti‐ZIKV IgG using six panels of 500 donor specimens each collected in March 2015, at the beginning, peak and end of the 2016 epidemic, and from March 2017 and April 2018.

Methods: We employed a commercially available ZIKV IgG ELISA antibody (Ab) assay based on the ZIKV NS1 antigen from Bio‐Techne to characterize ZIKV seroprevalence in the 6×500 cross‐sectional sample sets (anonymized with selected demographic information).

Results: 500 PR donor samples collected in April 2015 were initially evaluated using the manufacturer supplied cut‐off to confirm that the ZIKV Ab results were largely negative (3 positive, 3 equivocal) despite the high dengue virus seroprevalence (>90%) in PR that could potentially lead to false positive ZIKV Ab results. We then used this dataset, together with known positives collected 1–3months post‐detection from ZIKV NAT yield donors, to set a population‐specific cut‐off based on receiver operating characteristic (ROC) curve analysis. This cut‐off yielded sensitivity and specificity values of>99%, and an area under the curve (AUC) of 0.999, demonstrating a highly accurate assay. We used this new cut‐off to calculate final rates of seroreactivity in the additional 5 sample sets (2500 samples) and estimate seasonal incidence. Rates of reactivity, together with mean net OD for only the reactives (shown in parentheses), were calculated for each sample set: March 2015: 0.6% (0.4); April 2016: 4.1% (1.6); June 2016: 9.0% (1.6); October 2016: 17.8% (2.0); March 2017: 23.0% (1.3); and April 2018: 16.2% (0.7).

Summary/Conclusions: The peak seroprevalence of 23% shortly after the epidemic (March 2017) is consistent with our estimate of 22% seasonal incidence in PR during the 2016 outbreak derived from the yield of ZIKV RNA reactive blood donors and duration of the NAT detection period (Chevalier et al. Emerg Infect Dis (2017)23:790; Williamson et al., submitted). Another interesting finding was the rapid waning of ZIKV Abs between 2017 and 2018, both in terms of percentage of reactive donations and mean OD signal in reactive samples. These findings have implications for performing serosurveys in other populations and for potential risk of ZIKV reinfections in donors and pregnant women during recurrent outbreaks.

BENEFITS OF THAWED PLASMA IN BLOOD SUPPLY MANAGEMENT

JJ Zwaginga^1,2

¹Immunohematology and Bloodtransfusion, Leiden University Medical Center ²Sanquin‐ LUMC Center for Clinical Transfusion Research, Sanquin Research, Leiden, Netherlands

While in trauma patients major blood loss still accounts for 30‐40% of early deaths and postpartum haemorrhage for 1/5th of maternal death, optimizing transfusion support is of extreme importance. Bleeding patients in this respect should be immediately evaluated if at risk for or already having massive blood loss with associated morbidity and mortality. If so namely, early as possible initiation of massive transfusion protocols (MTP) to preserve their haemostatic capacity with possibly limiting blood loss and mortality seems justified. Most optimally, initiation of transfusions to trauma patients takes about 20minutes, however, in the majority of patients this will take more than an hour. Next to the associated delay in damage control intervention, this delayed initiation of transfusion will add to morbidity. In the meantime colloid and crystalloid mediated preservation of haemodynamics, namely will dilute the also by other factors compromised haemostatic system. In this respect, also blood product‐related solutions to tackle this problem are justified.

Against this background, we will first give guidelines which bleeding patients justify initiation with MTP and which product (ratios) should best be given. Second, we will review evidence on the mortality‐modifying effect of early initiation of such protocols. Thirdly, next to storage and associated direct availability of blood products in the trauma room and surgery ward, we will discuss the present evidence and ratio for organizing pre‐hospital use of cryopreserved or lyophilized plasma or fibrinogen.

CHALLENGING THE 30‐MINUTE RULE FOR THAWED PLASMA

S Ramirez‐Arcos¹, J Allen², V Bhakta³, L Bower⁴, R Cardigan⁴, M Girard⁵, A Howell⁶, Y Kou¹, C McDonald², M Nolin⁵, D Sawicka⁴, W Sheffield³

¹Center for Innovation, Canadian Blood Services, Ottawa, Canada ²NHS Blood and Transplant, London, United Kingdom ³Center for Innovation, Canadian Blood Services, Hamilton, Canada ⁴NHS Blood and Transplant, Cambridge, United Kingdom ⁵Hema‐Quebec, Quebec ⁶Center for Innovation, Canadian Blood Services, Edmonton, Canada

Background: Frozen plasma (FP) is manufactured within 24h of blood collection and stored at≤‐18^°C for up to 12months. FP thawing is performed prior to transfusion. In Canada and the UK, thawed plasma is stored for 5days at 1–6^°C and 2–6^°C, respectively. The ‘30‐minute rule’ for red blood cells (RBC) requires the discard of units that have been exposed to uncontrolled temperature for more than 30min during storage. This rule is applied to other products including FP, without evidence‐based data, leading to product wastage. The 30‐minute rule for RBC has been recently extended to a 60‐minute rule in Canada and the UK.

Aims: Determine the effect of temperature excursions on the quality and safety of thawed plasma during 5‐day storage.

Methods: A multi‐center study was conducted with quality and sterility arms each performed at NHS Blood and Transplant, Héma‐Québec, and Canadian Blood Services. Groups of four ABO matched plasma units were pooled, split, and stored at≤‐18^°C for≤90days. Each group of units comprised two test units (T30 and T60), which were exposed to room temperature (RT) for 30 or 60min, respectively, on day 0 (thawing day) and day 2 of storage; a negative control (NC) unit, which remained under refrigeration for 5days; and, a positive control (PC) unit, which was stored at RT for 5days. On day 5, the T30 and T60 units were exposed once to RT for 5h. Plasma samples were taken for testing prior each exposure. For the quality assays, stability of coagulation factors FV, FVII, FVIII, fibrinogen, and prothrombin time was measured. Bacterial growth experiments were performed in thawed plasma units inoculated with ˜1CFU/ml or ˜100CFU/ml of Serratia marcescens, Serratia liquefaciens, Pseudomonas putida or Staphylococcus epidermidis on day 0 and then stored under different conditions as described for the quality assays. Statistical analyses were performed to compare results between conditions at each testing center.

Results: Quality assays did not show significant differences between T30 and T60 units for any of the coagulation factors in samples taken after a 5‐h exposure to RT on day 5. Bacterial growth was observed in the PC units inoculated with either concentration for all species. The exception was P. putida, which self‐sterilized in all units tested at Héma‐Québec. NC, T30 and T60 units inoculated with 100CFU/ml showed bacterial survival but not proliferation during 5days of storage while units inoculated with 1CFU/ml showed variable survival and self‐sterilization within the three sites. There was no difference in bacterial concentration between T30 and T60 units for any of the species in any of the testing sites.

Summary/Conclusions: Multiple RT exposures for either 30 or 60min do not affect the stability of coagulation factors or promote bacterial growth in thawed plasma stored for 5days, even in the worst‐case scenario after a one 5‐h exposure to RT. These data suggest that it is safe to expose thawed plasma to uncontrolled temperatures for periods of 60min as demonstrated for RBC.

SEX HORMONE INTAKE IN FEMALE BLOOD DONORS MODULATES RED BLOOD CELL SUSCEPTIBILITY TO SPONTANEOUS AND STRESS‐INDUCED HEMOLYSIS DURING COLD STORAGE

F Fang¹, K Hazegh², D Sinchar³, Y Guo¹, G Page⁴, A Mast⁵, S Kleinman⁶, M Busch⁷, T Kanias²

¹Division of Biostatistics and Epidemiology, RTI International, Durham ²Vitalant Research Institute, Vitalant, Denver ³Vascular Medicine Institute, University of Pittsburgh, Pittsburgh, PA ⁴Division of Biostatistics and Epidemiology, RTI International, Atlanta, GA ⁵Blood Research Institute, Blood Center of Wisconsin, and Department of Cell Biology, Neurobiology and Anatomy, Medical College of Wisconsin, Milwaukee, United States ⁶Department of Pathology and Laboratory Medicine, The University of British Columbia, Vancouver, Canada ⁷Vitalant Research Institute, Vitalant, San Francisco, CA, United States

Background: Sex hormone intake in blood donors occurs in three demographic groups: premenopausal women who take contraceptive drugs (progestins with or without estrogens), postmenopausal women who receive estrogen replacement therapies that may be combined with progestins, and testosterone therapies in men. We hypothesized that sex hormone therapies may modulate the quality of red blood cell (RBC) products via alterations of RBC function and predisposition to hemolysis during cold storage.

Aims: The objectives of this study were to evaluate the association between sex hormone intake and RBC measurements of hemolysis, and to examine possible mechanisms by which sex hormones interact with RBCs.

Methods: Self‐reported sex hormone intake and menstrual status were evaluated in 6,636 female blood donors from theNational Heart, Lung and Blood Institute's RBC‐Omics study. The associations between hormone intake and donor scores of spontaneous storage, osmotic, or oxidative hemolysis were determined in all women and by menstrual state. The interactions between sex hormones and RBCs were determined by sex hormone (progesterone,17β‐estradiol,or testosterone) potency to inhibit calcium influx or hemolysis during incubations or cold storage. The calcium fluorophore, Fluo‐3AM, was used to define RBC calcium influx in response to treatments with sex hormones or drugs that modulate transient receptor potential cation (TRPC) channel activity including Hyp9 (a selective TRPC6 activator).

Results: Sex hormone intake by menstrual status was higher inpremenopausal women (25.3%) than in postmenopausal women (7.4%). Female hormone intake was significantly (all P<0.0001) associated with reduced storage hemolysis in all females (0.32±0.16% versus0.35±0.29% in controls), enhanced susceptibility to oxidative hemolysis (37.9±9.3% versus35.8±9.9% in controls), and reduced osmotic hemolysis in postmenopausal women (23.1±10.2% versus 26.8±12.0% in controls). In vitro, supraphysiological levels of progesterone (10 or 20μmol/L), but not 17β‐estradiol or testosterone, inhibited spontaneous or Hyp9‐induced calcium influx into RBCs, and were associated with lower spontaneous hemolysis after 30day cold storage (0.95±0.18% versus 1.85±0.35%, progesterone 10μmol/L versus solvent control (dimethyl sulfoxide, 0.1%), P<0.0001). Co‐incubations (2.5h, 37°C) of RBCs in the presence of progesterone and a TRPC6 activator (Hyp9, 25μmol/L) suggested that progesterone protected against Hyp9‐induced hemolysis (1.45±0.13% and 1.01±0.09% versus 2.63±0.19%; Hyp9+progesterone at 10 or 20μmol/L versus Hyp9 alone, P<0.05 by one‐way ANOVA).

Summary/Conclusions: This study revealed that sex hormone intake in blood donors is capable of modulating RBC predisposition to hemolysis and led us to propose new mechanistic pathways by which progesterone regulates calcium influx and hemolysis in human RBCs.Pre‐ and postmenopausal women respond differently to hormone intake and its effects on RBC responses to osmotic or oxidative stress. Progesterone modulates calcium influx into RBCs via a mechanism that may involve interactions with membrane TRPC6 channels, activation of which is associated with pre‐hemolytic events such as senescence and eryptosis.

DEHT PLASTICIZER PRESERVES RED BLOOD CELL QUALITY WELL DURING STORAGE IN PVC BLOOD BAGS

L Larsson^1,2, S Ohlsson¹, J Derving¹, P Sandgren^1,3, M Uhlin^1,2

¹Department of Clinical Immunology and Transfusion Medicine, Karolinska University Hospital ²Department of Clinical Science, Intervention and Technology (CLINTEC) ³Department of Laboratory Medicine, Karolinska Institutet, Stockholm, Sweden

Background: Polyvinyl chloride (PVC) plasticized with di(2‐ethylhexyl) phthalate (DEHP) is the material of choice for commercial blood bags since the mid‐20th century. Plasticizers are essential for material flexibility. In addition, DEHP is favourable for storage of red blood cells (RBCs). Historically, removal of DEHP from blood bags has been linked to unacceptable haemolysis levels during storage.

Concerns are, however, raised for the potential toxicity of DEHP. Oncoming stricter regulations by the European Commission, allowing maximum 0.1% DEHP in medical devices, increases the urgency to find a replacement plasticizer to DEHP that do not compromise RBC quality. Di(2‐ethylhexyl) terephthalate (DEHT) is one suggested substitute.

Aims: The aim of this study is to compare PVC‐DEHT to PVC‐DEHP blood bags using two different additive solutions (AS), to determine whether blood bags plasticized with DEHT can maintain acceptable RBC quality over time.

Methods: Sixty‐four leukoreduced RBC concentrates were produced from 450ml (± 10%) whole blood collected directly into either DEHT or DEHP plasticized PVC blood bag systems (Macopharma, Mouvaux, France). Using a pool‐and‐split study design, pairs of identical RBC contents were created within each plasticizer arm. One of the splits was assigned AS saline‐adenine‐glucose‐mannitol (SAG‐M) for storage, and the other phosphate‐adenine‐glucose‐guanosine‐saline‐mannitol (PAGGS‐M), generating four study arms: DEHT/SAG‐M, DEHT/PAGGS‐M, DEHP/SAG‐M and DEHP/PAGGS‐M (N=16 per arm). Bags of different plasticizer were handled separately and did not touch anything containing the other plasticizer throughout the study.

During 49days, the RBC concentrates were sampled weekly for assessment of RBC storage lesion through analysis of haemolysis, mean corpuscular volume (MCV), pH and concentrations of extracellular potassium (K⁺), extracellular sodium (Na⁺), ATP, glucose and lactate. Mean ± standard deviation was calculated, and statistical significance was tested using repeated measures 2‐way ANOVA.

Results: Haemolysis was higher in DEHT than in DEHP from day (d) 14 onwards (P<0.01), without AS influence until d28. From d35, all four study arms differed (P<0.05), with higher haemolysis in SAG‐M than in corresponding PAGGS‐M arm. At end of storage, d49, DEHT/SAG‐M displayed the highest haemolysis, 0.39±0.03%. DEHT/PAGGS‐M reached 0.27±0.03%, whereas DEHP/SAG‐M, current reference blood bag in most European countries, reached 0.23±0.04%. DEHP/PAGGS‐M was lowest, 0.18±0.04%.

K⁺ concentration remained lower in DEHT than in DEHP independently of AS from d28 onwards (P<0.01). Concentrations at d49 were 45.5±1.4 (DEHT/SAG‐M), 44.7±1.7 (DEHT/PAGGS‐M), 48.4±1.7 (DEHP/SAG‐M) and 48.7±1.45 (DEHP/PAGGS‐M) mmol/L respectively. Reversely, Na⁺ had a more distinct relation to AS than plasticizer.

Neither pH, MCV, glucose, lactate nor ATP was affected by choice of plasticizer (ns). A consistent higher metabolism rate was however seen during storage in SAG‐M.

Summary/Conclusions: Our study demonstrates that PVC blood bags plasticized with DEHT provide adequate RBC quality. Haemolysis levels remain well beneath the EDQM limit (0.8%) even after 49days of storage, in both SAG‐M and PAGGS‐M. The lower extracellular K⁺ levels during storage compared to PVC‐DEHP strengthen the satisfactory conclusion.

We conclude that DEHT is a strong future candidate for replacement of DEHP in RBC storage.

PROTEASOME MODULATION IN STORED RBCS: STORAGE AGE AND DONOR EFFECTS

V Tzounakas¹, M Dzieciatkowska², A Anastasiadi¹, D Karadimas¹, A Vergaki³, P Siourounis³, I Papassideri¹, A Kriebardis⁴, A D'Alessandro², M Antonelou¹

¹Department of Biology, School of Science, National and Kapodistrian University of Athens, Athens, Greece ²Department of Biochemistry and Molecular Genetics, University of Colorado Denver, School of Medicine, Aurora, CO, Colorado, United States ³Regional Blood Transfusion Center, “Agios Panteleimon” General Hospital of Nikea, Piraeus ⁴Department of Biomedical Sciences, School of Health and Caring Sciences, University of West Attica (UniWA), Egaleo City, Greece

Background: Proteasome is a fundamental multicatalytic enzyme complex in the protein homeostasis system. It affects stress responses, cell aging and lifespan. Proteasomes are retained within RBCs during maturation and display activity in vitro. However, the physiological function of RBC proteasomes and their modulation by genetic factors and aging, both in vivo and in RBC units have not been studied in depth.

Aims: The goals of this study were to gain insight into the composition, activity and subcellular/extracellular distribution of RBC proteasomes, and to their variation during storage of RBCs from genetically distinct donor groups.

Methods: RBCs from sixteen male donors (6 with G6PDH deficiency, class II Mediterranean variant) were studied before and during storage in leukoreduced CPD‐SAGM units. Proteasome components were analyzed by proteomics and immunoblotting of RBC membrane and extracellular vesicles. Proteasome activity assays were performed in membrane, cytosol and plasma samples and in transfusion mimicking conditions by using fluorogenic peptide substrates. Oxidative stress was measured by fluorometry. The statistically important correlations of proteasomal factors were topologically represented in undirected networks.

Results: 38 different proteasome‐associated components were identified by proteomics analysis in the membrane of stored RBCs. The 20S components (including the catalytic subunits b1, b2 and b5) were predominant, showing increased levels at the middle of storage. There was a trend for higher and earlier membrane binding of proteasomal components in G6PD^‐ vs. control RBCs throughout the storage period. Individual donor analysis revealed substantial inter‐donor differences in the levels of membrane association, and a triggering effect for storage. LLVY (b5, chymotrypsin‐like) and LRR (b2, trypsin‐like) were the predominant proteasomal activities in the control cytosol, that decreased after the middle of storage, in contrast to the LLE activity (b1, caspase‐like) that was not affected by storage. A different pattern was present in the G6PD^‐ cytosol, with higher LLVY and LRR levels but a storage‐driven drop in LLE activity. The membrane levels of the proteasome activities were not affected by storage in control RBCs. In G6PD^‐ cells however, increased levels of LLVY and LLE were detected at the middle of storage compared to either the control cells or the late storage. Higher membrane‐versus‐cytosol LLE activity was found in control and G6PD^‐ RBCs soon after storage. However, the G6PDH deficiency was associated with elevated levels of membrane than cytosolic activities, which had strong correlations with metabolites of the glycolysis, PPP and GSH pathways. Extremely low LLVY activity was detected extracellularly, that increased however, in proportion to storage time in both groups, with higher activity measured in the G6PD^‐ supernatant. When reconstituted with healthy plasma at transfusion mimicking conditions, the stored RBCs exhibited increased LRR and LLE cytosolic activities (and lower ROS) compared to the baseline (RBC unit) but lower LLE activity at the membrane compared to the cytosol.

Summary/Conclusions: The intracellular distribution of proteasomal proteins and the levels of proteasome activities are modulated in RBCs as a function of extracellular environment (temperature, antioxidant status, plasma vs. supernatant), storage age and G6PDH deficiency, in close association with metabolic and proteome stress factors.

REACH AND RETAIN DONORS IN RESOURCE POOR SETTINGS

L Indermuehle

International Cooperation, Swiss Red Cross, Wabern, Switzerland

Background: Red Cross and Red Crescent Societies were playing an important role in setting up blood transfusion establishments in many low resource countries. By the mid‐1970s, the Red Cross was active in the national blood programs in approximately 95% of countries – mostly in blood donor recruitment and education. Today, major organizational developments in blood transfusion services were made in high income settings, where nearly 50% of all worldwide donations take place (home to only 17% of the population). WHO data shows that the median annual blood donation rate in high‐income countries is 3.21% of the population compared to 0.46 % in low‐income countries. The factors for this low turnout are multilayered, but it is well‐known that most resource poor settings suffer from a low rate of regular donors and challenges to set‐up and financially sustain a national blood donor program. The Red Cross and Red Crescent (RC) Societies assume a key role by reaching and retaining donors from the communities and contribute significantly to better safety and availability of blood. Partnerships and international collaboration, such as the Swiss Red Cross (SRC) program, are aiming to strengthen national structures to improve blood safety and to face today's epidemiological, demographical, and technological challenges.

Aims: The present work aims to review the role, the mandate and the impact of RC Societies in improving blood safety through systematic “Voluntary Non‐Remunerated Regular Blood Donor” (VNRBD) programming and international partnerships.

Methods: Data and evidence is drawn from the SRC International Cooperation projects over the last 30years, more specifically partnering with three RC Societies, and the data from the Global Advisory Panel (GAP) of the IFRC including their 2018 Global Mapping.

Results: The promotion of VNRBD has been a specific objective in all SRC supported programs. Through the engagement of the RC Society, the training of volunteers and partnering with the health authorities, the projects significantly increased the blood donor rate by recruiting and retaining donors from the communities. For example, the RC Societies increased total donations by 25 % in Lebanon; VNRBDs by 71% annually in Kirgizstan, and from practically zero to 3'588 in South Sudan. The importance of RC Societies was also underlined in the 2018 published global mapping of GAP, which showed that 36 (19%) of them provide level A (Full Blood Service), 75 (39%) are Level B (Systematic Blood Donor Recruitment) and 47 (25%) are level C (VNRBD Blood Promotion) blood services. GAP has also commenced a new three year VNRBD support program aimed at establishing tools and materials for National Societies.

Summary/Conclusions: The Red Cross / Red Crescent Movement has a unique mandate and position in improving global blood safety at all levels; with its huge network of volunteers, even blood donors in remote communities can be reached and retained. RC Societies in low resource settings with a Level B or Level C role should further capitalize on partnerships with local and international actors to leverage technical assistance and funding for their activities.

RECRUITMENT OF FUTURE BLOOD DONORS FOR SUPPLYING THE SAFEST BLOOD AND NATIONAL SELF SUFFICIENCY OF BLOOD SUPPLY

T Ilbars¹, J Anagnan², M Jankulovska², I Yenicesu², P Heimer², N Solaz², D Yuce², K Ilisulu², M Piskin², J Jankulovski², N Togo², S Guzel³, B Durmus³, M Kocakoglu³, I Yildiz³, S Canpolat⁴, M Kalender⁴, L Sagdur⁴, S Ozden⁵, U Kodaloglu Temur¹, N Ertugrul Oruc⁶, M Ozturk¹, A Kapuagasi¹

¹General Directorate of Health Services, Ministry of Health ²WYG Turkey Consultancy ³General Directorate of Secondary Education, Ministry of National Education ⁴General Directorate of Blood Services, Turkish Red Crescent ⁵General Directorate of Health Promotion, Ministry of Health ⁶MoH Diskapi Yildirim Beyazit Training and Research Hospital, Ankara, Turkey

Background: It is essential to motivate and encourage the public to donate blood and be eager to help saving lives, in order to maintain safe and adequate national blood supply.

Aims: “Technical Assistance for Recruitment of Future Blood Donors (EuropeAid/132420/D/SER/TR)” project aimed to avoid problems in supplying the safest blood to contribute to the improvement of public health by (i) increasing the knowledge of primary and secondary school students regarding blood donation, (ii) creating sensitivity in school principals and teachers regarding voluntary non‐remunerated blood donation (VNRD), (iii) motivating family members of the students for blood donation and (iv) creating public awareness through media.

Methods: An effective coordination is established between Ministry of Health (MoH), Ministry of National Education (MoNE) and Turkish Red Crescent (TRC). The existing curricula and textbooks of primary and secondary schools were reviewed and revised, and corresponding materials on the importance of blood donation were created. The human resources capacity of MoH, MoNE and TRC to support raising awareness on blood donation were developed. To raise public awareness on blood donation nationwide, education and recruitment campaigns on were organized in 500 pilot schools. Additionally, media and public relation campaigns on blood donation were organized throughout the country.

Results: (1) Existing curricula and textbooks relevant to promoting blood donation were reviewed, revised and reported to the Board of Education of MoNE. (2) Corresponding educational materials for students and teachers were developed and distributed. (3) Blood Donation Clubs were established in 500 pilot schools. (4) Trainings were conducted for 688 personnel of MoH and TRC on blood donation regarding their responsibilities. (5) Cascade trainings were conducted for 3218 personnel of transfusion centers and 4399 school principals in 81 provinces. (6) Information seminars were delivered to 251.475 students and 15.739 teachers and family members of students during school campaigns. (7) Four animation films on blood donation were produced and broadcasted on the national TV channel (TRT). (8) Three different computer games targeting different age groups were developed and uploaded to the web portal of the project and distributed to the pilot schools. (9) Media spots were produced and broadcasted 3.851 times in 33 different TV and Radio channels. (10) Billboard posters and brochures were prepared and distributed to 81 provinces for raising public awareness. (11) Advertisements about the project and the importance of VNRD were displayed 386 times on national and local newspapers, 1.117 times on online news, and broadcasted on 6 national TV channels. (12) During the campaigns, 28.310 units of whole blood were collected in pilot schools. (13) Visibility kits to recruit future blood donors are prepared and distributed throughout the project activities. (14) Awareness and knowledge level of students and their teachers/parents on the importance of blood donation are increased to 32.07 % and 35.99 % respectively, assessed through pretest and posttest. Voluntary non‐remunerated donation rate of national demand increased from 73.66% to 82.54 in two years.

Summary/Conclusions: Training and campaign programmes successfully increased the knowledge on blood donation. To achieve national self‐sufficient safe blood supply, efforts for recruitment should be continued.

EFFECTIVE METHODS FOR REACTIVATING INACTIVE BLOOD DONORS: A STRATIFIED RANDOMIZED CONTROLLED STUDY

J Ou‐Yang¹, C Bei², H Liang², B He², J Chen², X Rong², Y Fu²

¹Department of Blood Source Management ²Guangzhou Blood Center, Guangzhou, China

Background: Blood donation is the only source of blood products that recruiting enough blood donors is vital for every country.

Aims: In this study we assessed the efficacy and cost‐effectiveness of telephone calls and SMS reminders for re‐recruitment of inactive blood donors.

Methods: This single‐centre, non‐blinded, parallel randomized controlled trial in Guangzhou, China included 11,880 inactive blood donors whose last donation was between January 1 and June 30, 2014. The donors were randomly assigned to either one of two intervention groups (received telephone or short message service [SMS] communications) or to a control group without intervention. SMS messages with purely altruistic appeal were adopted in the SMS group; in addition to altruistic appeal, reasons for deferral of blood donation were also asked in the telephone group. All participants were followed up for 1year. The primary outcome was re‐donation rates, compared by both intention‐to‐treat (ITT) and as‐treated (AT) analyses. Secondary outcomes were the self‐reported deterrents. Other outcomes included the re‐donation interval, the efficacy of each intervention, and the cost‐effectiveness of telephone calls versus SMS reminders on re‐recruitment.

Results: ITT analysis discovered no significant differences in the re‐donation rate among the three groups. However, AT analysis showed that the re‐donation rate was significantly higher in the telephone group than in the SMS group (11.7% vs. 8.6%, P=0.002) and in the control group (11.7% vs. 7.4%, P<0.001). Donor return behaviour was positively associated with receiving reminders successfully(odds ratio: 1.56, 95% confidence interval [CI]: 1.30–1.88, P<0.001), age (1.03, 95% CI: 1.02–1.04, P<0.001) and donation history (1.08, 95% CI: 1.06–1.10, P<0.001). The SMS reminder prompted donors to return sooner than no reminder within 6months (P=0.006), and it was much more cost‐effective than telephone calls (71.7 times cheaper). Time constraints, medical issues, and group‐sponsored donation were the main causes of self‐deferral in the telephone group. The altruistic appeal had a positive effect among donors who reported time constraint as the reason for deferral.

Summary/Conclusions: Interventions to reactivate inactive blood donor can be effective, with telephone calls prompting more donors to return but at a greater cost than SMS messages. SMS reminder with purely altruistic appeal can urge donors to re‐donate sooner within 6months than no reminder.

PROMOTION OF VOLUNTARY BLOOD DONATIONS IN PAKISTAN THROUGH THE FACEBOOK BLOOD DONATION FEATURE

H Zaheer, U Waheed, S Tahir, K Nasir

Safe Blood Transfusion Programme, Ministry of National Health Services, Government of Pakistan, Islamabad, Pakistan

Background: Despite 70% of Pakistan's population being under 29years, only 10% of blood supplies come from voluntary donors while remaining blood is collected from ‘Family Replacement Donors’. In Pakistan the system has outsourced the mobilization of blood donors to the patient families. As a result many people reach out to their networks including on Facebook to locate blood donors. There are thousands of posts each month in Pakistan seeking blood donors on Facebook.

To facilitate needy families, the global social media giant Facebook launched a special blood donation Feature for Pakistan in collaboration with SBTP, Pakistan. The Feature makes it easier for people to sign up to become blood donors and helps connect these voluntary donors with people and organizations in need of blood.

Similar Features have been launched by Facebook in India, Bangladesh and Brazil to address the problem of blood shortages in those countries. However, among these four countries Pakistan has unique position because of the existence of a national counterpart, SBTP which can facilitate Facebook in promoting its Feature and provide the feedback on the impact of this innovative effort for continuous improvement of the Feature.

Aims: To promote voluntary blood donations and blood safety in Pakistan through Facebook.

Methods: The Facebook and SBTP teams launched a pilot to study the impact and effectiveness of the Facebook Blood Donation Feature as a tool of community engagement. A six months plan has been chalked out to measure the impact of this tool in five selected blood centres. A checklist called “P0 checklist” was shared with these blood centres to fulfill some basic requirements for an official blood bank page including a display picture, cover photo, contact information, directions, etc. Regular Skype meetings are held between the teams of SBTP, Facebook (San Francisco and Singapore) and the blood centres to monitor the progress of the Pilot and generate feedback.

Results: The Facebook Blood Donation Feature has recorded remarkable success with over one million signups within few months in Pakistan. The blood centres participating in the Pakistan study have experienced enhancement in the voluntary blood donations trend with 3–10 walk‐in donors and an average of more than 20 telephonic queries regarding voluntary blood donation per month in each center. The trend is gradually surging as the Feature is being refined on the basis of feedback received. The Pilot will end in June 2019. The statistics generated since January 2019 are very encouraging and underscore the importance of social media in reaching out to the untapped potential blood donors. The study will be used to plan an effective nationwide strategy to increase donor mobilization, recruitment and retention.

Summary/Conclusions: Pakistan Facebook Feature for Voluntary Blood Donations can go a long way in shifting the current reliance of the system from the Family Donors to the internationally recommended voluntary blood donors. This paradigm shift in Pakistan can be emulated in other developing countries which face shortages of voluntary donors. The Pakistan experience can potentially modify global strategies for donor mobilization in future in a most cost effective manner.

RECRUITMENT OF BLOOD DONORS OF NON‐CAUCASIAN ETHNICITY: THE EXPERIENCE OF A SWISS BLOOD DONATION CENTER

L Infanti^1,2, T Ruefli¹, V Pehlic¹, M Argast¹, H Luescher³, N Borer¹, C Nobs¹, S Arnold¹, A Holbro^1,2, A Buser¹

¹Blood Donation Center Swiss Red Cross Basel ²Division of Hematology ³University Hospital, Basel, Switzerland

Background: In our region, an increasing number of patients of African or Asian origin with sickle cell disease (SCD) or transfusion dependent thalassemia (TDT) require red blood cell (RBC) transfusions, and many have RBC alloantibodies. Selecting optimally matched RBC units for these patients is essential for preventing not only acute hemolysis but also further alloimmunisation. Beside antigen‐matching for ABO, Rh D, C, c, E, e and K, patients with SCD and TDT should ideally receive RBC units matched also for M, S, s, Fya, Fyb, Jka and Jkb (extended phenotype). This is the policy at our center, which currently provides RBC products to 31 patients with hemoglobinopathies.

Because the vast majority of our blood donors are Caucasians, the selection of matched RBC units for patients of different ethnic origin can be difficult. Therefore, expanding the number of available African and Asian blood donors is becoming increasingly necessary.

Aims: Herebytherecruitment strategy of non‐Caucasian blood donors introduced at our center is described andthe results obtained duringsix years are reported.

Methods: Since 01.01.2013, whenever a first‐time blood donor of non‐Caucasian origin is registered, an alert is entered in the donor file to trigger the determination of the extended RBC phenotype along with routine testing. RBC antigen determination is performed in our laboratory with serologic methods. In selected cases (i.e. suspected RhD or RhCE variant), samples are sent for molecular analysis (SSP PCR). Rare RBC phenotypes relative to ethnicity are, among others, Fy(a‐b‐), S‐ s‐, Lu(b‐) and those with uncommon Rh phenotypes. If a rare RBC phenotype is detected, a coded comment is entered in the donor data and the donor is listed in the national Rare Donor File.

Results: From 01.01.2013 until 01.03.2019, an extended determination of RBC antigens was performed in 352 subjects presenting for blood donation. Twenty‐nine rare donors (8%) were identified and included in the Rare Donor File: 25 Fy(a‐b‐), 1 Lu(a‐b‐), 1 Lu(b‐), 1 Fy(a‐b‐) and S‐, 1 CCddee (r'r’). Overall, these 29 donors provided 105 RBC units (range 1–26). To date, all donors are still active and 14 are reserved for dedicated donations. The internal price of RBC antigen testing per donor is approximately 100. ‐ CHF, resulting in a total financial effort of around 35,200.‐CHF in the time since the project was started.

Summary/Conclusions: In our experience, a “passive” recruitment of non‐Caucasian blood donors based on ethnicity has an overall low efficiency from a logistic and financial point of view. Moreover, African and Asian blood donors may require investigations for hemoglobin variants and serology for malaria in addition to routine testing. Nevertheless, a targeted determination of extended RBC antigen phenotype does allow the identification of persons with rare phenotypes. Currently, measures for the active recruitment of potentially rare blood donors are being implemented at our center. After a pilot phase, a project for a nationwide recruitment strategy will be elaborated. A further goal is to build a national registry of patients with hemoglobinopathies requiring transfusions.

TRANSFUSION AND TREATMENT OF SEVERE ANAEMIA IN AFRICAN CHILDREN TRIAL (TRACT)

K Maitland^1,2

¹KEMRI Wellcome Trust Research Programme, Kilifi, Kenya ²Imperial College, London, United Kingdom

Background: In sub‐Saharan Africa, where infectious diseases and nutritional deficiencies are common, severe anaemia is a common cause of paediatric hospital admission, yet treatment options are very limited. To avert overuse of blood products the World Health Organization advocate a conservative transfusion policy and recommend iron, folate and anti‐helminthics at discharge. Outcomes are unsatisfactory with high rates of in‐hospital mortality (9–10%), 6‐month case fatality and relapse (6%) warrant a definitive trial to establish best transfusion and treatment strategies to prevent both early and delayed mortality and relapse.

Aims: We conducted a multicentre randomised controlled trial (Transfusion and Treatment of African Children trial; TRACT) in 3954 children aged 2months to 12years admitted to hospital with severe anaemia (haemoglobin<6g/dl). Children were enrolled at 4 centres in Uganda and Malawi and were followed for 6months. The trial simultaneously evaluated (in a factorial trial with a 3×2×2 design) three ways to reduce short and longer‐term mortality and morbidity following admission to hospital with severe anaemia in sub‐Saharan Africa. Trial Registration: Current Controlled Trials ISRCTN84086586.

Methods: The trial compared (i) R1: liberal transfusion (30ml/kg whole blood) versus conservative transfusion (20ml/kg) versus no transfusion (control). The control is only for children with uncomplicated severe anaemia (haemoglobin 4–6g/dl); (ii) R2: post‐discharge multi‐vitamin multi‐mineral (MVMM) supplementation (including folate and iron) versus routine care (folate and iron) for 3months; (iii) R3: post‐discharge cotrimoxazole prophylaxis versus no prophylaxis for 3months. All randomisations were open‐label. Enrollment to the trial started September 2014 and last patient was follow up in December 2017. The primary outcome is cumulative mortality to 4weeks for the transfusion strategy comparisons, and to 6months for the nutritional support/antibiotic prophylaxis comparison. Secondary outcomes include mortality, morbidity (haematological correction, nutritional and anti‐infective), safety and cost‐effectiveness.

Results: At the meeting results from the two transfusion randomisations will be presented. The two manuscripts relating to these randomisations are currently under review and we hope will be published ahead of the conference. Discussion If confirmed by the trial, a cheap and widely available ‘bundle’ of effective interventions, directed at immediate and downstream consequences of severe anaemia, would lead to substantial reductions in mortality in a substantial number of children in African hospitalised with severe anaemia every year and lead to revisions in World Health Organization guidelines.

GLOBAL PERSPECTIVE OF ETHICS IN HEALTH SUPPLY – CRITICAL APPRAISAL

S Hurst Majno

No abstract available.

Transfusion in Limited Infrastructure Locations

A Greinacher¹, T Nwagha², A Ugwu², D Gwarzo³, A Gwarzo³

¹University Medicine, Institute for Immunology and Transfusion Medicine, Greifswald, Germany ²Department of Haematology & Immunology UNTH Ituku, Ozalla ³Department of Hematology, Kano teaching hospital, Kano, Nigeria

Blood transfusion is an essential treatment. Transfusion safety consists of several components. Although all are important, ion richer countries the order of priority is typically: 1.) Avoidance of transfusion transmittable infections; 2.) Quality of the blood product with a strong focus on component therapy; 3.) Prevention of severe transfusion reactions; 4.) Avoidance of clerical errors; 3.) Sufficient availability of blood.

The keynote of this lecture will be that the order of priorities on transfusion safety should probably be different in resource limited environments. 1.) Sufficient availability of blood and proper utilisation; 2.) Avoidance of transfusion transmittable infections; 3.) Avoidance of clerical errors; 4.) Prevention of severe transfusion reactions; 5.) Quality of the blood product.

Most important, in regions with limited resources patients suffer from under‐transfusion because not enough blood is available. All efforts should be made to reduce wastage of the available blood either by inappropriate storage, handling, or non‐indicated transfusions. In addition, prevalence and number of pathogens transmittable by transfusion is much higher than in the richer parts of the world. This is aggravated by the fact that rejection of blood donors based on their history is problematic when the blood bank is empty.

The aim to develop centralized national blood services with few sites manufacturing cost effective (low personnel costs) high‐quality blood products, which are distributed to regional hospitals is not matching the reality of infrastructure, governmental support, and functionality. In healthcare systems with limited resources usually available personnel (hands) is not a problem, while reagents, equipment and even electricity are precious resources.

The lecture will propose to focus on staff training and education, establishing local hospital‐based transfusion services, which provide the blood products for the region, based on donor recruitment campaigns adjusted to the available technology, local culture, including replacement donor programs, with retention of safe donors as highest priority. Fractionation of blood into components had been standard in transfusion medicine, but recently whole blood has experienced revival for patients with acute blood loss. Given main transfusion indications such as postpartum hemorrhage, or severe trauma, in regions with limited infrastructure, whole blood might be the more appropriate product. Most patients requiring transfusion in these regions are younger and volume overload by whole blood is not a major issue. In addition, frequent electricity failures do not allow prolonged storage of plasma at ‐20°C (this is therefore mostly wasted), although this issue can be overcome by solar powered freezers. Ideally whole blood should be pathogen inactivated for which two methods are currently available.

To reduce frequencies of acute hemolytic transfusion reactions again education and training to minimize clerical errors in the transfusion process are most important. Extended testing for other Rhesus antigens and K beside ABO and Rh‐D in transfusion dependent hemoglobinopathy patients may help to reduce delayed hemolytic transfusion reactions.

Currently a leukodepleted pathogen‐inactivated whole blood product might mostly serve the needs for blood transfusion in regions with limited infrastructure . The developed world should invest research efforts to develop such a product available at affordable costs.

HLA‐ AND BLOOD GROUP‐TYPING WITH NGS – ONE‐STOP SHOP

G Schöfl

No abstract available.

A Customisable, Comprehensive Sequencing Panel for Red Cell, Platelet and Neutrophil Genotyping

E Roulis¹, E Schoeman¹, M Hobbs², T Powley³, R Flower¹, C Hyland¹

¹Clinical Services and Research, Australian Red Cross Blood Service, Kelvin Grove ²Garvan Institute of Medical Research, Sydney ³Red Cell Reference Laboratory, Australian Red Cross Blood Service, Kelvin Grove, Australia

Background: In modern transfusion medicine, serological investigations for blood cell antigens are complemented by genotyping arrays and PCR assays. Whilst these platforms are informative in the majority of reference investigations, they are limited in their ability to define nucleotide variants associated with rare antigens and unable to detect novel variants potentially affecting antigenicity. Next‐generation sequencing is increasingly being employed in reference settings, providing information that cannot be obtained through these methods.

Whole genome and whole exome sequencing have been successfully employed in many investigations of novel and rare antigens, however concerns remain regarding the collection of genomic data unrelated to reference investigations, and reporting clinically significant incidental findings. These concerns can be addressed through the use of targeted sequencing panels. We report on the design, testing and efficacy of a panel providing a comprehensive genotyping profile for red cell, platelet and neutrophil antigens in a single test.

Aims: Design a customised targeted exome sequencing panel for red cell, platelet and neutrophil antigen genes, and benchmark the efficacy against a commercial medical sequencing panel (Illumina TruSight One ‐ TSO).

‐ Test the panel and in‐house genotype prediction script on sequence outputs from 51 samples with known red cell, platelet and neutrophil genotypes and phenotypes and determine whether predictions are concordant.

Methods: The panel was designed with 2654 probes covering exons of 64 genes associated with red cell, platelet and neutrophil antigens.

Using Illumina Nextera rapid capture technology, 51 samples were tested over five sequencing runs, on standard and micro chemistries, to determine optimal sample plexity per standard run, and the efficacy of smaller flow cells for lower throughput applications.

An in‐house python script was used to predict star‐allele genotypes based on variants listed in ISBT and EMBL databases. These predictions were compared to results from serology, SNP array and previous TSO data.

Results: Coverage consistently averaged>250×, with 94% of target at a quality of Q30. Optimal sample plexity for a standard run was determined to be 14 samples, allowing for sufficient coverage of all clinically significant variants. For red cell samples with previous typing data (excluding RH structural variants), the script correctly predicted 99.5% of SNP based red cell genotypes. Script predictions were 100% concordant for platelet genotypes, and four of five neutrophil antigen genotypes. HNA2 genotypes defined by CD177 could not be reliably determined.

The increased target coverage of the panel allowed for detection of a clinically significant heterozygous variant in Scianna system, previously undetected by the TSO panel due to extremely low coverage. Additionally, a variant defining a potentially novel null allele was detected in the P1PK system.

Summary/Conclusions: The panel demonstrates considerably higher coverage, quality and throughput compared to the TSO and allows for detection of variants previously overlooked due to low sequencing coverage. Up to 14 samples can be reliably sequenced in a single run. Our script correctly predicts over 99% of SNP based alleles; however, RH structural variants require further manual analysis.

DONOR CHARACTERISATION: A NOVEL PLATFORM FOR COMPREHENSIVE GENOTYPING, RESULTS FROM A LARGE‐SCALE STUDY

N Gleadall^1,2

¹Haematology, University of Cambridge ²NHS Blood and Transplant, Cambridge, United Kingdom on behalf of the Blood transfusion Genomics Consortium

Background: To ensure the safety of a transfusion it is critical to identify the blood type of both donor and recipient. Serological methods for typing ABO, RH and KEL use monoclonal antibodies, however, typing reagents for rare blood groups are expensive, unavailable or unreliable. DNA‐based identification of human blood groups has been used to overcome these limitations and its application has reduced rates of alloimmunisation in chronically transfused patients. However, to date, the cost per sample has prevented the universal application of DNA‐based donor typing

Aims: To achieve universal adoption of DNA based donor typing, the Blood transfusion Genomics Consortium (BGC) set out to develop an affordable DNA based platform, capable of typing all red cell antigens, HLA class I and II and Human Platelet Antigens.

Methods: The UK Biobank Axiom array, previously used to type 600,000 UK citizens, was redesigned for donor typing using three approaches: i) Mining transfusion medicine knowledge, e.g. ISBT allele tables; ii) Inclusion of loci associated with donor health; iii) Extraction of all coding variants in relevant genes with a frequency>1:20,000 identified in large‐scale sequencing data. Samples from NHSBT and Sanquin blood donors (n=7,871) were used for performance assessment. Red cell and platelet antigens for each donor were inferred from genotypes using the bloodTyper algorithm and concordance with clinical serological typing results assessed.

Results: Concordance between genotypic and serological typing results was 99.9% for 95,022 comparisons; 29 of the 89 discrepancies were serologically negative and genotypically positive for a given antigen (K/k, Fy[a/b], Lu[a/b]). In all cases DNA variants known to modify or weaken antigen expression were detected, displaying the power of genotyping to detect variant ‘weak‐antigen’ expression. Across 48 antigens for which serology was available, genotyping provided a 3.6‐fold increase in the number of typing results available per donor (47.9 vs 13.2). Furthermore, genotyping provided data on an additional 224 clinically relevant antigens, allowing identification of antigen‐negative donors and blood group identification for which antibodies are not commercially available. The power of a genotyped panel of donors to support patient management was demonstrated by a retrospective analysis of clinical cases referred to NHSBT. From 3,146 patient referrals with>3 alloantibodies between 2014 and 2018, 772 unique alloantibody profiles were identified. We found that there was a 2.6‐fold greater likelihood of finding O negative compatible donors for these patients when using genotyping data from the 4,721 NHSBT donors. Importantly, the number of alloantibody combinations for which no compatible antigen‐negative donor could be identified fell from 293 to 246, representing an additional 164 patients that could be provided with directly compatible blood using the same donors.

Summary/Conclusions: Through the BGC efforts an affordable fully automated genotyping platform, including the processes for quality assurance and data analysis, has been developed. Furthermore, we have demonstrated the real‐world benefits more extensive donor characterisation can provide when selecting blood for patients with multiple antibodies. The results of this international collaboration provide opportunities to introduce fully‐automated genotype‐based donor typing in a safe and cost‐efficient manner in blood supply organisations.

FULL BLOOD GROUP GENOTYPE PROFILE BY NATIONAL BIOBANK WHOLE‐GENOME SEQUENCING ANALYSIS

P Wu¹, Y Chang², Y Lin³, P Chen³, M Chen², S Pai¹

¹Technical Division ²Department of Testing Lab, Taipei Blood Center ³Graduate Institute of Medical Genomics and Proteomics, National Taiwan University, Taipei, Taiwan ‐ China

Background: In transfusion practice, the accuracy of antigen typing and matching is essential, especially for transfusion‐dependent patients. To provide patients with antigen matching blood can greatly reduce alloimmunization and improve transfusion outcome. However, antigen testing is limited to antigens mostly within the first 10 systems due to the lack of antiserum. Also, antigen typing may be difficult due to altered or weak expressions. Hence, many blood banks and centers incorporate blood group genotyping to complement serotyping.

There are currently 36 blood group systems comprising over 300 antigens, with published phenotypic changes associated with genetic variations. Therefore, antigen expression can be predicted by analyzing blood group gene sequences. Taiwan Biobank performs whole‐genome sequencing (WGS) from individuals nation‐wide. These data are suitable for allele frequency analysis to demonstrate gene expression and genetic profile in our population, also to estimate the significance of each antigen in transfusion practice.

Aims: We aim to provide and verify population‐based blood group antigen profile using WGS and DNA samples from Taiwan Biobank.

Methods: A near 1500 WGS and demographic data were analyzed. Annotations of blood group antigen were performed according to variants from ISBT allele tables, including 2 transcription factors; variants for the Lewis system were obtained from previous studies. Annotations of blood group variants were verified by 4 DNA samples with targeted sequencing on Illumina MiSeq, and specific variants were verified by 40 DNA samples with the commercial genotyping kit or Sanger sequencing. Allele frequencies from WGS analysis were compared with population serology data using two‐proportion z test.

Results: Population‐wide blood group antigens were analyzed, revealed in‐depth antigen expression profiles in all systems (except CH/RG). The antigen frequencies from WGS were similar compared with published serology data, except for the antigens and possible explanations listed as follow, 1) M, N: insufficient sequencing reads, 2) C, c: identical RHCE exon 2 with RHD exon 2 for C allele, 3) Mur: insufficient read length/depth for GYPA/GYPB hybridization calling and individuals from high prevalence of Mur antigen in aboriginal tribes were not enrolled. Blood group antigen predictions and variants from WGS were accord to DNA verification. Furthermore, 13 systems shown no genetic variations, predicting uniform antigen expression in our population, and we can manage transfusion with minimum concerns for antigen mismatch in these systems. Moreover, we found weak and null alleles in our population for blood group systems that we had previously no knowledge of, such as LAN, JR, and VEL. These variants were helped to identify a patient with anti‐Jr^a carrying homozygous Jr^a null alleles.

Summary/Conclusions: Taiwan Biobank WGS is suitable for full blood group antigen profile determination with few adjustments required for specific antigens. The population antigen allele frequency provides valuable insight to antigen significance in transfusion practice, matching strategy for our patients, and estimation of the likelihood to obtain for specific antigen negative blood from mass population. Also, the genetic variants revealed in this study can help us to locate rare donors, and to integrate variations into routine donor blood group screening to provide suitable blood at a low cost efficiently.

EXTENDED BLOOD GROUP GENOTYPING OF IMMORTALISED ERYTHROID CELL LINE BEL‐A2 USING NEXT GENERATION WHOLE EXOME SEQUENCING

R Javed^1,2, L Tilley³, J Frayne¹, V Crew³

¹University of Bristol, Bristol, United Kingdom ²Clinical Haematology and BMT, TATA Medical Center, Kolkata, India ³International Blood Group Reference Laboratory, NHS Blood and Transplant, Bristol, United Kingdom

Background: Providing adequate amounts of safe, appropriately matched blood to meet the demands of an expanding and aging patient population presents a challenging global problem. For these reasons, sustainable in vitro sources of red cells may offer a desirable alternative to reliance on donor blood. The firststable immortalized early adult erythroblast cell line, BEL‐A2, has been shown to differentiate efficiently into mature, functional reticulocytes (Trakarnsanga et al., Nat Commun. 8:14750, 2017) and consequently could provide a readily available tool for diagnostic use and proof of principle for future therapeutic use.

Aims: At IBGRL, next generation whole exome sequencing (WES) has been used previously to accurately predict blood group phenotype in a number of blood group systems, including ABO, Rh and MNS (Tilley & Thornton, Transfusion Medicine 27 (Suppl.2):42, 2017). Here we have used it to analyse and document BEL‐A2 blood group‐related genotypes and predict blood group phenotypes. Additional genes involved in cell‐growth and enucleation were also analysed in order to further elucidate the characteristics of the BEL‐A2 cell‐line.

Methods: BEL‐A2 cells (day 196) were cultured in expansion medium and genomic DNA (gDNA) was isolated from 1×10⁶ cells on day 5. For WES, gDNA libraries were prepared using Nextera^® Rapid Capture Exome enrichment and sequenced on Illumina^® MiSeq. Sequence alignments for 41 genes encoding all 36 known blood group systems and 12 further genes encoding transcription factors and cell enucleation‐associated proteins were visualised using Integrative Genomics Viewer, whilst Illumina^® Variant Studio was used to identify observed mutations. Mutations in coding regions were used to determine BEL‐A2 genotype and predicted phenotype.

Results: Good coverage of most of the selected genes was achieved. Alignment of homologous blood group genes including RHD/RHCE, GYPA/GYPB and C4A/C4B was problematic and additional analysis of coverage of these genes was required for accurate interpretation. Despite a number of polymorphisms observed across the tested genes, BEL‐A2 did not express any novel or rare blood group antigens. Genotyping results predicted a common antigenic profile, in agreement with previous serological and genotyping results where available. Although a number of missense single nucleotide variations were detected in analysed genes, including CR1, CDAN1 and TMX4, these were common polymorphic variants and unlikely to be of any functional significance.

Summary/Conclusions: WES was used to determine BEL‐A2 genotype in relation to blood group genes and selected genes encoding transcription factors and proteins associated with cell enucleation. WES allowed accurate prediction of blood group phenotypes, showing full concordance with available serological data (Trakarnsanga et al, 2017). A small number of mutations were identified which are of unknown significance and require further work to determine any potential phenotypic effects. This complete record of the BEL‐A2 blood group‐related exome will enable reliable gene editing strategies for future diagnostic and therapeutic purposes. Additionally, knowledge of the full cell line exome will allow analysis of any emerging genes of interest and provide better insight into the mechanisms of erythroid differentiation and enucleation.

ITP TREATMENT IN PAEDIATRIC PATIENTS

S Holzhauer

No abstract available.

INTERNATIONAL VARIATION IN PLATELET TRANSFUSIONS PRACTICES AMONG CRITICALLY ILL CHILDREN

M Nellis¹, R Goel², O Karam³, S Stanworth⁴

¹Pediatrics, Weill Cornell Medicine, New York ²Simmons Cancer Institute at SIU School of Medicine, Springfield ³Pediatrics, Virginia Commonwealth University, Richmond, United States ⁴John Radcliffe Hospital, Oxford, United Kingdom

Background: Emerging evidence, especially in neonates, has shown potential harm associated with liberal platelet transfusion strategies. Very little evidence exists regarding optimal platelet transfusion thresholds in critically ill children. Randomized controlled trials may be difficult due to lack of equipoise from providers. If regional variation in practice exists, comparative effectiveness studies may be an alternative approach.

Aims: To describe regional variation in platelet transfusion practices in critically ill children.

Methods: Secondary analysis of a prospective, observational study. Subjects were grouped according to region (North America, Europe, Middle East, Asia and Oceania) and nation. Transfusions were analyzed as prophylactic (given to prevent bleeding) or therapeutic (given to treat bleeding). The primary outcome was the total platelet count (TPC) prior to transfusion. Sub‐groups analyses were performed in children with an underlying oncologic diagnosis and those supported by extracorporeal life support (ECLS). The dosing and processing of the platelet transfusions were analyzed as secondary outcomes.

Results: Five hundred and forty‐nine children from 16 countries were enrolled (67% in North America, 17% in Europe, 7% in Oceania, 5% in Asia, and 4% in the Middle East). Overall, the median (IQR) TPC prior to prophylactic transfusions (n=360) differed significantly on a regional basis (P=0.04) and ranged from 12 (8–41) x10⁹ cells/L in the Middle East to 45 (20–66) x10⁹ cells/L in Asia. The median TPC prior to prophylactic transfusions did not significantly differ between countries (P=0.08), nor did the TPC prior to therapeutic transfusions (n=189) differ on either a regional (P=0.16) or national (P=0.57) basis. For children supported by ECLS (n=90), there were no regional (P=0.06) or national (P=0.40) differences for prophylactic transfusions. However, significant differences in the TPC prior to therapeutic transfusions were observed on both a regional (P=0.02) and national (0.04) basis with the Middle East, in particular Israel, transfusing at the lowest median (IQR) TPC [28 (16–81) x10⁹ cells/L]. For children with an underlying oncologic diagnosis (n=233), no differences were seen in the TPC for prophylactic transfusions (n=175) on a regional (P=0.19) or national (P=0.20) basis. Nor were differences seen in the TPC prior to therapeutic transfusions on a regional (0.86) or national (P=0.33) basis. There was significant variability in the dosing of platelet transfusions on both a regional (P<0.001) and national basis (P<0.001). The median (IQR) dose based on volume ranged from 8.4 (5.0–11.2) ml/kg in North America to 12.6 (9.8–16.0) ml/kg in Europe. The vast majority of transfusions were leukoreduced and irradiated but significant variation exists in storage duration on both a regional (P<0.001) and national (P<0.001) basis.

Summary/Conclusions: Regional and national variation exists in platelet transfusion practices among critically ill children, especially in those given therapeutic transfusions while supported by ECLS. Considering this variation, comparative effectiveness studies may be an appropriate approach to gain evidence to optimize platelet transfusion thresholds.

PROPHYLACTIC PLATELET TRANSFUSION IN PEDIATRIC PATIENTS WITH CANCERS: BAMBINO GESU’ CHILDREN'S HOSPITAL EXPERIENCE

P Berti, F Lamura, L Costantino, C Damico, A Paoloni, E D'Agostino, L Livesi, I Ferruzzi, M Conte, A Cappelli, M Montanari

Immunohematology and Transfusion Medicine Unit, IRCCS Bambino Gesu’ Children's Hospital, Rome, Italy

Background: The optimal threshold for prophylactic platelet (Plts) transfusion in pediatric patients with cancer is still controversial and current clinical practice comes from studies on adults and on inpatient setting. The international guidelines (ICMTG, 2015) recommend, for all age patients, a prophylactic platelet transfusion when Plts count is≤10×10¹¹/L and a platelet dose of 1.1×10¹¹ per square meter (sm) of body‐surface area (BSA) in inpatient and 2.2×10¹¹/sm in outpatient setting.

Aims: In January 2018 we started in our Children's Hospital a prospective protocol in order to evaluate the impact on bleeding risk of current clinical practice of prophylactic platelet transfusion in inpatients and outpatients onco‐haematological patients.

Methods: BSA was calculated from age‐standardized weight. Inpatients received a dose per transfusion of 1.1×10¹¹/sm and Outpatients a dose per transfusion of 2.2×10¹¹/sm.

Platelets were transfused when the count was≤10×10¹¹/L or in presence of bleeding signs; pediatric aliquots were obtained from Buffy coat derivedpooledplatelet concentratesor apheresisplatelet concentrates, according disponibility.

Results: From January 2018 to December 2018 a total of 9839 platelet pediatric aliquots were transfused: 5534 (56.2%) were obtained from apheresisplatelet concentrates and 4305 (43.8%) from Buffy‐coat‐derivedpooledplatelet concentrates. The majority of platelets pediatric aliquots (7505–76.3%) were transfused to onco‐hematological patients undergoing hematopoietic stem cells transplant (HSCT) or conventional chemotherapy. Among them, 6365 aliquots were transfused in inpatient setting: 1675 (17%) in the hematology unit, 2772 (28.2%) in the oncology unit and 1918 (19.5%) in HSCT unit.

A total of 1140 (11.5%) aliquots were transfused in outpatient setting: 840 (8.5%) to patients affected by hematological malignancies and 300 (3%) to patients with solid tumors.

Five major bleeding events (WHO grade≥3) were observed during the study period and all of them occurred in hospitalized patients. Two patients with solid neoplasm developed a WHO grade 3 bleeding event. Two patients with hematologic malignancies and a patient with neuroblastoma (n=3, 1.2%) developed intracranial bleeding (WHO grade 4). The platelet count at the time of the event was 22×10⁹/L, 25×10⁹/L and 8×10⁹/L, respectively.

Summary/Conclusions: Our results showed the efficacy, in onco‐hematological pediatric patients, of a prophylactic platelet transfusion protocol based on international guidelines: a very low incidence of WHO grade 4 bleeding has been observed in inpatients setting only (1.2% vs 1% of PLADO Trial, SJ Slichter, NEJM, 2010), while in outpatients setting the double platelet dose prevents the major bleeding event (WHO grade≥3) occurrence.

THE RESULTS OF TRANSFUSIONS OF PATHOGEN‐REDUCED RED BLOOD CELL SUSPENSIONS IN CHILDREN WITH ONCOLOGICAL AND HEMATOLOGICAL DISEASES

I Kumukova, P Trakhtman, N Starostin, L Kadaeva, O Chaykina

Transfusiology, National Medical Research Center for Pediatric Hematology, Oncology and Immunology, Moscow, Russia

Background: The problem of blood‐borne infections remains relevant in transfusion medicine. Pathogen reduction technologies (PRT) provide a preventive approach to a wide range of transfusion‐transmitted infectious diseases. To date, PRT widely used for platelet concentrates and blood plasma, however, the use of these technologies for the treatment of red blood cell‐containing blood products undergo research.

Aims: The aim of our study was to evaluate the safety and efficacy of transfusions of pathogen‐reduced (test group) red blood cell suspensions (RBCS) and compare these data with gamma‐irradiated RBCS (control group).

Methods: The technology based on the combined action of riboflavin and ultraviolet (Mirasol PRT, Terumo BCT, Belgium) was used to reduce pathogens in whole blood. Subsequently, the RBCS of the test group were derived from pathogen‐reduced whole blood.

The control RBCS were irradiated at the GammaCell 3000 Elite (Best Theratronics, Canada)at a dose of 25 Gray.

All RBCS were used for transfusion for 14days from the harvest day.

70 pediatric patients with various oncological and hematological diseases wererandomized to 2 groups of 35 members in each group. The test group of patients received transfusions of a pathogen‐reduced RBCS; the control group received transfusions of a gamma‐irradiated RBCS.

The next day after transfusion were assessedhemoglobin and hematocrit increment, the level of potassium and haptoglobin in the patients’ serum, the frequency and severity of transfusion reactions. 3–5days after the transfusion, the direct antiglobulin test (DAT) was performed and after 14–21days the indirect antiglobulin test (IAT) was performed. The interval to the next need for transfusion was also evaluated.

Results: The increase in hemoglobin and hematocrit (P=0.2), as well as the concentration of potassium (P=0.44) and haptoglobin (P=0.25) in the patients’ serum after the transfusion did not differ between groups. None of the patients in both groups had hyperkalemia after transfusion. In each group, two patients hadfebrile non‐hemolytic transfusion reactions of comparable severity (P=1). All DAT and IAT tests were negative in both groups. The interval between transfusions were not significantly different between groups (P=0.39). Only in the test group was found the correlation between the increase in the hemoglobin and hematocrit values with the volume of transfusion, with the dose and the adjusted dose of hemoglobin obtained for the transfusion on body weight. And in this group was found inverse correlation between the hemoglobin and hematocritincrement with the level of hemolysis in the RBCS.

Summary/Conclusions: We found that the clinical efficacy and safety of RBCS of the compared groups did not differ. There was no evidence of immune elimination and allo‐sensitization caused by pathogen‐reduced RBCS. According to our data, the spectrum of efficiency and safety indicators of pathogen‐reduced RBCS is no worse than that of gamma‐irradiated RBCS, provided that RBCS is used for 14days of storage.The founded correlation suggests that the efficiency of pathogen‐reduced RBCS transfusions is more dependent on the characteristics of the RBCS.

5‐YEAR TEMPORAL TRENDS IN PERIOPERATIVE RED CELL TRANSFUSIONS IN CHILDREN: NATIONALLY REPRESENTATIVE DATA FROM A LARGE PROSPECTIVE NORTH AMERICAN REGISTRY

A Gupta⁶, C Josephson³, M Nellis⁴, O Karam⁵, LV Vasovic⁷, A Tobian⁸, R Goel^1,2

¹SIU School of Medicine, Springfield ²Johns Hopkins, Baltimore ³Pathology, Emory University, Atlanta ⁴Pediatrics, Weill Cornell, New York ⁵Pediatrics, Virginal Commonwealth University, Richmond ⁶Vanderbilt University, Nashville ⁷Weill Cornell Medical College, New York ⁸Pathology, Johns Hopkins, Baltimore, United States

Background: Patient blood management (PBM) programs are expanding at an international level. A recent nationally representative study from United States observed pediatric age group as the only age group showing lack of objective evidence of PBM initiatives (Goel et al, JAMA 2018).

Aims: This study aims to identify trends in peri‐operative blood utilization in children undergoing elective and non‐elective surgeries over 5years duration from 2012 to 2016.

Methods: Using 5years data (2012–2016) from pediatric database of the American College of Surgeons National Surgical Quality Improvement Program (PEDS ACS‐NSQIP), temporal trends of peri‐operative RBC transfusions in children (<18years) undergoing elective/urgent/emergent surgeries were assessed and subgroup analyses by case type per performed. Trend analysis was performed to assess for significance.

Results: A total of 369,176 children (43.2% males and 56.8% females) were analyzed. Of these, 77,521 were infants (<1year) and 15,858 neonates (<28days)]. 73.2% patients underwent elective, 10.0% urgent and 16.8% emergent surgical procedures respectively.

Overall, perioperative transfusion of RBCs was documented in 23,835 patients (6.5%). [7,926 (10.2%) infants; 2,487 (15.7%) neonates]. Of these, about 0.7% (n=2,425) children received pre‐operative transfusions (within 48h of surgery). About 5.4 percent (n=19,968) of children received RBC transfusions intra/post operatively (start of surgery until 72hrs post‐op).

Perioperative transfusions decreased steadily per year from 6.4% in 2012 to 5.5% (14% cumulative decline) in 2016 for children of all ages (OR 0.966; 95% CI 0.957–0.976; p trend<0.001). The cumulative change in elective procedures was 9.2% versus 27.2% decrease in urgent/emergent procedures (p trend<0.001).

Summary/Conclusions: In this large prospective registry study of>350,000 children undergoing elective/non‐elective surgeries, a statistically significant decrease in utilization of peri‐operative RBC transfusions was seen across 5years from 2012 through 2016 with more significant decrease in urgent/emergent procedures than elective procedures

While these findings need evaluation for non‐surgical indications of transfusion, these results may provide first evidence of peri‐operative pediatric patient blood management strategies being implemented to optimize transfusions in pediatric population.

TTI AND PATIENTS WITH IMMUNE DEFICIENCY: SELECTED BLOOD PRODUCTS OR CLOSE MONITORING

H. H. Hirsch

University of Basel, Basel, Switzerland

Transfusion‐transmitted infections (TTI) are a long‐standing and well recognized concern in medicine, which is tackled on the highest level to guarantee the safety of the transfusion procedure for all stake holders. These include the recipient patients, the donating volunteers, the health care workers involved, and their respective contact persons. Accordingly, current national and international guidelines including expert societies and the WHO provide medical, technical, and legal frameworks, which are the basis for the standard operating procedures. Nevertheless, there are important challenges, which render TTI a “moving target”, and reflect the dynamics in three main areas. First, a change in the type and number of recipient patients with past or ongoing immunomodulatory / immunodeficiency component (examples being HIV/AIDS, SOT, allogenic HCT, monoclonal antibody therapies, small molecule inhibitors). Second, changing exposure to known agents in donors due to global travel, migration and displacement, as well as environmental/climate change. Third, discovery and diagnostics of old and new agents with their known or presumed impact as TTI. These aspects will require careful review of data and studies, and judicious discussion of the potential action such as selection versus close monitoring to keep TTI rates as low as possible, to deliver maximal safety of patients and stakeholders.

CHARACTERIZATION OF HEPATITIS B VIRUS STRAINS INFECTING BLOOD DONORS WITH HIGH HBSAG LEVELS AND INCONSISTENTLY DETECTABLE HBV DNA: IMPLICATIONS FOR BLOOD SAFETY AND SCREENING POLICY

D Candotti¹, M Vermeulen², A Kopacz³, M Miletic⁴, A Saville², P Grabarczyk³, C Niederhauser⁵, L Boizeau¹, S Laperche¹

¹DATS CNR RIT, National Institute of Blood Transfusion, Paris, France ²The South African National Blood Service, Johannesburg, South Africa ³Virology, Institute of Haematology & Transfusion Medicine, Warsaw, Poland ⁴Research & Development, Croatian Institute of Transfusion Medicine, Zagreb, Croatia ⁵Research & Development, Inter‐regional Blood Transfusion SRC, Berne, Switzerland

Background: The implementation of nucleic acid testing (NAT) and the development of sensitive and specific serologic assays to detect HBsAg and anti‐HBc antibodies significantly reduced the risk of HBV transfusion‐transmission. The apparent redundant testing for two direct viral markers prompted debates on maintaining HBsAg screening, particularly in low endemic countries where blood donations are screened for anti‐HBc. However, frequencies of 2–20% of HBsAg‐confirmed positive/NAT negative donations have been reported depending on the sensitivity limit of the molecular assays used. The nature of this discrepancy between HBsAg and DNA remains largely unknown and it is essential to evaluate any potential negative impact on blood safety before considering removing HBsAg testing.

Aims: The prevalence in blood donors and the molecular mechanisms responsible for a persistent undetectable or barely detectable level of viral replication in the presence of a sustained HBsAg production were investigated in a collaborative study including five laboratories/blood centers in Europe and South Africa. Discrepancy between viral DNA and HBsAg levels suggested the presence of mutations that may negatively affect HBV replication and/or infectious viral particle production.

Methods: Donor samples from France, South Africa, Poland, and Croatia were selected for having HBsAg levels≥100IU/ml and being ID‐NAT (Procleix‐Ultrio Plus™ [95% LOD: 3IU/ml]) non‐reactive/non‐repeatable reactive (NR/NRR) with undetectable viral load (VL) or<6IU/ml (n=44) or NAT repeat reactive (RR) with VL<6IU/ml (n=32). French samples initially tested NAT NR/NRR with Procleix‐Ultrio (LOD 95%: 11IU/ml) were retested with Ultrio Plus prior inclusion in the study. HBV DNA load was quantified (Cobas TaqMan HBV [LOQ: 6IU/ml]). HBV DNA was purified from 5 to 12ml of plasma after ultracentrifugation. The whole HBV genome, Pre‐S/S, PreCore/Core and BCP regions were amplified and sequenced.

Results: Following viral concentration, HBV DNA presence was confirmed in 79% of all samples with undetectable or VL<6IU/ml. HBV genotypes were A1 (33.3%), A2 (18.4%), A3 (3.3%), B (3.3%), C (1.7%), D (25%), and E (15%). All samples were anti‐HBc positive and 73% of Ultrio‐negative samples tested positive with Ultrio Plus. Unusual 1–2 nt insertions/deletions identified in BCP regulatory elements (TATA boxes, pgInr, epsilon domain) suggest altered viral replication. Amino acid substitutions (n=16) or deletions (n=4) at positions reported involved in nucleocapsid formation, particle envelopment and virion formation were observed in the core protein of 30 samples. The replicative properties of the BCP and core variants are currently evaluated in vitro as a surrogate model for direct infectivity testing. Preliminary results indicate that the variants tested so far have replicative capabilities similar to those of control viruses. Analysis of Pol, S, and HBx proteins is ongoing.

Summary/Conclusions: These data confirmed the presence of extremely low level of circulating DNA‐containing viral particles in ID‐NAT non‐reactive or non‐repeated reactive blood donations with concomitant high HBsAg levels and anti‐HBc reactivity. Despite the presence of mutations in the viral genomes potentially affecting virion production, preliminary data indicate that some of the viruses in plasma retain the ability to replicate in vitro and to constitute a potential infectious risk.

HEPATITIS B VIRUS INFECTION AFTER VACCINATION IN A BLOOD DONOR: A CASE REPORT

A Brassel¹, M Stolz¹, C Tinguely¹, P Gowland¹, M Jutzi¹, C Niederhauser^1,2,3

¹Interregional Blood Transfusion SRC, Switzerland, Bern ²Faculté de Biologie et de Médecine, Université de Lausanne, Lausanne ³Institute of Infectious Diseases, University of Bern, Bern, Switzerland

Background: In Switzerland highly sensitive nucleic acid screening in an individual donation format for hepatitis B virus (HBV ID‐NAT) and hepatitis B surface antigen (HBsAg) detection is mandatorily performed (guidelines of Swiss Transfusion SRC, Switzerland). Since 1998, HBV (HB) vaccination is recommended in Switzerland for children and adolescents until the age of 15 and for adults belonging to known risk groups.

Aims: To highlight that low anti‐HBs titers several years following HBV vaccination still confer protection and enable the host immune system to clear HBV DNA without development of serologic markers of disease.

Methods: A retrospective donor interview was conducted to complete information not covered by the questions included in the standard donor questionnaire. Routine HBV serological donor screening was performed on a Quadriga system (Diasorin, former Siemens) with the Enzygnost HBsAg assay (Diasorin, former Siemens). Further HBV tests were performed on the Abbott Architect i1000 analyser (HBsAg neutralisation, HBeAg, anti‐HBc IgG/IgM, anti‐HBc IgM, anti‐HBe and anti‐HBs). Routine ID‐NAT screening for HIV/HCV/HBV was performed with the Roche cobas MPX test on a Roche cobas 8800 platform. HBV ID‐NAT positive samples were confirmed with a quantitative HBV NAT assay (Abbott).

Results: Testing of the implicated donation and pre‐and post donation samples: During the screening at December 4th 2018 a 57‐year‐old male blood donor tested positive for HBV DNA by ID‐NAT with a low titer of HBV (6IU/ml) and negative for HBsAg. In a follow‐up blood sample at December 11th 2018 the positive HBV NAT result was confirmed (<1IU/ml) and HBsAg was again negative. Further serologic testing for active or past HBV infection (anti‐HBc IgM/IgG, HBeAg, anti‐HBe) were negative in both samples. All previous 26 donations were negative in screening for HBV DNA and HBsAg. A further blood sample, obtained on February 19th 2019, had undetectable levels of HBV‐DNA while serologic HBV markers including anti‐HBc IgM/IgG remained negative. The anti‐HBs antibody level was 13.86IU/ml on May 25th 2018, 500 mIU/ml on December 4th 2018 and>1000IU/ml on December 11th 2018.

Retrospective review of the donor interview and medical records: The donor had received HBV vaccination (Engerix B 20μg/ml) in 2001/2002 according to a standard 3‐dose schedule as verified in his medical records. He indicated a possible exposure to Hepatitis B virus by sexual contact in Thailand after the last negative donation in May 2018, greater than 6months prior to the positive donation. The donor confirmed to have had no other potential exposure to HBV in the relevant period.

Summary/Conclusions: In this previously HBV vaccinated blood donor with an anti‐HBs titer of 13.86 mlU/ml before suspected infection, we detected a transient viremia without clinical evidence of disease, no expression of HBsAg and complete absence of seroconversion for the typical HBV infection markers. An increase of anti‐HBs titer was observed following infection accompanied by the complete clearance of HBV DNA. These findings demonstrate the protective effect 17years after vaccination despite breakthrough infection and anti‐HBs far below the level considered protective and successful elimination of HBV without seroconversion.

ANALYSIS OF SERUM HEPATITIS B CORE‐RELATED ANTIGEN (HBCRAG) IN BLOOD DONORS WITH OCCULT HEPATITIS B VIRUS INFECTION

M Spreafico¹, B Foglieni¹, I Guarnori¹, S Frison¹, A Berzuini², C Bonato³, A Gerosa¹, D Prati²

¹Department of Transfusion Medicine and Hematology, Azienda Socio‐Sanitaria Territoriale (ASST) di Lecco, Lecco ²Department of Transfusion Medicine and Hematology, Fondazione IRCCS Cà Granda Ospedale Maggiore Policlinico, Milan ³Department of Laboratory Medicine, Azienda Socio‐Sanitaria Territoriale (ASST) di Lecco, Lecco, Italy

Background: Hepatitis B core‐related antigen (HBcrAg) is a structural antigen of HBV, consisting in HBCAg, HBeAg and the p22cr precore protein. Quantitative HBcrAg measurement is a sensitive marker of viral replication reflecting the cccDNA content and persistence of disease. HBcrAg positivity was found to be a significant risk factor of HBV reactivation in HBsAg‐, anti‐HBc+, HBV DNA‐ patients (occult HBV infection, OBI) undergoing immunosuppressive therapy.

Aims: No data about HBcrAg status in apparently healthy subject with OBI are available. The aim of this study was to analyse this marker in our cohort of OBI blood donors.

Methods: HBcrAg was measured in 69 blood donors confirmed to be carriers of OBI (HBsAg‐, HBV DNA+). Of them, 59/69 (85.5%) donors were anti‐HBc positive, and 10 (14.5%) negative. 18 donors had both anti‐HBc and anti‐HBe reactivities. A group of 11 young blood donors vaccinated for HBV infection (HBsAg‐, HBV DNA‐, anti‐HBc‐), and 9 patients with chronic HBV infection (HBsAg+, HBV DNA+) were used as negative and positive controls group, respectively. Serum HBcrAg was measured using a chemiluminescent enzyme immunoassay on the Lumipulse G600 automated analyzer (Fujirebio, Tokyo, Japan). The lower limit of detection (LOD) of the quantitative assay is 2 logU/ml and the lower limit of quantification (LOQ) is>3 logU/ml, due to nonlinearity results between 2 and 3 logU/ml. Levels of HBcrAg were tested in the three groups and analysed in comparison to the presence of anti‐HBc and anti‐HBe. Statistical analysis was performed by the IBM Statistics SPSS 19.0.0.

Results: All donors in the negative control group had undetectable HBcrAg levels, whereas all patients in the positive control group have detectable HBcrAg (mean value: 3.0 logU/ml, range 2.0–4.9), confirming that individuals without prior exposure to HBV would not have detectable HBcrAg. HBcrAg was detectable in 34/69 OBI donors (49.3%), with a mean value of 2.21 logU/ml (range 2.0–3.20). HBcrAg could be measured only in 2 OBI donors (3.0 and 3.2 logU/ml), being below the LOQ of the test in the majority of OBI (32/34). Considering the presence of anti‐HBc, HBcrAg was detected in 29/59 (49.1%) anti‐HBc+ and in 5/10 (50%) anti‐HBc‐ OBI, with no significant difference in their mean levels (2.21±0.32 vs 2.14±0.26; P=0.44). Interestingly, the presence of anti‐HBe (18/62) was independently associated with higher HBcrAg levels (2.38±0.42 vs 2.14±0.22; P=0.03).

Summary/Conclusions: Identification of donors with OBI is critical to prevent the risk of HBV transfusion‐transmission. Being HBcrAg associated with the cccDNA content and replication, our results suggest that the presence of HBcrAg, even if not quantifiable, could be useful marker to confirm the occult infection status, even in anti‐HBc negative donors. The association between HBcrAg, anti‐HBc and anti‐HBe could also be a useful marker to identify OBI donors with a higher risk of HBV reactivation.

ANALYSIS OF T CELL IMMUNOLOGICAL CHARACTERISTICS OF OCCULT HEPATITIS B INFECTION

X Rong^1,2, W Zhang², Y Fu^1,2

¹Institute of Transfusion, Guangzhou Blood center ²School of Laboratory Medicine and Biotechnology, Southern Medical University, Guangzhou, China

Background: Occult hepatitis B infection (OBI) is manifested as positive HBV DNA in liver tissue, low or negative serum HBV DNA, and negative serumhepatitis B surface antigen(HBsAg), which may be accompanied by positive serum hepatitis B surface antibody(HBsAb), HBeAb and/or HBcAb. Most OBI have a strong inhibitory effect on HBV replication and gene expression. However, OBI has not been clearly described in terms of virological characteristics, host immune response, and epigenetics. In particular, the molecular and cellular immune functions of OBI carriers are not clear.

Aims: To investigate the molecular and cellular immune function of T lymphocyte, including characteristics of T cell proliferation and IFN‐γ secretion capacity of OBI, chronic hepatitis B infection (CHB) and healthy blood donor control (HC).

Methods: HBsAg was detected by chemiluminescence in study population. HBV DNA was detected by nucleic acid test (NAT). Three groups of study population were established, including 37 cases in OBI group, 53 cases in CHB group and 23 cases in HC group. Human peripheral blood mononuclear cells (PBMCs) from blood donors were stimulated with HBV polypeptides pool in vitro. T cell proliferation assays (CFSE) was used to detecting T cell proliferation, enzyme‐linked immunospot assay (ELISPOT) was used to detecting the frequency of HBV‐specific IFN‐γ secreted T cells. SPSS 20.0 statistical analysis software was used for statistical analysis. The measurement data of normal distribution were tested by two independent samples t test; and the comparison between multiple groups was analyzed by one‐way ANOVA. Mann‐Whitney U test was used for comparison between non‐normal data sets. P<0.05 was considered statistically significant.

Results: 1. Proliferation characteristics of T cells. The proliferation of CD4⁺ T lymphocytes was mainly stimulated by specific HBV polypeptide pool, and the proliferation rates of OBI group and CHB group were significantly higher than those of HC group (3.0%, 3.3% vs. 1.7%), with significant difference (3.0% vs. 1.7%, P=0.016, 3.3% vs 1.7%, P<0.001).

2. The frequency of specific IFN‐γ secreted T cells. The response intensity of the OBI group (25 SFC/10⁶ PBMCs) and CHB group (25 SFC/10⁶ PBMCs) was higher than that of the HC group (5 SFC/10⁶ PBMCs) under the stimulation of HBV polypeptide pool, and the positive rate of T cell response to the stimulation of HBV polypeptide pool was the highest in the OBI group (64.0%).

Summary/Conclusions: Both OBI and CHB had higher rates of HBV‐specific T effector cell proliferation and IFN‐γ secretion than the healthy control group. Compared with the CHB group, OBI group had a higher positive rate of T cell response, which may be one of the causes of host immunity resulting in OBI. Further studies on other immune factors are required.

DEVELOPMENT OF SCENARIOS FOR THE FUTURE DEMAND OF BLOOD PRODUCTS IN THE NETHERLANDS: SCENARIO SESSIONS

P Langi Sasongko¹, M van Kraaij², K van den Hurk¹, M Janssen¹

¹Donor Medicine Research ²Donor Affairs, Sanquin, Amsterdam, Netherlands

Background: Western blood transfusion practices are currently changing due to various drivers such as blood management policies, ongoing technological developments, and new therapeutic options. In the Netherlands, as in many high‐income countries, these have resulted in a diminishing trend of red blood cells. Therefore, it is important for blood bank management to anticipate the future demand of blood products for the sake of medium and long term decision making. To support this decision making, we have employed scenario development, which is used in many other sectors (such as finance and transportation) and can also be applied to blood transfusion. Building upon a prior literature review and semi‐structured interviews of international experts, we gathered experts together for scenario sessions to assess the opportunities and threats for Sanquin's medium‐term (15–20years) strategy using an online platform and face‐to‐face discussions.

Aims: To assess for opportunities, threats, and the organizational implications thereof for the medium‐term future of Sanquin, the Dutch national blood bank.

Methods: Twenty‐one multidisciplinary experts in blood transfusion agreed to participate and were separated into two groups for half‐day interactive sessions. Using an iterative process through an online platform, experts brainstormed opportunities and threats for Sanquin, which were categorized into themes. These themes were ranked according to importance and certainty, and through consensus, experts chose two themes with high impact and high uncertainty. For these chosen themes, specific actions for the blood bank were listed to mitigate and/or enhance the threat or opportunity. Discussions were ample throughout.

Results: With regards to opportunities and threats for Sanquin's medium term strategy, experts brainstormed many ideas and categorized them under 10 themes: political context/ changing legislation, novel products and alternative applications, donors, international markets, commercialization, digitalization, change in perceptions, research, demand, and organizational structure. After ranking for importance and certainty, six themes were chosen: change in perceptions, international markets, political context (opportunities), demand (opportunities), research (vulnerabilities), and donor (vulnerabilities). For each of these themes experts provided specific actions for the organizations to mitigate threats or stimulate opportunities accordingly. These actions included increased transparency and improved communication with the (donor) public, lobbying in political spheres, increased activities in educational institutes and large funding organizations, and creating and collaborating on novel blood products on an international level, to name a few.

Summary/Conclusions: These results show that mapping and assessing a blood bank's future using a multi‐disciplinary group of experts is conducive as an effective means of collection a diverse range of opportunities and threats. This provides an opportunity for blood bank management to become proactive towards these potential opportunities and threats and possibly evolve future strategies for the organization.

POLYGENIC RISK SCORES BASED ON PLASMA FERRITIN DO NOT PREDICT TIREDNESS/LACK OF ENERGY IN FEMALE BLOOD DONORS – RESULTS FROM THE DANISH BLOOD DONOR STUDY

J Dowsett¹, M Didriksen¹, A Rigas¹, L Thørner¹, E Sørensen¹, K Burgdorf¹, C Erikstrup², O Pedersen³, K Banasik⁴, H Ullum¹

¹Clinical Immunology, Copenhagen University Hospital, Copenhagen ²Clinical Immunology, Aarhus University Hospital, Aarhus ³Clinical Immunology, Næstved Hospital, Næstved ⁴NNF Center for Protein Research, University of Copenhagen, Copenhagen, Denmark

Background: A Danish Blood Donor Study (DBDS) analysis presented at ISBT 2018 showed that iron‐deficient female blood donors were more likely to have depressive symptoms than non‐iron deficient female blood donors. Among participants with depressive symptoms, females with low plasma ferritin levels had significantly increased odds for reporting a “feeling of lacking energy and strength” (OR=2.11; 95% CI: 1.03–4.31). As it is known that blood donors are at an increased risk of iron deficiency, it is important to determine whether those genetically predisposed to lower plasma ferritin levels have a higher risk of experiencing the tiredness/lack of energy symptom.

Aims: To investigate whether there is an association between polygenic risk scores (PRSs) based on plasma ferritin levels and the tiredness/lack of energy symptom in blood donors.

Methods: The DBDS is an ongoing nationwide blood donor cohort, of which genome‐wide genotype data are available for 72,000 participants. Genotyping was performed using the Infinium Global Screening Array (Illumina^®) and imputation was achieved based on a Scandinavian reference genome. Ferritin PRSs, based on an Icelandic ferritin GWAS (N=150,000), were calculated for all DBDS participants. 12,903 female donors were available for the analysis. Data on depressive symptoms were obtained using the validated Major Depression Inventory scale (MDI), a self‐report mood questionnaire, which assesses the presence of 10 depressive symptoms. A donor was classified as “tired” if they responded “all the time” or “most of the time” to the question “how often do you feel that you lacked energy and strength?”. Logistic regression analysis was performed, adjusting for age. For generating the quantile plots, the participants were distributed evenly into six quantiles based on their PRS, whereby quantile 1 contained the donors with the lowest PRSs (genetically predisposed to lower ferritin levels) and was set as the reference quantile with OR=1 from the age‐adjusted regression analysis (tiredness˜quantile).

Results: PRSs in females ranged between ‐0.42 and 0.50 (mean 0.01). A total of 12,523 female donors were classified as “not tired” and 380 (2.9%) were classified as “tired”. No significant difference in ferritin PRS was found between “tired” and “not tired” female donors (tired mean PRS: 0.00; not tired mean PRS: 0.01). An age‐adjusted logistic regression model found this to be insignificant (OR: 0.74, 95% CI: 0.33–1.66), P=0.47). To visualise the lack of association, a quantile plot was created, separating the female donors into six equal quantiles based on their PRS. No clear trend was observed; donors with the highest PRSs (in quantile 6) had OR=1.02 (P=0.93) of being tired when compared to those in quantile 1 (OR set as 1).

Summary/Conclusions: No significant association was found between the ferritin PRSs of female blood donors and the tiredness/lack of energy symptom. Further studies are needed to understand the effect of blood donation versus genetic constitution on tiredness among female iron‐deficient blood donors.

HIV‐1 MOLECULAR EPIDEMIOLOGY AND PREVALENCE OF RESISTANCE ASSOCIATED MUTATIONS IN TREATMENT NAIVE BLOOD DONORS

J Zhao^1,2,3, L Chang^1,2, H Ji^1,2,3, L Zhang^1,2,3, X Jiang^1,2,3, F Guo^1,2, L Wang^1,2,3

¹National Center for Clinical Laboratories, Beijing Hospital, National Center of Gerontology, Beijing, P. R. China ²Beijing Engineering Research Center of Laboratory Medicine, Beijing Hospital, Beijing, P. R. China ³Graduate School, Peking Union Medical College, Chinese Academy of Medical Sciences, Beijing, P. R. China, Beijing, China

Background: Antiretroviral therapy (ART) is critical for the control of clinical progression of human immunodeficiency virus (HIV) infections. However, the outcome of ART could be limited by drug resistance‐associated mutations (DRMs), even lead to the transmission of drug‐resistant HIV to treatment naïve patients such as blood donors, which is a huge concern to ART. DRMs surveillance in HIV infected groups is strongly recommended by World Health Organization. Characteristics of genetic diversity and DRMs of HIV among blood donors may provide comprehensive data to monitor viral evolution and optimize ART, play important roles in blood safety.

Aims: Limited data concerning the epidemic of HIV‐1 subtypes and DRMs of blood donors is available in China. This study is to investigate genetic characteristics and DRMs of HIV‐1 infected blood donors.

Methods: From 2016–2018, 177 blood donations collected from 24 blood centers, covering almost the whole of China, were confirmed as HIV‐1 positive by National Centers for Clinical Laboratories using Abbott RealTime HIV‐1 assay or COBAS TaqMan HIV‐1 Test, version 2.0. Then HIV‐1 Gag (973bp, HXB2: 1074‐2044), pol genes (1454bp, HXB2: 2068‐3521) (encoding the whole protease (PR) and a part of reverse transcriptase (RT)) was sequenced after viral RNA extraction and amplification. HIV‐1 subtype based on Gag and PR‐RT regions was determined by comprehensive analyses of Los Alamos HIV BLAST tool, REGA HIV‐1 Subtyping Tool, phylogenetic trees and online jpHMM program. DRMs analysis was performed in the Stanford HIV Drug Resistance Database.

Results: Among 177 donations, Gag and PR‐RT regions of 160 samples were sequenced successfully. The distribution of HIV‐1 genotype was as follows: CRF_07BC=66 (38.8%), CRF_01AE=58 (36.3%), B=9 (5.6%), CRF_08BC=2 (1.3%), CRF55_01B=4 (2.5%), CRF59_01B=1 (0.6%), CRF65_cpx=1 (0.6%), CRF67_01B=2 (1.3%), CRF79_0107=1 (0.6 %), CRF85_BC=1 (0.6 %), URF_0107=6 (3.8%) and URF=9 (5.6%). 26 of 160 HIV‐1 isolates were identified to have DRMs. There were 4 (15.4%, 4/26) Protease Inhibitors (PI) accessory DRMs, 3 PI major DRMs and 20 (76.9%, 20/26) Non‐nucleoside Reverse Transcriptase Inhibitors (NNRTI) DRMs. Most of blood donors with DRMs were CRF01_AE and CRF07_BC (61.5%, 16/26). 3 of 4 PI accessory DRMs were Q58E. The PI major DRMs included M46L, M46I and N88S. N88S could result in HLR to atazanavir(ATV) and NFV, LLR to indinavir(IDV) and saquinavir(SQV).V179D/E is main NNRTI DRM (70.0%, 14/20). A combination of V179D and K103R among two samples acted synergistically to reduce efavirenz (EFV) and nevirapine(NVP) susceptibility. Furthermore, two blood donors with K103N mutation in reverse transcriptase gene had high level‐resistance to EFV and NVP.

Summary/Conclusions: Overall, the most prevalent subtypes among blood donors in the study were CRF07_BC (38.8%), CRF01_AE (36.3%). Besides, other rare CRFs and several URF_0107 and URFs were also found in these HIV‐1 isolates, which suggested theepidemic of HIV has been shifted from high risk populations into general populations, including blood donors in China.DRMs were observed in 16.3% donors in the study, which may result in resistance to PIs and NNRTIs, especially the HIV‐1 variants with N88S mutation in PR gene and K103N mutation in RT gene.In summary, our findings indicate that increasing diversity of HIV‐1 in blood donors and remind us the necessity of timely genotypic drug resistance monitoring and molecular epidemiology surveillance of HIV‐1 among blood donors.

BIOTINYLATION OF PLATELETS FOR TRANSFUSION PURPOSES‐ A NOVEL METHOD TO LABEL PLATELETS IN A CLOSED SYSTEM

S de Bruin^1,2, E van de Weerdt², D Sijbrands³, R Vlaar¹, E Gouwerok¹, B Biemond⁴, A Vlaar², R van Bruggen¹, D de Korte^1,3

¹Blood cell research, Sanquin Research and Landsteiner Laboratory ²Department of Intensive care, Amsterdam University Medical Centers, location AMC ³Product and Process Development, Sanquin ⁴Department of Hematology, Amsterdam University Medical Centers, location AMC, Amsterdam, Netherlands

Background: Labeling of platelets is required to measure the recovery and survival of transfused platelets in vivo. Currently a radioactive method is used to label platelets. However, its’ application is limited, due to safety issues and the inability to isolate transfused platelets out of the circulation. Biotin‐labeling of platelets is an attractive non‐radioactive option, however, no validated protocol to biotinylate platelets is currently available for clinical purposes.

Aims: The aim of this study is to develop a simple, standardized, reproducible method to label platelets with biotin as a non‐radioactive alternative to trace transfused platelets in vivo.

Methods: Six pooled buffy coats derived platelet concentrates (PCs) stored in 100% plasma were biotinylated at day 1 and day 7 of storage. To distinguish the effect of the processing steps from the effects of biotin incubation, ‘sham’ samples were processed. For the biotinylation procedure, 50ml of PCs was washed twice and incubated with 5mg/L biotin, dissolved in phosphate buffered saline‐PAS‐E (1:9), for 30min. Stability of the biotin labeled platelets after irradiation was tested. Annexin V and CD62P expression were assessed as measures of platelet activation. Applicability of this method to other platelet products was assessed in three pooled PCs stored in 65% PAS‐E and three single donor apheresis PCs.

Results: The method was reproducible performed in a closed system. After biotinylation, 98.4% ± 0.9% of platelets were labeled. Platelet counts, pH and ‘swirling’ were within the range accepted by the Dutch blood bank for standard platelet products. The number of annexin V positive cells was not significantly altered by the biotinylation procedure in both fresh and stored platelets. In contrast, CD62P expression was increased in biotinylated platelets 48.4% IQR(41.7–56.2%) compared to the control samples 12.3% IQR(9.5–12.7%) on day 1 of storage. However, biotinylated platelets were not more activated compared to sham samples 50% IQR(41.7–56.2%). Thus only the procedural steps led to increased CD62P expression and not the biotin label itself. All samples showed maximal response to thrombin receptor‐activating peptide. For platelets labeled at day 7, a similar pattern was observed. Irradiation of biotin labeled platelets did not alter the stability of the biotin label nor cell quality. Furthermore this method is also applicable to pooled PCs stored in PAS‐E and apheresis PCs, with similar patterns in annexin V and CD62P expression.

Summary/Conclusions: We developed a standardized and reproducible protocol according to Good Practice Guidelines (GPG) standards, for biotin‐labeling of platelets for clinical purposes. The procedural steps, which are similar to the steps used for production of hyperconcentrated platelet products, led to an increased CD62P expression, but did not alter the annexin V expression. This method can be applied as non‐radioactive alternative to trace and recover transfused platelets in vivo.

THE A1 INSULATOR REDUCES THE GENOTOXICITY OF A BETA‐GLOBIN LENTIVIRAL VECTOR

G Kaltsounis¹, P Papayanni^1,2, P Christofi^1,2, I Valianou¹, G Karponi¹, M Yiangou², A Anagnostopoulos¹, M Sadelain³, I Rivière⁴, G Stamatoyannopoulos⁵, E Yannaki^1,5

¹Gene and Cell Therapy Center, Hematology‐BMT Unit, George Papanicolaou Hospital ²Department of Genetics, Development and Molecular Biology, School of Biology, Aristotle University of Thessaloniki, Thessaloniki, Greece ³Center for Cell Engineering ⁴Cell Therapy and Cell Engineering Facility, Memorial Sloan‐Kettering Cancer Center, New York ⁵Department of Medicine, University of Washington, Seattle, United States

Background: Insertional oncogenesis remains a major limitation for gene therapy (GT). Self‐inactivating lentiviral vectors (SIN‐LVs), although safer than γ‐retroviral vectors, still carry the risk of insertional mutagenesis, especially when, like the globin‐vectors, incorporate strong enhancers. The use of chromatin insulators (CIs) as a means to minimize vector‐mediated genotoxicity through their ability to act as enhancer‐blockers preventing the activation of endogenous genes by the vector enhancer, has been limited mainly by their large size affecting titers and suboptimal insulation. Recently, 27 human elements have been functionally characterized as robust enhancer‐blocking insulators, with the majority of them exhibiting superior enhancer‐blocking activity over the prototypic cHS4 insulator in cell lines and substantially reducing genotoxicity in a γ‐retroviral vector‐mediated carcinogenesis mouse model. In contrast to cHS4, these insulators are small‐sized (119–284bp vs 1.2kb) and can be easily accommodated in GT vectors without detrimentally affecting vector titers.

Aims: We aimed to test whether A1, one of the newly discovered CIs, could reduce vector‐mediated genotoxicity in the challenging context of SIN‐LVs, by insulating a therapeutic globin‐vector.

Methods: We tested the genotoxicity effect in the IL‐3‐dependent 32D cells, which upon transduction with oncogenic vectors become IL‐3‐independent, leading to transformation. 32D cells were transduced with SIN‐LVs: the β‐globin‐ΤΝS9.3.55‐, the insulated β‐globin‐A1‐TNS9.3.55‐ and the oncogenic SFFV‐GFP‐vector. Transduced cells were expanded in 10% IL‐3 and transduction efficiency was determined by vector copy number (VCN). Transduced 32D cells were seeded in methylcellulose with 10% or 0–1% IL‐3 to detect the IL‐3‐independent and potentially transformed clones. The IL‐3‐independent clones were further expanded in 10% IL‐3 and infused in partially myeloablated and IL‐3‐treated C3H/HeJ mice. WBC analysis, blood smears and bone marrow(BM) cytospins were performed.

Results: The A1 insulator did not negatively affect vector titers (ΤΝS9.3.55, A1‐TNS9.3.55, SFFV‐GFP: 2.8, 1.8, 2.5X10^8IU/ml, respectively). 32D cells were successfully transduced with all vectors (%VCN positive colonies: 40–100%) and expanded up to 400‐fold. The A1‐insulator decreased the number of IL‐3‐independent colonies by 78–93% over the uninsulated vectors. The uninsulated vector‐transduced, IL‐3‐independent colonies, were greatly expanded in culture with 10% IL‐3 over the A1‐transduced colonies (SFFV, ΤΝS9.3.55, A1‐TNS9.3.55: 48, 40, 10 fold change, respectively). IL‐3 independence as a transformation event was confirmed in vivo by the development of overt leukemia (hyperleukocytosis, splenomegaly, BM‐ and extramedullary site‐infiltration) in mice transplanted with the IL‐3‐independent and expanded colonies.

Summary/Conclusions: Under forced oncogenic conditions, the A1 insulator effectively protected a therapeutic vector from vector‐mediated genotoxicity. A1 may serve as a safety feature in the construction of globin‐SIN‐LVs.

WEAK D ANTIGEN EXPRESSION COMMONLY RESULTS FROM SPLICING ALTERATION: CHARACTERIZATION OF NOVEL SPLICE SITE VARIANTS AND CORRELATION BETWEEN IN SILICO PREDICTIONS AND EXTENSIVE FUNCTIONAL STUDIES

L Raud^1,2, C Ka^1,2,3, L Vigneron¹, I Gourlaouen¹, I Callebaut⁴, C Férec^1,2,3, G Le Gac^1,2,3, Y Fichou^1,2

¹UMR1078 Genetics, Functional Genomics and Biotechnology, Inserm, EFS, UBO ²Laboratory of Excellence GR‐Ex ³Laboratory of Molecular Genetics and Histocompatibility, University Hospital (CHRU), Brest ⁴Sorbonne Université, Muséum National d'Histoire Naturelle, UMR CNRS 7590, Institut de Minéralogie, de Physique des Matériaux et de Cosmochimie (IMPMC), Paris, France

Background: Novel rare nucleotide substitutions are frequently identified in RHD, the gene encoding the immunogenic D antigen of the clinically‐relevant Rh blood group system, resulting in D variant phenotype. So far, it has been commonly accepted that substitutions of amino acids located either in a transmembrane or intracellular domain of the RhD protein induce weak D phenotype, i.e. reduced D antigen density at the surface of red blood cells. Recently we showed by functional analysis using a “minigene splicing assay” (MSA) that a decrease in D antigen expression may be due also to alteration of cellular splicing.

Aims: Here we pay attention to the general disruption of this mechanism and the related phenotypic consequences in novel and previously reported single‐nucleotide variations in RHD. We then sought to characterize functionally by MSA novel candidate splicing variants in RHD. Then we extended the project by studying prospectively all single‐nucleotide variations reported in RHD exons, in order to assess globally the correlation between in silico prediction and functional analysis and to gain insights into the reliability of bioinformatics tools in line with the available phenotypic and/or clinical data.

Methods: Seventeen novel or uncharacterized RHD variations, including missense, synonymous and intronic substitutions, were selected for functional analysis by MSA in human cell models. A second set, including 46 missense variants reported in RHD exons 6 and 7, was further analyzed. Functional data were compared with an algorithm derived from the QUEPASA method and tools available in the Alamut suite. A published 3D protein model was used to visualize the location of missense amino acid substitutions and to assess potentially their respective phenotypic consequence.

Results: A novel “universal” minigene was validated and used successfully to characterize eleven novel splicing variants. Those variants include six intronic and four missense substitutions close to the consensus dinucleotide splice sites, as well as the c.1065C>T synonymous variation associated with a weak D phenotype, which creates a de novo splice site. Very interestingly, c.1154G>T (Gly385Val; D‐negative) disrupts totally normal splicing, while c.1154G>C (Gly385Ala; weak D) and c.1154G>A (Gly385Asp; D‐negative) only partially alter the mechanism. Further visualization of amino acid changes in a 3D model suggests that Gly385Asp, but not Gly385Ala, dramatically impair RhD protein structure/folding. Subsequently the global analysis of mutations in RHD exons 6 and 7 by MSA showed that inclusion of whole exon sequence in the mature transcript is significantly reduced in 15/46 (32.6%) variants, which correlates well with the QUEPASA‐like prediction (sensibility=0.93, specificity=0.94). Additionally, while normal exon inclusion is affected by c.1012C>G (weak D type 70), the associated Leu338Val substitution does not seem to be deleterious to the protein.

Summary/Conclusions: On the basis of our functional data, this work shows that splicing disruption in the presence of RHD variants is a common and general mechanism that may act independently or synergistically with alteration of protein structure through amino acid substitutions, resulting in a weak D phenotype. It also illustrates the potency of combining functional tests and in silico tools towards the phenotypic/clinical interpretation of rare variants.

INCENTIVES FOR BLOOD DONORS: DONOR RETURN RATE AMONG COMPENSATED AND NON‐COMPENSATED BLOOD DONORS

M Müller‐Steinhardt^1,2, C Weidmann³, H Klüter¹

¹Institute of Transfusion Medicine and Immunology, Medical Faculty Mannheim, Heidelberg University ²Red Cross Blood Service Baden‐Württemberg ‐ Hessen, Mannheim ³Hochschule Furtwangen University, Furtwangen, Germany

Background/Aims

Monetary and non‐monetary incentives may support blood services in recruiting blood donors but have also been criticized for violating ethical principles and threatening blood safety by attracting donors with a high risk for infectious diseases. Although incentives for blood donors have been discussed extensively over the past decades, empirical research on this topic remains limited. The aim of this study was to describe attitudes towards incentives for blood donors in Europe and show donor return rates of compensated and non‐compensated blood donors in South‐West Germany.

Methods: First, we present results of a secondary analysis of the eurobarometer, a nationally representative survey in all member states of the European Union. In 2014, participants were asked to evaluate eight potential incentives for blood donations as acceptable or unacceptable. These incentives were refreshments (e.g. coffee), physical check‐ups (e.g. blood pressure), free (testing) laboratory parameters, free medical treatment, complimentary items (e.g. first aid kits), monetary travel reimbursements, additional cash reimbursements, and release from work.

Second, we conducted a retrospective analysis of donor return patterns of 6.210 compensated and 68.205 non‐compensated donors who started donating blood at mobile and fixed donation sites. Compensated donors received either 25 EUR as a regular reimbursement for their expenses (at a fixed donation site), in accordance with the German transfusion law, or a singular free entrance for an amusement park (at a mobile donation site). These compensated donors were compared with non‐compensated donors who started either at a fixed or mobile donation site. Chi‐square statistics were used to test for differences in regular donor status after 24, 36, 48 and 72months between compensated and non‐compensated first‐time donors.

Results: Among German participants of the eurobarometer, physical check‐ups (51.9%), refreshments (50.0%) and free (testing) laboratory parameters (38.3%) showed the highest acceptance as an incentive for blood donors. Travel reimbursements and free medical treatment were rated as acceptable by 28.6% and 25.3%, respectively. The lowest acceptance was for release from work (17.8%), complementary items (13.9%) and additional cash reimbursement (11.8%). Interestingly, the acceptance of potential incentives varies considerably across Europe.

In South‐West Germany, donor return of first‐time donors differed significantly by type of compensation. Among compensated first‐time donors, who received 25 EUR as a monetary reimbursement, the proportion of regular donors after 48months (19.9%) was significantly higher than among comparable non‐compensated donors (9.2%). However, a non‐monetary compensation (free entrance) did not increase donor return rates.

Conclusion: The eurobarometer survey indicates that in most European countries monetary incentives are only accepted by a small minority. Refreshments, check‐ups, free (testing) laboratory parameters and free medical treatment were most popular as incentives for blood donors. However, results of our four non‐randomized donor samples from South‐West Germany suggest that monetary compensation may increase the likelihood of donors returning to fixed donation sites. Regular monetary reward may therefore help to recruit regular donors especially in urban settings. Incidentally, non‐monetary compensation by a free entrance, however, may not affect donor return.

DEFERRAL AS THE POINT OF NO RETURN: A MIXED METHODS APPROACH TO UNDERSTANDING WHOLE BLOOD DONOR LAPSE AFTER TEMPORARY DEFERRAL

ML Spekman^1,2, T van Tilburg², E Merz^1,2

¹Donor Medicine Research, Sanquin ²Sociology, VU Amsterdam, Amsterdam, Netherlands

Background: Previous research showed that whole blood (WB) donors that are temporarily deferred on‐site are at higher risk of lapsing, yet very little studies have focused on differentiating the effects that different deferral reasons (e.g., travel, hemoglobin [Hb]) may have on donor lapse. In addition, donor experience (i.e., first‐time or repeat donor) has also previously been found to affect donor lapse, yet novice (1–5 prior donations) and reactivated donors (returning after years of not donating) may respond differently. Finally, it is currently unclear how and why different deferral reasons and donor experience interact in influencing donor lapse.

Aims: Our aims were to understand 1) how deferral reasons and donor experience jointly affect donor lapse, and 2) why donors may lapse after temporary deferral.

Methods: A mixed methods approach was used. First, we used Sanquin's donor database for a quantitative analysis of return behavior of all Dutch WB donors between 2013 and 2015 (N=343,564). The first WB donation for each donor was identified as the target donation. Lapse was defined as non‐return within a follow‐up period of two years after the target donation. Target donations included 25% new donors, 41% novice donors, 30% experienced donors, and 4% reactivated donors. Deferral reasons included travel, Hb, medical short‐term (<28days duration), medical long‐term (>28days duration), and miscellaneous. Next, we interviewed 31 temporarily deferred donors to understand the deferral process from their perspective. Semi‐structured interviews were used to understand how these donors cognitively and emotionally experienced on‐site temporary deferral. We analyzed the interviews (using the Framework approach, cf. Hillgrove et al., BMC Public Health, 2012) to identify key topics and underlying themes.

Results: Of target donations, 11% were deferred, mostly for travel (40%), medical short‐term (23%), and Hb (19%). Survival and time‐to‐events methods showed that the different deferral reasons and donor experience levels differentially impacted donor return or lapse. Importantly, experience and deferral interacted in influencing return (rate). For instance, deferred new donors were more likely to lapse than eligible or experienced donors (ORs<0.75, p’s<0.001). Even though deferral also affected return of experienced donors, this effect was smaller or even non‐existent for certain deferral reasons (e.g., travel‐ and Hb‐related deferrals).

Qualitative results showed that almost all donors experienced temporary deferral as disappointing, particularly when it was unexpected (e.g., first‐time deferral). Not all donors (fully) understood the aims of deferral or how to prevent on‐site deferral. Donor beliefs about why deferral would lead to lapse were related to recurring deferrals, (mistakenly) interpreting deferral as permanent, or feeling all the effort did not pay off.

Summary/Conclusions: Reasons for temporary deferral differently impact risks of donor lapse at different levels of donor experience. For new donors all reasons for deferral are related to higher risks of lapse, whereas some reasons for deferral seem not to affect lapse among more experienced donors. Unexpected or recurring deferrals may explain why donors lapse after temporary deferral. Blood banks may tackle disappointment after deferral by explicitly showing that the donor is still valued, for instance by using personalized communication or offering an alternative good deed.

PREDICTION OF HEMOGLOBIN IN BLOOD DONORS USING A LATENT CLASS MIXED‐EFFECTS TRANSITION MODEL

E Lesaffre¹, K Nasserinejad², J van Rosmalen³

¹I‐Biostat, KU Leuven, Leuven, Belgium ²HOVON Data Center Clinical Trial Center, ³Department of Biostatistics, Erasmus MC, Rotterdam, Netherlands

Background: Blood donors experience a temporary reduction in their hemoglobin (Hb) value after whole blood donation. In the Netherlands, the Hb value is measured before each donation, and a too low Hb value (cut‐off values: 8.4mmol/l (135g/l) for men and 7.8mmol/l (125g/l) for women) leads to a deferral for donation, in order to prevent iron deficiency and anaemia. The minimum interval between two donations is internationally set at 8weeks, but over time donors exhibit iron deficiency so that blood donors are temporarily deferred from donation each year. In the US 35% ‐ 75% of deferrals are due to low Hb, especially in women (Editorial, Transfusion, 2012). Due to the recovery process after each donation and the unobserved heterogeneity of donors, advanced statistical methods are needed to model the longitudinal data of Hb values of blood donors.

Aims: To estimate the shape and duration of the recovery process of Hb until the Hb value has returned to its pre‐donation level, to assess whether one can distinguish between donors with fast and/or slow recovery of their Hb level and to predict future Hb values.

Methods: The study is based on data of the Donor InSight study, which was a prospective cohort study performed by Sanquin in the Netherlands from 2007 to 2016. We employed three statistical models for the Hb value: (i) a mixed‐effects models, (ii) a latent‐class mixed effects model, and (iii) a latent‐class mixed‐effects transition model. In each model, a flexible function was used to model the recovery process after donation. The latent classes identify groups of donors with fast or slow recovery times, and donors whose recovery time increases with the number of donations. The transition effect accounts for possible state dependence in the observed data. All models were estimated in a Bayesian way, using data of a sample of 1000 new entrant donors (500 males and 500 females). Prior information from the clinical literature (Boulton, Vox Sanguinis 2004) about the recovery process three days after blood donation was incorporated into the analysis since these values were not identified in the observed data.

Results: The results show that the latent‐class mixed‐effects transition model fits the data best. We also found that the recovery process shows a concave process (initially fast followed by slower recovery). The estimated recovery time is much longer than the current minimum interval of 58days between donations. Namely, depending on the subgroup that the donor belonged to, males showed a recovery time of 100 to 419days, while the estimated recovery time for females varies between 54 to 503days. These results suggest that an increase of this interval may be warranted.

Summary/Conclusions: The analysis shows the usefulness of the sophisticated statistical models that make use of historical information to model complex processes in time, in this case the Hb trajectory over time across repeated donations. In addition, our results suggest a (much) longer time lag between subsequent donations to avoid anemia.

AN INTERVENTION STUDY FOR THE PREVENTION OF VASOVAGAL REACTIONS: ANALYSIS OF DONORS’ RETURN FOR SUBSEQUENT DONATION

JC Wiersum‐Osselton¹, F Prinsze², T Marijt‐van der Kreek³, E van den Brekel⁴, F Hermans⁵

¹Unit Donor Affairs, Sanquin Blood Bank, Leiden ²Donor Studies ³Unit Donor Affairs, Sanquin Blood Bank, Amsterdam ⁴Unit Donor Affairs, Sanquin Blood Bank, Nijmegen ⁵Unit Donor Affairs, Sanquin Blood Bank, Deventer, Netherlands

Background: Complications of blood donation are known to reduce donors’ return for future donation. The EPISoDe study (Experience Success in Donation) showed that water drinking shortly before donation had an effect of 23% reduction of self‐reported vasovagal reactions (VVR) in younger novice whole blood donors (Wiersum‐Osselton, Transfusion, 2019).

Aims: In this study we analysed the return for a subsequent donation of the donors participating in the EPISoDe study. This was a predefined secondary outcome of the EPISoDe study.

Methods: The EPISoDe study was conducted in young (<30years) whole blood donors making their first, second, third or fourth donation in geographically selected collection centres. The study interventions were: 330ml water drink, 500ml water drink or squeezing a ball (placebo intervention) during the wait after the screening interview and before phlebotomy, and a control group without intervention. Participating donors were sent an online questionnaire about their experience within a week following their donation attempt.

In the Netherlands donors are usually invited for blood donation in accordance with hospitals’ needs; the aim is to invite eligible donors at least once a year. Donors were included in the return analysis if they had received at least one invitation within 400days after the index donation and we analysed their return for a donation attempt within 421days. Associations with the interventions and donors’ donation status, gender and reported symptoms at their index donation were analysed by calculating return percentage of eligible donors and by binomial logistic regression.

Results: Out of the 8199 EPISoDe participants who had received an invitation, 6538 (79.7%) returned within the study period. There was no difference in donor return between the two water groups. The likelihood of return was significantly increased in both water and placebo intervention donors compared to the questionnaire group (OR 1.2, 95% CI 1.06–1.4 and 1.3, 1.12–1.5 respectively). Return was slightly lower in women (OR 0.88, 0.78–0.99) and lower in first‐time donors (OR 0.86, 0.77–0.96) than after a 2^nd–4^th donation. A staff‐recorded or self‐reported VVR at the index donation reduced donor return (OR 0.47, 95% CI 0.37–0.60 and OR 0.53, 0.46–0.61 respectively). Other symptoms following donation were also associated with a lower return percentage.

Summary/Conclusions: In this cohort of younger new and novice blood donors, 79.7% returned for a subsequent donation. A VVR (either staff‐recorded or self‐reported) reduced donor return. Donors who received a study intervention, either water or placebo, were more likely to return, whether or not they had suffered a VVR. It is conceivable that the mere fact of study participation could also have increased donor return, even in de questionnaire group; this will be examined in the total population of target group donors.

HEALTH STATUS, DONATION PATTERNS, AND IRON‐DEFICIENCY IN OLDER BLOOD DONORS‐ EVIDENCE FROM A LARGE POPULATION‐BASED COHORT AND HEALTH SERVICES USE DATA

S Karki, S Wright, D Irving, T Davison

Research and Development, Australian Red Cross Blood Service, Sydney, Australia

Background: The contribution of older blood donors to the blood supply is substantial. In Australia, donors aged>50years contributed 41% of all donations made in 2017. However, with ageing, the general health status of older donors changes relatively faster, thus progressively affecting their ability to donate. An in‐depth understanding of the relationship between older donors’ health status, future donation patterns, and risk of iron‐deficiency could be of a great value to inform the Blood Service to predict the number of future donations, and manage the risk of iron‐deficiency.

Aims: To understand the relationship between self‐reported health, blood donation patterns, and the management of identified iron‐deficiency in older blood donors.

Methods: We linked the Sax Institute's 45 and Up study baseline data collected between 2006 and 2009 to the Blood Service donation records, inpatient records, and Medicare records*. The data‐linkage was conducted by Centre for Health Record Linkage. Using these linked data, we examined the relationship between health, donation patterns, and iron‐deficiency and its management.

Results: We followed up 22,058 active whole blood donors for 200,403 eligible person‐years (average age at recruitment 56.4years, 56.3% female, average follow up 9.1years per‐person). After adjusting for the effect of age, sex, body‐mass index, education, non‐English language spoken at home, country of birth, smoking, physical activity, regular use of multivitamins, alcohol consumption at enrolment, and total number of whole blood donations in the 2years prior to enrolment, participants with better self‐reported health at recruitment showed significantly higher rates of donation. Excellent, very good, good, and fair/poor health status donors made 1066 (95% CI 1056–1075), 987 (981–993), 900 (891–909), and 710 (690–730) donations per 1000 person‐years, respectively.

Iron‐deficiency was identified in 8.9% of donors in the study (n=1964, 95% CI 8.5–9.3). Sixty percent of those with iron deficiency (n=1,175, 95% CI 57.6–62.0) visited their general practitioner (GP) within 60days of the identification of iron‐deficiency, and 48.4% (95% CI 45.5–51.3) of those visiting GP underwent further iron status examination and monitoring.

After adjusting for several potential confounders including the total number of donations made during the follow‐up period, excellent self‐reported health status was independently associated with lower risk of iron‐deficiency (p for trend= 0.004).

Summary/Conclusions: Information on self‐reported health status can be an effective indicator to estimate the future donation yield of an older blood donor panel, and risk of developing iron‐deficiency. Donors with better self‐reported health had a higher number of future whole blood donations and a lower risk of iron‐deficiency. Donors referred to GPs for management of their iron status utilised the health services as expected, however there is an opportunity to improve their contact with their GPs.

* Medicare records was provided by Australian Government Department of Human Services.

THE IMPORTANCE OF ANAEMIA MANAGEMENT AND IMPACT IN THE HEALTHCARE SETTING

A Mo^{1, 2, 3}, Z McQuilten^{1, 2}, E Wood^{1, 2}

¹Transfusion Research Unit, Monash University, Melbourne ²Department of Haematology, Monash Health, Clayton ³Departments of Clinical and Laboratory Haematology, Austin Health, Heidelberg, Australia

Anaemia is a major public health issue, affecting 25% of the population worldwide according to the World Health Organization. Iron deficiency is responsible for approximately half of all cases globally, with other causes including anaemia of chronic disease, other nutritional deficiencies, haemoglobinopathies, renal impairment, malignancy and bone marrow disease. In the elderly, where anaemia is even more common, the cause is frequently multifactorial.

Anaemia is associated with increased mortality, decreased cognitive and physical function, depressive symptoms and fatigue, particularly in older adults. Poor outcomes have also been reported in anaemic patients with underlying comorbidities such as cardiac and renal disease, and cancer.

Within a hospital setting, anaemia is highly prevalent. Preoperative anaemia, affecting up to 30% of patients, is associated with poor clinical outcomes including higher in‐hospital mortality, longer length of stay and higher ICU admission rates.

Anaemia management requires a proactive and multi‐faceted approach, typically involving a multi‐disciplinary team in which the transfusion practitioner plays a vital role. This includes screening of high‐risk patients and pre‐admission clinics to identify and manage patients at high risk of peri‐operative anaemia. Implementation of patient blood management (PBM) guideline recommendations has been shown to be effective to prevent and optimally manage anaemia within the community and hospital settings. The transfusion practitioner has key roles in the coordination, monitoring and auditing of PBM programs. Active patient involvement and engagement of all members of the multidisciplinary team, including primary care clinicians, are also key to enhance the success of such programs.

THE ROLE OF THE TRANSFUSION PRACTITIONER IN ANAEMIA ASSESSMENT AND MANAGEMENT: PROCESSES, TIPS AND RESOURCES FOR CREATING CHANGE

L Bielby¹, R Moss²

¹Dept of Health & Human Services, Victoria and Australian Red Cross Blood Service, Blood Matters, West Melbourne, Australia ²Department of Laboratory Medicine, Great Ormond Street Hospital for Children NHS Foundation Trust, London, United Kingdom

Background: Patient blood management (PBM) is an evidenced based integrated multi‐disciplinary approach aimed to improve clinical outcomes by effectively managing and conserving the patient's own blood, thus reducing unnecessary exposure to transfusion. PBM has the patient as the central focus with the aim being to improve their outcomes and include them in the process.

PBM includes three pillars: 1) optimising the patient's own blood, 2) minimising blood loss and 3) optimising a patient's physiological tolerance of anaemia.

Delayed assessment/management of anaemia contributes to increased health costs and unnecessary blood transfusions, and transfusion has been recognised to be associated with increase morbidity and mortality.

The term Transfusion Practitioner (TP) includes those known as transfusion nurses, transfusion safety officers, haemovigilance officers, or patient blood management (PBM) coordinators. A key aspect of the role is driving and influencing clinical blood management activities to help align practice to internationally recognised guidelines and standards, including PBM.

Aim: To demonstrate the TP role in anaemia assessment & management and discuss strategies, processes, tips and resources for creating organisational and cultural change to implement PBM.

Context: Literature outlines the importance of a multidisciplinary team to implement PBM related changes, and TPs play a fundamental role within these teams to support ‘buy in’. TPs are seen as enablers, pulling resources together, engaging with those involved, providing education and facilitating change. They are often the ones to conduct audits, collating data and evaluating outcomes.

Approaches to implement PBM should be tailored to suit individual organisations. The authors will outline different approaches, highlighting where the TP can support or lead activities. One approach to anaemia assessment is to undertake an audit, examples of available tools will be shown. With this data, the TP along with the PBM team can explore options for corrective action. These could include interventions such as developing a pathway where all or a specific group of patients are assessed and or treated either at a preoperative clinic, or with their local general practitioner; through to more complicated strategies such as establishing anaemia clinics.

The skills of the TP are a valuable asset to analyse clinical specialties/patient mix who should be targeted to achieve best outcomes, they know the organisation and as such are well placed to help develop a process/concept that will suit, and they can provide education and support to promote and embed these practices.

Conclusion: Appropriate assessment and management of anaemia requires a multi‐disciplinary approach. The TP plays an active and crucial role in this team. Examples of processes, tips and resources to support change and embed a PBM culture across the clinical spectrum will be shared.

INNATE IMMUNITY AND IMMUNE HAEMOLYTIC ANAEMIA

S Zeerleder

Department of Hematology and Central Hematology Laboratory, Inselspital Bern, Bern, Switzerland

Immune haemolytic anaemia (IHA) is characterized by an increased breakdown of red blood cells (RBCs) due to allo‐ and/or autoantibodies directed to RBC antigens with or without complement activation. Clinical and laboratory signs of haemolysis in concert with the presence of a positive direct antiglobulin test characterize IHA. Alloantibodies formed during pregnancy and/or after prior transfusions may cause acute or delayed haemolytic transfusion reaction after transfusion of a RBC product incompatible with the specificity of the alloantibody. Autoantibodies to RBCs reduce the survival of endogenous and hamper the recovery of donor RBCs after transfusion. Lymphoproliferative disease, autoimmune disease, infection or drugs often cause autoantibodies to RBC, but frequently no obvious cause can be identified.

Besides the antigen specificity, the isotype critically determines the biological activity of RBC antibodies in vivo. The isotype defines the affinity to Fc‐gamma receptors on cells of the reticuloendothelial system as well as the capacity to activate the classical pathway of complement, IgM being the most effective. Antibody‐mediated complement activation results in the opsonisation of RBC with C3bc/C3d with subsequent complement receptor‐mediated removal by phagocytes (extravascular haemolysis). Occasionally, complement activation proceeds via the activation of C5 to the formation and insertion of the membrane attack complex resulting in intravascular haemolysis.

There is growing evidence that the innate immune system plays an important role in the pathogenesis of IHA. The process of complement‐mediated haemolysis results in systemic inflammation, which contributes to morbidity and mortality of patients suffering from IHA. Complement activation results in the release of anaphylatoxins, which are strongly vasoactive and mediate chemotaxis, inflammation and formation of radical oxygen species. Release of cell‐free haemoglobin and cell‐free haeme upon haemolysis induces endothelial cell activation, NO‐depletion, cytotoxicity, ROS formation and neutrophil activation. Natural plasma scavengers, such as haptoglobin and hemopexin complex with their target molecules, cell‐free haemoglobin and haeme, with subsequent removal of the complexes via CD163 and CD91‐mediated phagocytosis. Although being positive acute phase proteins due to consumption the plasma scavengers become exhausted during chronic haemolysis thereby failing to prevent the adverse biological effects of cell‐free haemoglobin and haeme in the circulation. Inducible haeme oxygenase‐1 (HO‐1) is an efficient cellular scavenger by breaking down haeme into biliverdin with subsequent formation of bilirubin, CO and ferrous iron with subsequent oxidation to ferric iron and storage by the ferritin H chain. HO‐1 has an established role in the systemic protection from systemic inflammation induced by haemolytic and non‐haemolytic diseases. The lecture will emphasise the role of innate immunity with a special focus on different plasma‐ and cellular systems involved in the pathogenesis of systemic inflammation in patients suffering from IHA.

UNDERSTANDING ERYTHROCYTE CLEARANCE

C Roussel, P Amireault, P Ndour, P Buffet

Research and teaching, Institut National de la Transfusion Sanguine, Paris, France

The clearance of erythrocytes is essential in physiology, disease and transfusion. Elimination of erythrocytes altered because of senescence or pathological processes is expected to protect the microcirculation from obstruction by adhesive or rigid erythrocytes. It also contributes to the harmful consequences of anemia and hemolysis in hereditary and acquired red blood cells diseases as well as in conditions associated with auto‐ or allo‐immunization. Immunobiology has explored in great details antibody‐mediated clearance of erythrocytes but conventional approaches may not be fully operational to explain Delayed Hemolytic Transfusion Reactions. Some important clearance processes are independent from the recognition of molecules or antigens on the erythrocyte surface. Increased erythrocyte stiffness triggers their clearance in hereditary spherocytosis, malaria and possibly also in the context of autoimmune anemia. Knowns and unknowns on the mechanisms and sites of erythrocyte clearance will be presented based on a critical review of old and recent contributions.

RED BLOOD CELLS AS MODULATORS OF IMMUNITY

R Riganò, B Buttari, E Profumo

Cardiovascular and Endocrine‐metabolic Diseases and Aging, Istituto Superiore di Sanità, Rome, Italy

Existing literature indicates that red blood cells (RBCs), beyond gas transport, exert a complex role in human physiology, being involved in many functions essential to maintain ion, metabolic and immunological homeostasis. RBCs display an immunomodulatory activity on adaptive immune cells by promoting T cell growth and survival and inhibiting activation‐induced cell death. The balance between cell death and survival controls T cell homeostasis and anomalies in this balance account for diseases linked to excessive or faulty T cell growth. RBCs are able to modulate innate immunity by binding endogenous molecules such as chemokines and mitochondria‐derived DNA, as well as external agents such as pathogens. RBCs can also directly modulate innate immune cell activation or tolerance by controlling the maturation of the circulating pro‐inflammatory subset of dendritic cells (DCs). These cells are potent inducers of primary antigen‐specific T cell responses, produce TNF‐a when stimulated by LPS and are the principal IL‐12p70‐producing cells among leukocytes. The pro‐inflammatory capacity of circulating DCs is controlled by RBCs that are able to inhibit their maturation and IL‐12 production. In diseases characterized by local Th1 inflammatory response such as psoriasis vulgaris and rheumatoid arthritis, pro‐inflammatory DCs play a role in the induction and perpetuation of inflammation. Collectively, literature data indicate that RBCs exert important modulatory functions that may result in immune activation or quiescence, depending on the environmental conditions. When RBCs encounter a microenvironment characterized by an intense production of ROS, the RBC defenses get overwhelmed or are unable to counteract the new pro‐oxidant status and become themselves a source of ROS, which cause the generation of senescent signals on RBCs. The major feature of oxidized RBCs is the clustering and/or the breakdown of Band 3. Other features are the complexation of Hb with spectrin, the loss of glycophorin A, the externalization of phosphatidylserine and the reduction of the “marker of self” integrin‐associated protein CD47. A similar senescence phenotype has been documented in RBCs during the storage period. Oxidized, senescent or stored RBCs, due to surface antigen modification and to the release of pro‐inflammatory molecules, fail to control immune cell homeostasis thus contributing to the perpetuation of inflammation and to the pathogenesis of immune‐mediated diseases associated to oxidative stress, such as autoimmune diseases and atherosclerosis. Our research group demonstrated that RBCs from patients with carotid atherosclerosis presented a senescent phenotype similar to that acquired by RBCs from healthy subjects following to in vitro oxidation. Oxidized erythrocytes fail to control T lymphocytes apoptosis and lipopolysaccharide‐induced monocyte‐derived DC maturation, thus representing dangerous signals for adaptive and innate immunity and contributing to the pathogenesis of atherosclerosis. In conclusion, the crosstalk between RBCs and the immune system represents a mechanism to maintain immunological homeostasis. However, in high oxidative stress conditions, that can take place during a prolonged storage period or in particular diseases, RBCs can acquire a pro‐oxidant behaviour and lose their functional and homeostatic features. By interfering in immune system homeostasis, RBCs become a potential tool that can be manipulated to improve or reverse pathological situations characterized by anomalies in the control of adaptive and innate immunity.

TRANSFUSION THERAPY IN SICKLE CELL DISEASE: WHICH INDICATIONS AND FOR HOW LONG?

K Yazdanbakhsh, H Zhong, Y Liu, F Vinchi, A Mendelson

New York Blood Center, New York, NY, United States

Transfusion therapy remains an important treatment modality for patients with sickle cell disease (SCD). Transfusions are given to lower the percentage of circulating sickle RBCs, and to decrease blood viscosity and have been shown in clinical trials to reduce the risk of stroke by 90%. However, many indications for transfusion in SCD remain controversial partly due to insufficient randomized clinical trials data and in part because of our limited understanding of the complex pathologic networks leading to diverse disease complications in SCD despite the common single mutation. Similarly, we have incomplete mechanistic understanding of why chronic transfusion protocols must be continued for those indications supported by clinical data. The beneficial effects of transfusion therapy in SCD need to also be weighed against potential transfusion risks including alloimmunization associated with life‐threatening delayed transfusion reactions, increased iron stores associated with increased oxidative stress and exposure to infectious agents. We believe that a deeper understanding of the benefits as well as harmful effects of transfusions is crucial to optimize our current transfusion therapy protocols in SCD. This knowledge may provide highly needed guidance, which is currently lacking, for expansion or limiting existing indications for chronic transfusions in SCD.

TREATMENT OF THALASSAEMIA

Y Aydinok

Department of Pediatric Hematology, Ege University, Faculty of Medicine, Bornova/Izmir, Turkey

Thalassaemia is a devastating blood disease with a significant worldwide burden. Annually, 60,000 children are born with a major thalassemia. Life‐time RBCC transfusions and iron chelation remain standard of care treatment in thalassaemia. Transfusion therapy still account for significant iron overload related morbidity and mortality despite chelation therapy which is associated with poor adherence, safety concerns and varied efficacy. Higher risk for transfusion transmitted infections (TTIs) exists for thalassemia patients whose transfusion exposure sustains lifelong. Although, the risk of transmission for traditional viruses is exceedingly rare in the modern era, emerging infectious diseases continue to be recognized as potential threats to transfusion safety. The inadequacy of blood safety points to the necessity for an additional layer of security for the blood supply in the developing world. Pathogen reduction technologies for RBCC may imply a proactive, more generalized approach against new and re‐emerging pathogens in the developed world and may be an ultimate safeguard for transfusion safety in the developing countries. RBC alloimmunization may become a major challenge in thalassaemia management. Prevention is the key reducing the burden of alloimmunization. While the recommendation is to transfuse thalassaemic patients with C/c,E/e,Kell compatible blood, it is not universally practiced. Extended molecular RBC typing may be an appropriate adjunctive test in addition to serological typing before embarking on transfusion therapy. If a complete RBC antigen profile has not yet been performed in an alloimmunized patient, genotyping is the only option for accurate detection of RBC antigens that may guide the antibody identification. Allogeneic stem cell transplantation (A‐SCT) is the only available curative therapy in children with HLA matched sibling which is available to approximately 20% patients. In the absence of MSD, MUD transplant with high compatibility criteria has still limited experience. Mismatch related, cord blood and haploidentical donor SCTs are considered experimental. A‐SCT carries a substantial risk of SAEs and mortality, both increasing with recipient age and disease severity. DFS is 88% in paediatric and 65% in adults. Gene therapy for correction of the α‐globin chain imbalance overcomes the problems of donor availability and immunologic complications associated with A‐SCT. Multicenter clinical studies on gene addition therapy by using self‐inactivating lentiviral vector are currently underway. Recently, gene editing by either gene disruption or gene correction emerged as a potential alternative to gene addition therapy in beta‐thalassaemia. A new era of novel therapeutics is unfolding in thalassemia management. Several targets have been identified that can improve alpha/beta chains imbalance, ineffective erythropoiesis, or iron dysregulation and a number of those now have agents in preclinical and clinical development. Hydroxyurea may improve globin chain imbalance and be beneficial for reducing or omitting transfusion requirement in selected group of patients. Ruxolitinib has shown the limited effect on pretransfusion haemoglobin and reduction in transfusion needs, but allowed steady decrease in spleen volume that may serve for avoiding splenectomy in beta thalassaemia. Luspatercept may restore normal erythroid differentiation and improves anaemia and hepcidin mimetics or TMPRSS6 inhibitors may modulates ineffective erythropoiesis by iron restriction and improves anaemia and organ iron loading.

SAFEGUARDING BLOOD SAFETY FOR PATIENTS WITH THALASSAEMIA

C Politis¹, A Eleftheriou², C Richardson³, E Zervou⁴, M Angastiniotis², P Englezos²

¹Hellenic Coordinating Haemovigilance Centre, Hellenic Centre for Disease Control and Prevention, Athens, Greece ²Thalassaemia International Federation, Nicosia, Cyprus ³Panteion University, ⁴Hellenic Centre for Disease Control and Prevention, Athens, Greece

Background: Thalassaemia major (particularly β‐type) and Sickle Cell Disease (SCD) are the commonest clinically important haemoglobinopathies, representing major sources of morbidity. Recommended therapy is regular transfusion of safe, good quality blood, and monitoring of related complications. Thalassaemia International Federation (TIF) guidelines, in place since 1987, include strategies for precautionary measures and use of scientific progress in detection, inactivation and elimination of transfusion transmissible pathogens. Antigen–matching strategies to avoid alloimmunization against RBC antigens and other measures including haemovigilance are key components for safe blood, alongside voluntary, non‐remunerated blood donation and laboratory quality assurance programmes.

Aims: We present the contribution of TIF and the Greek experience in ensuring safety and availability of blood for thalassaemia patients applying internationally accepted standards and recommendations.

Methods: TIF ‐ a non‐profit, patient‐driven organization with 204 national thalassemia associations in 62 countries ‐ promotes national control programmes for prevention and management contributing to the achievement of final cure. The main working methods are provision of education, expert support, networking, communications and projects to support improvements in the quality of health, social and other care.

In Greece, technical standards for blood donor selection and testing are applied in compliance with Directive 2004/33/EC as well as haemovigilance programmes and traceability procedures for recording adverse reactions and events associated with the transfusion of RBCs (Directive 2005/61/EC). Pre‐transfusion and transfusion measures recommended by the Council of Europe are applied. In particular, measures for transfusion of “the right blood at the right time for the right patient”, leucodepletion, RBC washing and accurate cross‐matching and antigen and antibody screening for an extended matching policy are practised. Fresh (up to 15days old) RBCs are used. Molecular testing for ABO and Rh D is performed in cases with blood group discrepancies.

Haemovigilance in Greece covers 95% of total blood supply. Data on TTIs in 3,067 patients with thalassaemia and SCD‐thalassaemia in 2010–2017 are analysed.

Results: TTI prevalence in thalassaemia syndromes was: HBV 1.8% (occult type 1.3%), HCV 54%, HIV 0.3%, HTLV 0.8%, WNV 0.5% and HEV 3%. Most frequent adverse reactions in 2015–17 were allergic (incidence 1:854), non‐haemolytic febrile reactions 1:2,631, “Other” 1:5,883, alloimmunisation 1:6,667, TACO 1:100,013, TAD 1:50,006, TT‐HEV 1:100,013. Hyperhaemolysis was diagnosed in two SCD patients, delayed haemolytic transfusion reaction in one thalassaemia intermedia patient.

Trends in 2010–2017 show reduced incidence of alloimmunisation against RBCs. Rates of allergic and pyrexial ARs remained stable. No major ABO incompatibility case was reported and no fatal transfusion reaction of transfusion has been recorded.

Summary/Conclusions: Blood safety in transfusion has significantly improved in high and upper‐middle income but unfortunately not in lower and low income countries. Blood shortages and lack of stringent protective measures for thalassaemia patients is the reality for many developing countries.

TIF focuses particular attention on the provision of support and the promotion of initiatives promoting the safety and adequacy of blood – a key component of the lifelong management of patients with transfusion‐dependent thalassaemia.

HEPCIDIN GENE POLYMORPHISMS AND IRON OVERLOAD IN MAJOR Β‐THALASSEMIA PATIENTS REFRACTORY TO IRON CHELATING THERAPY

P Zarghamian, A Azarkeivan, A Khazaeli, S Majid

Blood Transfusion Research Center, High Institute for Research and Education in Transfusion Medicine, Tehran, Iran

Background: β Thalassemia is the most common group of hereditary hemoglobinopathy diseases. Affected people with major thalassemia are dependent on regular blood transfusion which leads to iron overload. Hepcidin is a peptide and an important regulator of iron homeostasis. Expression of this hormone is influenced by polymorphisms within the hepcidin gene, HAMP.

Aims: This study aimed to analyze the association of three polymorphisms in promoter of HAMP, rs10421768, rs117345431, and rs142126068 with iron overload in major β thalassemia patients who do not respond to iron chelating therapy.

Materials and Methods

a total of 102 samples from major β thalassemia patients were collected. Genomic DNA was extracted and sequenced for SNPs rs10421768, rs117345431, and rs142126068. Statistical analysis was performed on IBM*SPSS* STATISTIC 23 using Independent t test and Fisher test.

Results: Our analysis revealed statistically significantly difference between the level of cardiac iron concentration and c.‐582A>G variant (P=0.02). For rs117345431 statistical analysis was on the edge of significant relationship between minor allele and serum ferritin (P=0.058). All samples were homozygous for allele T of rs142126068.

Summary/Conclusions: different factors affect iron overload in thalassemia. Our findings and others emphasize the role of hepcidin polymorphism as a key component in iron homeostasis.

ENDING AIDS IN AFRICA ‐ VISION OR ILLUSION?

ND Labhardt^1,2

¹Infectious Diseases and Hospital Epidemiology, University Hospital Basel ²Medicine, Swiss Tropical and Public Health Institute, Basel, Switzerland

Ten to twenty years ago, countries in South Eastern Africa faced the peak of the devasting HIV/AIDS epidemic leading to an up to 20years drop in general life expectancy. With the burden of HIV/AIDS falling mainly on the economically active population of young and medium‐aged adults, the epidemic endangered social and economic stability in nations most heavily affected. Today, despite AIDS still being a major cause of death in South Eastern Africa, the epidemic has become an example of public health gains that can be achieved through programmatic, evidence‐based approaches that are endorsed by globally aligned policy and funding strategies.

Based on his work from Lesotho, where one out of four among adults is infected by HIV, Niklaus Labhardt will take the auditors through the history of HIV programs in South Eastern Africa and show how innovative, pragmatic and evidence based implementation brought the region to a stage where the goal to end the AIDS epidemic by 2030 might be in reach.

EVOLVING BLOOD DONOR DEFERRAL POLICY FOR MEN WHO HAVE SEX WITH MEN IN FRANCE: IMPACT ON THE RISK OF HIV TRANSMISSION BY TRANSFUSION

J Pillonel¹, C Pelat¹, C Sauvage¹, B Danic², C Martinaud³, F Barin⁴, I Sainte‐Marie⁵, B Coignard¹, S Gross⁶, P Tiberghien⁶, S Laperche⁷, F Lot¹

¹Direction des Maladies Infectieuses, Santé publique France, Saint‐Maurice ²Etablissement Français du Sang, Rennes ³Centre de Transfusion Sanguine des Armées, Clamart ⁴Centre National de Référence du VIH, Centre Hospitalier Régional Universitaire, Tours ⁵Agence Nationale de Sécurité du Médicament ⁶Etablissement Français du Sang, Saint‐Denis ⁷Centre National de Référence Risques Infectieux Transfusionnels, INTS, Paris, France

Background: In France, the deferral for men who have sex with men (MSM) was reduced from permanent to 12months in July 2016. Since this change has not impacted the residual risk (RR) of undetected HIV among blood donations, the Ministry of Health is considering a greater access of blood donation to MSM. Two scenarios have been studied: S1. deferral of MSM during the 4months preceding the donation; S2. deferral of MSM who have had more than one sexual partner in the 4months preceding the donation, similarly to all other blood donors in France.

Aims: To assess the impact of these two scenarios on the HIV RR estimated over the period July 2016‐December 2017 which is the baseline RR with the current 12‐month deferral for MSM.

Methods: Baseline HIV RR was calculated with the classical Incidence‐Window‐Period method, where HIV incidence was derived from a detuned assay (EIA‐RI) detecting recent infections (≤180days) since all HIV‐1 antibodies positive blood donations are tested with this test. Theassessment of the impact of both scenarios on the baseline HIV RR was based on (i) data obtained from surveys among MSM in the general population and in blood donors (compliance survey), to estimate the number of additional MSM who would give blood in each scenario, and on (ii) HIV incidence estimate among these additional donors. This incidence was estimated: for S1, from MSM blood donors with the current deferral policy (12months) and for S2, from monogamous MSM of the general population.

Results: From July 2016 to December 2017, 8/28 (29%) HIV‐1 positive blood donors tested with the EIA‐RI were identified as recently infected, allowing to estimate the baseline HIV RR at 0.16 in 1 million donations [95% CI: 0.04–0.34], or 1 in 6,380,000 donations. For S1, the number of additional MSM donors was estimated at 733 and the number of additional HIV positive donations at 0.09, resulting in an HIV RR of 0.16 in 1 million donations [95% CI: 0.04–0.35] or 1 in 6,300,000 donations. For S2, the number of additional MSM donors was estimated at 3,103 and the number of additional HIV positive donations at 4.92, resulting in an HIV RR of 0.23 in 1 million donations [95% CI: 0.05–0.56] or 1 in 4,300,000 donations. Sensitivity analysis shows that if both the number of MSM and the HIV incidence were multiplied by 1.5, the risk would be 1 in 6,225,000 donations for S1, and 1 in 3,000,000 for S2.

Summary/Conclusions: For both scenarios, the HIV RR remains very low. For S1 (4‐month deferral), the risk is identical to the baseline RR and is very robust to variations in the model parameters. For S2 (no more than one sexual partner, 4months), the risk is 1.5 higher than the point estimate of the baseline RR and sensitivity analysis shows that this estimate is less robust than for S1, since the risk could be 2 times higher than the baseline RR. For both scenarios, there was a modest increase in eligible MSM donating.

RISK FACTORS FOR RECENTLY ACQUIRED HIV INFECTION IN BLOOD DONORS IN SOUTH AFRICA

K van den Berg^1,2, M Vermeulen³, G Jacobs³, R Swanevelder⁴, J Hemingway‐Foday⁵, D Creel⁵, U Jentsch⁶, E Murphy^7,8, B Custer^7,9, for the NHLBI REDS‐III South Africa Program.

¹Department of Translational Research, South African National Blood Service, Port Elizabeth ²Department of Clinical Haematology, University of Cape Town, Cape Town ³Department of Operations Testing ⁴Department of Business Intelligence, South African National Blood Service, Roodepoort, South Africa ⁵RTI International, Research Triangle Park, United States ⁶Specialised Laboratory Services, South African National Blood Service, Roodepoort, South Africa ⁷Department of Laboratory Medicine, UCSF, San Francisco, United States ⁸Research and Scientific Programs, Vitalant Research Institute, San Francisco, South Africa ⁹Research and Scientific Programs, Vitalant Research Institute, San Francisco, United States

Background: Recruiting safe blood donors amongst the largest HIV‐positive population in the world is a major challenge for South African blood transfusion services. South African donor deferral criteria and deferral periods for perceived high risk activities have evolved over time, but current risk factors for infection have not been formally assessed. In addition, most studies have reported risk factors for prevalent HIV infection whereas risk behaviours for incident infection are more informative as donations with these infections could occur during the window periods of available screening assays.

Aims: To identify the demographic and behavioural risk factors associated with incident HIV infection among blood donors in South Africa.

Methods: We conducted a case‐control study with incident HIV‐infected blood donors compared to infectious marker negative controls. Incident HIV cases and controls seronegative for HIV, hepatitis B and C viruses and syphilis were accrued from a donor pool covering 8 of 9 provinces in South Africa. Controls were frequency matched at a 3:1 ratio to cases on race, age and geography. Incident HIV‐infections were HIV RNA positive by individual donation nucleic acid amplification testing (ID‐NAT; Procleix, Grifols) but antibody (Ab) negative (PRISM, Abbott) as well as those RNA+/Ab+ donors with recently‐acquired HIV based on Limiting Antigen Avidity (LAg) assay results with normalized optical density values of<1.5. Eligible cases and controls completed a confidential audio computer assisted structured interview (ACASI) on motivations for blood donation and behavioural factors, including behaviours in the 6months before donation. Frequencies and measures of statistical association for risk behaviours comparing cases and controls are reported after adjusting for multiple comparisons.

Results: From November 2014 to January 2018, we enrolled 323 incident HIV cases and 877 controls; 202 (62.5%) cases and 544 (62%) controls were≤29years old. There were significantly more female cases 230 (71.2%) than female controls 439 (50.1%) (P<0.0001). Significant HIV risk factors (all P<0.0001) reported within the 6‐months before donation included: having a primary sex partner who is male; reporting increasing numbers of male sexual partners for both females and males; frequency of vaginal sex; frequency of vaginal sex without condoms; use of methods to clean, dry, or tighten one's anus before sex; and having visited a traditional healer for medical care. Lack of medical aid (private health insurance) and reports of injury or accident with blood loss were also associated with an incident HIV infection.

Summary/Conclusions: Our study has identified a set of novel, putative risk factors for incident HIV infection among South African blood donors while confirming a number of previously known sexual risk behaviours. Not having private health insurance and being injured may be markers of socio‐economic context that place individuals at higher risk rather than behaviours that directly increase HIV transmission risk. The detection of risk behaviours by ACASI in donors who passed pre‐donation questionnaires and interviews suggests that ACASI has the potential to improve risk behaviour identification.

HIV‐1 CLUSTERS AND RISK FACTORS IN BLOOD DONORS FOUND POSITIVE FOR HIV IN FRANCE

P Cappy^1,2, A Chaillon³, A Essat⁴, J Pillonel⁵, M Chaix^6,7, P Tiberghien⁸, L Meyer⁹, F Barin^10,11, S Laperche^1,2

¹Departement des Agents Transmissibles par le Sang ²CNR Risques Infectieux Transfusionnels, INTS, Paris, France ³Division of Infectious Diseases, UCSD, La Jolla CA, United States ⁴INSERM CESP U1018, Université Paris Sud, Le Kremlin‐Bicêtre ⁵Département des maladies infectieuses, Santé publique France ⁶Service de Virologie ‐ CNR VIH, Hôpital Saint‐Louis – APHP ⁷INSERM U944, Université Paris Diderot, Paris ⁸UMR 1098 INSERM, Université de Franche‐Comté, EFS, Besançon ⁹Service de Santé Publique, Hôpital Bicêtre ‐ APHP, Le Kremlin‐Bicêtre ¹⁰Laboratoire de Virologie associé au CNR VIH, CHRU de Tours ¹¹INSERM U1259, Université de Tours, Tours, France

Background: In France from 2000 to 2016, among male blood donors (mBDs) found HIV‐1 positive at blood donation screening, 31% did not disclose any risk factor for HIV infection during post‐donation interviews, while 38% reported having sex with men (MSM), and 24% and 7 % reported heterosexual sex (HTS) and other risk factors, respectively.

Aims: In order to gain new insights into the risk factors for HIV‐1 infection in mBDs, we performed an HIV‐1 genetic network analysis, including HIV‐1 positive mBDs and patients included in the French primary HIV infection ANRS CO6 PRIMO cohort (PC).

Methods: 284 mBDs, who donated blood between 2000 and 2016, and 1340 PCs, included between 1998 and 2014, were studied. Epidemiological data were collected by the French Blood Service (EFS) upon blood donation or post‐donation interviews for mBDs, and upon inclusion for CPs. Viral strains were sequenced and genotyped in pol gene, and a recent infection assay was performed to date infection in mBDs (recent: <6months). A partial transmission network was computed based on Tamura‐Nei 93 nucleotidic distance (threshold for HIV‐1 s/t B=1.5%; for non‐B s/t=0.8%) and assortative mixing was evaluated for mBDs epidemiological data, including risk factors for HIV infection (MSM, HTS, others and unknown). Self‐reported data were then compared to assortativity‐enhanced data.

Results: HIV‐1 strains from 81 mBDs and 126 PCs were linked into 54 clusters including at least one mBD. PRIMO‐only clusters were excluded from the analysis. Compared to mBDs who did not cluster, those found linked to the network were younger (30 vs. 36year‐old; P<0.01) and were more likely to have a recent infection (48% vs. 34%; P=0.03). Assortative mixing indexes showed that paired individuals were more likely to live in the same area (P<0.001) and to have the same risk factor for HIV infection (P<0.001) compared to a random distribution. Imputing MSM risk factor to non‐MSM individuals paired with MSM changed the distribution of risk factors as follows: MSM: 38% vs. 49%, HTS: 24% vs. 21%, other: 7% vs. 6% and unknown: 31 vs. 24%.

Summary/Conclusions: After validating the assortativity of risk factors between paired individuals, and imputing MSM risk factor to individuals self‐reported as non‐MSM (including those with no identified risk factor), up to 18% (31/176) of mBDs could be reclassified as MSM. This is a worst‐case scenario, as the network analysis does not exclude the possibility of one or several persons between two paired individuals (missing link). Altogether, these results could help reevaluate the HIV residual risk linked to MSM mBDs, especially in the frame of the evolution of blood donor deferral criteria.

FROM UNIVERSAL TO SELECTIVE HTLV SCREENING IN THE UK

K Davison¹, H Harvala², C Reynolds³, S Brailsford³

¹NHSBT/PHE Epidemiology Unit, Public Health England ²Microbiology Services, ³NHSBT/PHE Epidemiology Unit, NHS Blood and Transplant, London, United Kingdom

Background: Although most individuals remain asymptomatic, HTLV infection can lead to adult T‐cell leukaemia/lymphoma (ATLL) and HTLV‐1 associated myelopathy (HAM). The serious nature of these diseases, evidence of transmission via non‐leucodepleted blood, and concern about a high prevalence among donors originating from endemic areas led to the UK blood services introducing universal blood donation screening in 2002. Monitoring through routine surveillance commenced and HTLV‐infected donors were invited to participate in the HTLV National Register cohort study to assess disease progression. These data together with evidence from lookback to previously untested donations and cost‐effectiveness analysis were reviewed by an expert working group in 2013 and 2015.

Aims: To describe the epidemiology of HTLV among UK blood donors and evidence of disease progression from long term follow up of asymptomatic donors.

Methods: UK blood donations screened, and infected donors identified are reported to a national surveillance scheme. These donors are contacted, their results explained and information about clinical history and possible sources of infection are collected. Where appropriate, HTLV‐infected donors are consented to the register, with participants completing a baseline questionnaire about their health, flagged in registries for cancer or death, and followed up about every 2–3years.

Results: In the UK 2002 – 2017, 254 HTLV‐infected donors were identified. Prevalence among new donors was steady around 5/100 000 donations. Prevalence among repeat donors peaked in 2002 (2.7/100 000 donations), with most in previously untested. From 2004 to 2016, prevalence of 0.07 per 100,000 donations (average of one positive/year) was recorded. In 2017, prevalence among new donors increased to 9.9/100,000 donations (17 positives), with increased numbers associated with Asian ethnicity and coinciding with an increase in collections from BAME groups.

Overall, most were women (182/254, 72%), UK‐born (125/254;49%) and HTLV‐1 infections (228/254;90%). Mean age was 43years. Almost all positive donations were from previously untested donors (240/254), with seroconversion within a year of previous donation confirmed for only 5 of the 14 previously tested donors. Typically, infections were associated with endemic countries (including Caribbean region, West Africa, Iran, India and Japan), acquired through breast feeding or from their heterosexual partner originating from these countries. Interestingly, three were thought to have been infected through self‐flagellation. A total of 114 HTLV‐positive asymptomatic blood donors have already been recruited to the HTLV National Register, and during over 800‐person years follow‐up, none had developed ATLL or HAM.

Summary/Conclusions: Over 16years of testing, few seroconverters were identified, suggesting very little ongoing transmission among UK blood donors. The lack of disease among the cohort study was also reassuring, although it is likely too early to detect associated symptoms of a slow progressing disease. Recruitment to this unique dataset continues, also outside of the blood donation setting. As a result of these surveillance data, evidence from lookback, and cost‐effective analysis, in 2017 NHSBT ceased to test donations from previously tested donors unless the donation was being used to manufacture a non‐leucodepleted component.

MANAGEMENT OF BLOOD SUPPLY IN A COUNTRY WITH LONG DISTANCES

M Syrjälä

Blood Service, Finnish Red Cross, Helsinki, Finland

Finland lies in Northern Europe between the 60^° and 70^° N latitude. The length of the country is 1150km and width 550km. By surface area it is the fifth largest country in EU. The population of the country is 5.5 million resulting in the lowest population density in EU (15.8 inhabitants/km3). The whole country is inhabited, although most of the population is packed in the south. The climate of Finland is influenced mainly by its latitude, but the warm waters of the Gulf Stream and the North Atlantic Drift Current also play a role. Due to Finland's northern location, winter is the longest season. The southern portions of the country are snow‐covered about three or four months of the year, and the northern regions for about seven months. Long distances, low population density and the extreme climate give logistical challenges. It is estimated that these logistical costs can be as much as 10–20% of GDP in Finland.

The Finnish Red Cross Blood Service (FRC BS) has been the nationwide blood service provider in Finland since 1948. FRC BS collects annually about 200 000 whole blood units of which 50 % are collected in 10 fixed sites and 50 % in mobile sessions around the country. Central activities (donor recruitment, medical support, production, testing, supply chain management, digital services and administration) are located in Helsinki.

Management of transfusion is highly dependent on the logistical arrangements from blood donation sites to the central facilities and from the central inventory to the hospitals.

The logistics is outsourced to three major partners all of whom have their roots in nationwide public transportation and logistics services. Posti Ltd is a state owned company having its roots in the national postal and telecom office. Today it is the leading postal and logistics service company having the widest network coverage in Finland. Blood units collected at different fixed sites and mobile sessions are transported overnight by Posti Ltd to the FRC BS central facilities by 8 am on the day following the blood donation. Posti Ltd is also used for the regular deliveries of blood products to the hospitals. The other important partner is Matkahuolto Ltd, which was founded in the 1930s to maintain bus stations and to serve as a common marketing company for the bus and coach services in Finland. It maintains a nation‐wide package delivery system based on the scheduled bus route network. Matkahuolto LTD is used to transport donor testing samples from the donation sites to the central laboratory. By this arrangement it is possible to obtain most of the donor samples to the laboratory around midnight, which significantly speeds up the completion of laboratory results. The third logistics partner is JetPak Finland Ltd, which operates the air freight for the national flight company Finnair.

Blood transfusion services can be managed centrally in a large sparsely populated country in a manner that is of high quality, safe and cost effective. However, the supply chain has to be planned carefully.

EDUCATING THE MASSES: THE USE OF ELEARNING IN TRANSFUSION MEDICINE

D Peterson, T Clark, T Verrall, L English, S Ogley

Centre for Education and Training, BloodSafe eLearning Australia, North Adelaide, Australia

Background: Elearning is a divisive topic. It is often criticised as an inferior form of education while simultaneously being promoted as a means to provide education to large numbers of people in a consistent, cost‐effective manner.

BloodSafe eLearning Australia (BEA) is a government‐funded blood transfusion education program that commenced in 2007 and provides courses in clinical transfusion practice and patient blood management (PBM) including:

• ‐

  Clinical Transfusion Practice (4 courses)

• ‐

  PBM: General (7 courses)

• ‐

  PBM: Medical (6 courses)

• ‐

  PBM: Acute Care and Surgical (4 courses)

• ‐

  PBM: Obstetrics and Maternity (3 courses)

• ‐

  PBM: Neonates and Paediatrics (7 courses)

Aims: To determine the engagement, outcomes and impact of learning of BloodSafe eLearning Australia courses.

Methods: A retrospective analysis of user registrations, course completion records, course evaluation data and red cell usage in Australia to determine learner demographics, and the impact on acquisition of knowledge and application to clinical practice.

Results: In the period from 1 July 2007 to 31 January 2019:

• ‐

  489,600 people registered as learners

• ‐

  1,072,299 courses were completed

• ‐

  These learners came from 182 countries, with 11,141 (2.3%) of them from outside of Australia.

Analysis by profession shows that:

• ‐

  82.4% are nurses and/or midwives

• ‐

  11.2% are medical

• ‐

  6.4% are laboratory, anaesthetic technicians or other.

Analysis of user evaluation data (n=2,885) from 1 April 2015 to 31 January 2018 shows that these courses have a positive impact, with 88.7% of respondents stating they gained additional knowledge, 65.2% able to make changes to clinical practice, and 88.2% reporting that these changes will improve patient safety and outcomes. Analysis of international participants shows greater benefits with 93.5% gaining knowledge, 76.4% able to change their clinical practice and 91.9% believing this will improve patient outcomes.

Analysis of red cell usage in Australia shows that since 2012 there has been a 21.1% reduction in red cells issued. This has been achieved through a number of PBM activities including development of guidelines, research and audits, education, waste reduction strategies, and promotional campaigns. BloodSafe eLearning Australia courses on PBM were released in 2012 and are one part of this PBM activity, and it is notable that these courses have the widest reach as they are undertaken by a large proportion of doctors, nurses and midwives in Australia who are not directly involved with the blood sector.

Stakeholder feedback shows that the program provides credible, consistent education that is cost‐effective, reduces duplication, is ‘best‐practice’ elearning, is readily accessible, and allows institutions to focus on the development of practical transfusion skills.

Summary/Conclusions: This analysis shows that elearning is a well‐accepted, well‐utilised form of education for healthcare workers to learn about clinical transfusion practice and patient blood management, and learners gain knowledge that can change their clinical practice and improve patient outcomes. It is also likely that these courses have contributed to better utilisation of a scarce, freely‐donated resource.

This approach has global reach and availability, and is a cost‐effective model for improving transfusion practice in the developing world by providing education for millions of healthcare workers.

PROSPECTIVE PLATELET AUDITING: ANALYSIS OF TRAINEE COMPLIANCE WITH GUIDELINES

S Vossoughi, J Schwartz

Pathology, Columbia University, New York, United States

Background: Apheresis platelets are a component product with high cost and limited supply. Furthermore, there is a potential for severe transfusion reactions associated with this product such as transfusion related lung injury (TRALI), and sepsis due to bacterial contamination. Therefore, transfusion guideline compliance is closely monitored by many centers. This quality assurance analysis describes our experience with prospective platelet auditing performed by physicians in pathology residency training.

Aims: This study aims to evaluate the ability of physicians in training to perform prospective auditing and compare policy compliance for different levels of experience.

Methods: This is a quality assurance analysis of a prospective platelet audit program for a 12‐month period (January 2018‐December 2018). The blood bank paged the on call physician any time an order was placed for a patient with a platelet count of>50,000/μl, ≥2 doses of platelets with no interim repeat count, or an unknown platelet count. Audit records created by physician trainees in their first post graduate year (PGY1) were compared to subsequent years (PGY>1). Information collected included the total number of doses requiring approval, number of products approved, training year for the approving physician, and transfusion indication. Cost analyses assumed $500 for a dose of platelets. Descriptive statistics and comparative analysis using a Pearson's Chi‐Square were used with a difference of P<0.05 considered statistically significant.

Results: There were 1446 platelet doses requiring approval with 670 (46%) routed to the PGY1 group and 776 (54%) to the PGY>1 group. There were 847 (59%) ordered doses that were in compliance with hospital transfusion policy and 599 (41%) that were not in compliance with hospital policy. Of the 847 appropriately ordered doses, the PGY1 group declined release of 7 necessitating the clinical team to insist upon release without approval, and there were zero such instances in the PGY>1 group. When paged by the blood bank, PGY1 physicians approved product release not in compliance with policy for 191/670 (29%) doses while PGY>1 physicians approved not indicated products for 79/776 (10%) of doses (P<0.01). Products not indicated by hospital policy were held from release by PGY1 physicians for 113/670 (17%) doses and 216/776 (28%) doses by PGY>1 physicians (P<0.01). The ordered doses not in compliance with hospital policy had an estimated cost of $299,500. Of this cost, there was a calculated $164,500 savings of products not released due to prospective auditing. There was an additional potential savings of $135,000 for products not indicated but released ($95,500 from the PGY1 and $39,500 from the PGY>1 group).

Summary/Conclusions: Despite a higher number of requests being routed to the more senior PGY>1 group, there were a disproportionately higher number of out of compliance platelet orders being released by the PGY1 group in addition to withholding needed products on several occasions. Potential mitigation strategies for this could include a closer level of oversight for PGY1 physicians, and the potential monetary savings could justify a hiring a dedicated patient blood management team or quality assurance manager to monitor compliance and provide feedback to clinicians.

WHAT CAN WE LEARN FROM HOW ADVERSE EVENTS ARE DETECTED?

Ø Flesland, C Steinsvåg, A Espinosa

Norwegian Directorate of Health, Oslo, Norway

Background: The primary aim of reporting systems, such as haemovigilance systems, should be learning and improvement and to identify risk areas, not simply counting errors. To understand and learn from adverse events the description of how, where and why they occur, and how they are detected, is important. To support our understanding, we use a predetermined classification that is required for reporting to EU, supplemented by classification suggested by IHN, WHO and ourselves. In 2015 we started asking the blood establishments what steps they would take to prevent recurrence of the event, and we added a simple classification to tell how the adverse event had been detected.

Aims: This study aims to analyze how different types of adverse events reported to the haemovigilance system were detected, whether the current quality management systems used in Norwegian blood establishments had effective barriers and whether new barriers should be considered.

Methods: Adverse events reported to the Norwegian haemovigilance system in 2017 and 2018 were analyzed with focus on how the adverse event had been detected. In all cases classification had been performed by the reporter of adverse events and were confirmed, or reclassified if necessary, by the haemovigilance team before analysis. For analysis based on classification we used PowerBI (Microsoft).

Results: A total of 188 adverse events were reported from Norwegian blood establishments. All had been classified according to how the adverse event had been detected. Twenty (10.6 %) adverse events were detected because of alarms or warnings from IT‐systems or equipment. Routine checks by blood establishment staff detected 39 (20.7 %) events and formal internal or external reviews detected one event. Seven (3.7 %) events were detected because the donor became ill shortly after donation, but the illness was not caused by the donation. Sixty‐four percent of events were detected in a way that did not fit our present classification and hence were classified as “Other”.

Twelve out of 16 wrong blood in tube were detected by an alarm from the IT‐system or routine check, as were six of 22 events related to blood ordering, two of seven errors in testing, six of 17 events where incorrect blood had been transfused, and eight of 64 events related to donor selection.

In 80 reports human error was listed as the cause of the event and 27 of these were detected by alarms or routine checks.

Summary/Conclusions: Detection of adverse events by alarms or routine checks are highly efficient when the blood establishment has historic data to check against, as exemplified with wrong blood in tube or a patient require irradiated blood components. When no historical data exists or when the quality management systems do not require routine checks, events are usually detected by chance. Further analysis is needed to see if and where the quality management systems should be improved. The wide variety of adverse events can make it difficult to select which area to prioritize in the improvement work.

HEMOGLOBIN MEASUREMENT – HOW TO EVALUATE NURSES’ COMPETENCE AND THE ACCURACY OF THE POINT‐OF‐CARE (POC) METHOD

J Castrén, P Korkalainen, M Arvas, A Valkeajärvi, S Bäckman

Finnish Red Cross Blood Service, Helsinki, Finland

Background: Hemoglobin measurement from the finger prick is an important working step of blood donors’ eligibility screening. Too low hemoglobin is the most common reason of donor deferral in Finnish Red Cross Blood Service (FRCBS). 3.5 % of donation attempts were rejected due to low Hb in 2018. A proper and correct technique in donor Hb measurement could affect significantly the Hb deferral rate.

Aims: The aim of this study was to ensure the nurses’ competence to measure the hemoglobin from the finger prick. The aim was also to gather a representative data set of Hb measurements in order to evaluate the accuracy of the method currently in use for donor Hb measurements.

Methods: Each nurse participated in the practical skill test (n=168) documented POC measurements of donor's Hb from five donors eligible for donation. A total of 845 POC Hb measurements were analyzed in this study. Additionally venous blood samples were taken from the blood bag's sample pouch and analyzed in the laboratory using a cell counter in order to evaluate the accuracy of the POC method.

Results: The Hb measurements from the finger prick were on average 1.2g/L (0.9%) higher than from the venous blood samples. The range of the difference was ‐21 ‐ +22g/L. These results were used in order to add novel information to determine the measuring uncertainty of Hb measurement in FRCBS. In 1.4 % (12/845) of the donors in this study the venous hemoglobin measurements were below the cut‐off point of donor eligibility. In those measurements the difference of the finger prick and venous hemoglobin measurement was at most+9g/L.

92 % of the hemoglobin results from the finger prick were in the range ±10g/L compared to the venous hemoglobin results. 63 % of the results from the finger prick were between ±5g/L (the precision of the device) compared to venous hemoglobin results.

In 13 cases the difference between finger prick and venous measurements was outside 2 standard deviations from the mean i.e. 2.5 % from the bottom (n=9) or top (n=4) of distribution. Systematic errors were seen in some nurse's results both towards too low or too high Hb result in the finger prick measurement and some nurses had random errors in both directions.

The batch of cuvettes, donors’ age, gender or the time of sampling were not detected to have an impact on the difference between finger prick and venous Hb measurements in this study.

Summary/Conclusions: The results of the POC measurements compared to the cell counter were in agreement with published data and with manufacturers’ information on the device.

The practical skill test is a workable way to develop competence and operations to measure the hemoglobin from the finger prick. It offers an opportunity to give personal feedback to nurses concerning their personal performance in the use of the current Hb measurement technic. It also provided data on the accuracy of the POC method in the everyday donor selection process.

BLOOD DONATION AND IRON LOSS, WHAT ARE THE IMPLICATIONS?

MG van Kraaij¹, M Sweegers², M Janssen², K van den Hurk²

¹Units Transfusion Medicine and Donor Affairs ²Donor Medicine Research, Sanquin Blood Supply, Amsterdam, Netherlands

Background: Whole blood donation has frequently been related to iron deficiency. A blood donor loses per donation about 8% (men) to 81% (menstruating women) of iron stores. To replenish the iron lost by blood donation in a donation interval of 56days, a donor needs to absorb 4.5mg iron per day. This amount exceeds the reported maximal amount of absorbed iron of 3–4mg/day, eventually leading to iron deficiency, with consequences such as donor deferral and possibly iron deficiency‐related symptoms (decreased physical endurance, fatigue, pica, restless legs syndrome, and cognitive functions).

Since Hb levels do not reflect donors’ true iron status, measuring ferritin is a better way to detect low iron stores in whole blood donors. Studies from USA and Denmark showed that on the introduction of ferritin measurement with either extension of donation intervals or iron supplementation in case of low iron stores, deferral percentages for low Hb declined in both male and female donors.

Aims: To gain more insight in iron status of whole blood donors during their donor career, how this affects donor health and which measures may prevent low iron stores in donors.

Methods: In the Netherlands, Sanquin Blood Bank is currently implementing a policy with ferritin‐guided donation intervals. In brief, ferritin levels are measured in all new donors and in repeat donors every 5th donation or in case of an Hb below the deferral threshold. Donation intervals are extended if ferritin levels are<15μg/L, or≥15 and≤30μg/L (for 12 and 6months respectively). We anticipate that routine ferritin measurement will ultimately result in a lower prevalence of iron deficiency, less Hb deferrals and improved donor retention. This will be further evaluated in a stepped wedge cluster‐randomized trial ‘FIND'EM’, which may also identify subgroups of donors prone to develop (symptoms of) iron deficiency. In addition, implementing ferritin screening may lead to a decreased donor availability. For this purpose, we modeled the impact of the implementation of our ferritin deferral policy on donor availability over time, which provides insight for both the expected size of the impact of the ferritin deferral policy and the time and rate at which this impact is expected to occur. This allows the blood bank to timely plan actions to counterbalance possible donor shortage and ensure an adequate blood supply. Lastly, iron supplementation can be an alternative measure instead of donation deferral. As the used and recommended dosage of iron supplementation varies widely across blood services, Sanquin is planning to start a new study in whole blood donors to gain evidence on the dosage and frequency of iron supplementation and its effect on ferritin and hemoglobin levels and donor health.

Results: The before‐mentioned studies are ongoing and results will be expected from 2020 onwards.

Summary/Conclusions: Iron deficiency is a frequent side effect of whole blood donation. To prevent iron deficiency and its consequences, like donation deferral and health issues, more evidence‐based insight in iron management of whole blood donors is being generated.

SUPERDONORS – GENETIC RISK PROFILE AND RISK OF LOW HEMOGLOBIN DEFERRAL

AS Rigas¹, C Erikstrup², O Pedersen³, K Rostgaard⁴, G Edgren⁵, L thørner¹, E Sørensen¹, K Banasik⁶, K Burgdorf¹, H Hjalgrim⁴, H Ullum¹

¹clinical immunology, university hospital Copenhagen, University Hospital Copenhagen, Copenhagen ²clinical immunology, Åarhus University Hospital, Åarhus ³clinical immunology, Næstved Hospital, Næstved ⁴epidemiology research, Statens serum institut, Copenhagen, Denmark ⁵clinical epidemiology unit, Karolinska Institute, Stockholm, Sweden ⁶NNF center for protein research, University of Copenhagen, Copenhagen, Denmark

Background: No reliable method exists for stratifying new blood donors into those who can maintain sufficient hemoglobin (Hb) levels and those who will be deferred because of a low Hb (<7.8mmol/l [<12.57g/dl] for women and<8.4mmol/l [<13.54g/dl] for men). Polygenic risk scores (PRSs) have shown great promise in predicting complex disease risk. PRSs could also prove useful for identification of donors genetically predisposed to low Hb levels, and, thus, to an increased risk of deferral.

Aims: The objective of the study was to evaluate the association between PRS (modelled to predict Hb level as a quantitative trait) and risk of deferral as a binary outcome.

Methods: The Danish Blood Donor Study (DBDS) is an ongoing nationwide blood donor cohort since 2010with more than 110,000 participants. Extensive genotyping has been performed on approximately 72,000 DBDS participants using the Infinium Global Screening Array (Illumina^®) and extended by use of imputing based on the pan‐Scandinavian reference genome. Based on Hb and genetic data on more than 150,000 Icelandic individuals (an independent discovery cohort), we constructed 9 different weighted PRSs for individuals from DBDS. Information on the donors’ whole blood donations following inclusion into DBDS unto end of 2017 was obtained from a nationwide donation database, SCANDAT. The best predictor of Hb among the nine PRSs was chosen and used in all subsequent analyses. We performed multilevel mixed‐effects linear regression analysis with Hb as outcome, and PRS as factorized explanatory variable with cutoffs at 5, 10, 25, 40, 60, 75, 90, and 95^th percentiles, respectively. Moreover, the models had a two‐level clustering on donor ID and donation site and an ID‐specific random intercept; and further adjusted for: sex(binary), age(continuous), year of donation(factorized), and time since last donation (continuous). Lastly, risk of deferral was evaluated in random effects logit models with similar covariables and clustering structure.

Results: Mean number of donations per donor after DBDS inclusion was 6.9 donations. Generally, we observed a statistically significant positive association between PRS(Hb) and current Hb levels. Compared with donors in the 40‐60 PRS percentile group, donors below the 5^th percentile had lower (‐0.23 Hb mmol/L (95% CI: ‐0.25; ‐0.21)) and donors above the 95^th percentile higher (+0.19 Hb mmol/L (95% CI: 0.18; 0.21) Hb levels. In the random effects logit models we observed a marked increase in deferral risk with decreasing PRS percentile strata. With the 40–60 PRS percentile stratum as reference, donors below the 5^th percentile and donors above the 95^th percentile had odds ratios of deferral of OR=2.58 (95% CI: 2.27; 2.93) and OR=0.46 (95% CI: 0.39; 0.54), respectively.

Summary/Conclusions: We found a statistically significant positive association between PRS(Hb) and Hb levels and a markedly increased risk of deferral with decreasing PRS(Hb). From a scientific point of view, it is unsurprising that a genetic score for Hb from an independent cohort is associated with Hb in another cohort. However, from a practical perspective, PRSs may be the first step in a personalized donation approach to donors and their risk of deferral.

MODELING OPTIMAL INTER‐DONATION INTERVALS WITH PERSONALIZED RISK ASSESSMENT FOR ADVERSE IRON OUTCOMES

W Russell^1,2, D Scheinker^1,3,4, B Custer^2,5

¹Management Science and Engineering, Stanford University, Stanford ²Vitalant Research Institute, San Francisco ³Lucile Packard Children's Hospital ⁴School of Medicine, Stanford University, Stanford ⁵Laboratory Medicine, University of California San Francisco, San Francisco, United States

Background: Individually calibrated inter‐donation intervals for repeat blood donors have the potential to minimize the risk of iron related adverse outcomes (e.g., hemoglobin deferral or collecting a donation from a donor with low or absent iron stores) without unduly impacting the donated blood supply. Machine learning has shown promise for personalized clinical risk assessment.

Aims: Our aim is to use machine learning to develop donor‐specific, personalized inter‐donation intervals that minimize the risk of adverse outcomes while maintaining or improving the adequacy of the donated blood supply.

Methods: Using a public use dataset from the REDS‐II Donor Iron Status Evaluation (RISE) study (Cable, Transfusion, 2012) we defined donor profiles with physiological measures including hemoglobin, ferritin and soluble transferrin receptor along with questionnaire responses regarding diet, reproductive health indicators, and demographics. We used these profiles (58 features, 3,162 donations from 1,025 repeat donors) and the time until the next donation attempt to predict iron‐related outcomes of the next donation attempt. Possible outcomes were no adverse outcome, hemoglobin deferral, low‐iron donation (ferritin<20ng/ml for women and<30ng/ml for men), or absent‐iron donation (ferritin<12ng/ml for men and women). We trained multiple machine learning models on 2,543 of the donations and selected the model with the best performance (lowest cross‐entropy loss in cross validation). We assessed the best model's performance on a hold‐out test set of 620 donations, which were not used to train or select the model. We then used our model to generate risk estimates for these 620 test donors as a function of days since their last donation, which varied from 56days to 250days. To show individual variation, we generated graphical representations of individual donors’ risk over time.

Results: Ferritin, log ferritin, body iron, and time since last donation were most useful for predicting iron‐related adverse outcomes at the next donation attempt. The estimated risk of adverse outcomes at the next donation attempt varied considerably across donors. As expected, the risk of adverse outcomes 250days after the last donation was lower than the risk 56days after the last donation for most donors (risk of hemoglobin deferral decreased for 84% of donors; risk for low‐iron donation decreased for 61%; and risk for absent‐iron donation decreased for 94%).

Summary/Conclusions: The risk of iron‐related adverse outcomes as a function of time since last donation varies considerably between donors. Machine learning models trained on relevant donor profiles can effectively estimate how an individual's risk will change over time. Individual risk estimates could allow blood centers to protect high‐risk repeat donors while continuing to allow more frequent collections from low‐risk donors. Further study is needed to ensure this approach works well for donor classes that are not well‐represented by the RISE dataset, to assess risk prediction outside of the physiological measures collected in the RISE study, and to determine the viability of assigning an optimal inter‐donation interval to a first‐time donor using this approach.

A COMPOSITE MEASURE OF HEME IRON CONSUMPTION PREDICTS INCIDENT IRON DEPLETION IN REPEAT BLOOD DONORS

BR Spencer¹, M Fox², L Wise², R Cable³

¹Scientific Affairs, American Red Cross, Dedham, MA ²Epidemiology, Boston University School of Public Health, Boston, MA ³Scientific Affairs, American Red Cross, Farmington, CT, United States

Background: Iron depletion is common among repeat blood donors, who contribute a large proportion of the blood supply in many countries. Exogenous iron from multivitamins with iron or iron‐only supplements helps prevent donation‐induced iron depletion, but whether dietary iron protects against iron depletion in repeat donors has not been rigorously evaluated. Available data from the REDS‐II RISE study in the US (Cable, Transfusion, 2011) and from the Danish Blood Donor Study (Rigas, Transfusion, 2014) suggest minor impact of dietary iron consumption on blood donor iron status in multivariable regression models. Both studies, however, analyzed food items singly, such as beef or fish, rather than in aggregate, so precision was limited.

Aims: To evaluate whether a composite measure of dietary heme iron consumption, weighted for frequency and iron content, was associated with incident iron depletion among repeat blood donors.

Methods: A re‐analysis of the RISE cohort was undertaken to test the hypothesis that reported levels of animal protein consumption was associated with lower risk for incident iron depletion among repeat blood donors. The six blood centers participating in RISE enrolled first‐time and frequent donors for 15–24month follow‐up of donation frequency and iron status. A brief checklist of 8 food categories was administered at baseline to assess frequency of consumption of several categories of animal protein that are rich in heme iron, the biochemical form of iron most readily absorbed. An Iron Composite Score (ICS) weighted for frequency and heme iron content was derived and subjects were grouped into tertiles (thirds) of ICS. Iron status was assayed at enrollment and study completion and at roughly one‐third of donation visits in between. Modified Poisson regression with Generalized Estimating Equations was used to generate risk ratios controlling for donation frequency and other covariates.

Results: Of 2425 enrolled donors, 1406 were iron replete at baseline and completed the food checklist. The median value of the ICS for each tertile (lowest to highest) was 7.2, 13.1, and 22.0mg of heme iron weekly. These values are equivalent to approximately 3, 6, and 9 servings of beef per week, or alternately twice as many servings of chicken or pork. Across 2236 follow‐up visits with iron outcomes assayed, almost 33% of donor visits were associated with intermediate iron depletion (serum ferritin<26ng/ml) and 8.5% with complete depletion of iron stores, representing serum ferritin<12ng/ml. After controlling for demographic factors and donation frequency, the lowest tertile of ICS was associated with a greater than 2‐fold higher risk for complete iron depletion during all follow‐up visits (RR 2.40, 95% CI 1.51, 3.81, compared to the highest tertile).

Summary/Conclusions: In this longitudinal evaluation of dietary iron and iron status, blood donors with low intake of heme iron had an elevated risk for developing advanced iron depletion. These results suggest that blood centers should continue to recommend iron‐rich diets to repeat blood donors.

DIETARY INTAKE OF HEME IRON IS ASSOCIATED WITH FERRITIN AND HB LEVELS IN DUTCH BLOOD DONORS: RESULTS FROM DONOR INSIGHT

T Timmer¹, R deGroot¹, J Rijnhart², J Lakerveld², J Brug³, C Perenboom⁴, M Baart⁴, F Prinsze¹, S Zalpuri¹, W de Kort⁵, K van den Hurk¹

¹Donor Medicine Research, Sanquin Research ²Epidemiology and Biostatistics, Amsterdam UMC ³Amsterdam School of Communication Research (ASCoR), University of Amsterdam, Amsterdam ⁴Division of Human Nutrition and Health, Wageningen University & Research, Wageningen ⁵Public Health, Amsterdam UMC, Amsterdam, Netherlands

Background: Blood donors lose approximately 250mg of iron with every blood donation. As a result, frequent blood donors are at risk of iron deficiency and low hemoglobin (Hb) levels, which may affect their health and eligibility to donate. Lifestyle behaviors such as dietary iron intake and physical activity, may influence iron stores and thereby Hb levels.Gaining insight into associations between lifestyle behaviors and Hb levels is valuable for blood supply organizations, as lifestyle behaviors can potentially be considered to prevent Hb deferrals. Examining the mediating role of ferritin, a measure reflecting iron stores, in these associations will help to gain insight into whether iron stores could be the limiting or enabling factor that links lifestyle behaviors to Hb recovery after donation.

Aims: To investigate associations between lifestyle behaviors (dietary heme and non‐heme iron intake and physical activity) and Hb levels, and whether ferritin mediates these associations.

Methods: Donor InSight‐III (DIS‐III) is a Dutch cohort study of blood and plasma donors andincluded 2,552 donors. Participants who were pregnant, had hemochromatosis, used iron supplements/medication, got a hysterectomy orbilateral oophorectomy were excluded (n=292). Hb levels were measured in EDTA whole blood samples using a hematology analyzer (XT‐2000, Sysmex, Japan) and ferritin was measured in plasma from lithium heparin tubes (Architect Ci8200, Abbott Laboratories, U.S.A.).Dietary heme and non‐heme iron intake (grams/day) were assessed using a food frequency questionnaire adapted to measure iron intake. Moderate‐to‐vigorous physical activity (MVPA, minutes/day) was assessedusing the International Physical Activity Questionnaire (IPAQ)‐Short Form.

Results: In total,2,260 (1,186 female) participants were included.Donors with higher intakes of heme ironhad significantly higher Hb levels (regression coefficient (β) (95% confidence interval (95% CI))in men and women respectively: 0.147 (0.069 to 0.225) and 0.094 (0.013 to 0.175) mmol/L), independent of age, smoking, menstruation, number of donations in previous two years, donation interval, sedentary behavior, the other lifestyle variable (i.e. (non‐)heme iron intake or MVPA), and initial Hb level. Non‐heme iron intake was negatively associated with Hb levels (‐0.015 (‐0.026 to ‐0.004) and ‐0.018 (‐0.032 to ‐0.005) mmol/L for men and women respectively). Ferritin mediated associations between dietary iron intake and Hb levels(indirect effectin men and women respectively: 0.073 (0.046 to 0.107) and 0.065 (0.031 to 0.100) μg/L for heme and ‐0.003 (‐0.008 to 0.000) and ‐0.007 (‐0.013 to ‐0.002) for non‐heme). More MVPA was negatively associated with Hb levels in men only (‐0.004 (‐0.008 to ‐0.001)), which was not mediated by ferritin.

Summary/Conclusions: In conclusion,higher heme and lower non‐heme iron consumption are associated with higher Hb levels in donors via higher ferritin levels, indicating that donors with high heme iron consumption may be more capable of maintaining iron stores to recover Hb levels after blood donation.More MVPA was associated with lower Hb levels, although effect sizes were small, independent of ferritin. Taking a donor's lifestyle behaviors into account may be useful in preventing low Hb levels in blood donors.

DETECTION OF PLATELET AUTOANTIBODIES TO IDENTIFY ITP STATE OF THE ART

M de Haas^1,2,3, L Porcelijn¹

¹Immunohematology Diagnostics, Sanquin, Amsterdam ²CCTR, Sanquin ³IHB, LUMC, Leiden, Netherlands

Immune Thrombocytopenia (ITP) is still diagnosed by exclusion of other causes for thrombocytopenia. Sensitive and specific detection of platelet autoantibodies may support the clinical diagnosis and prevent misdiagnosis of ITP. For example, the direct monoclonal antibody immobilization of platelet antigens (MAIPA) assay, performed with in vivo sensitized patient platelets, offers platelet glycoprotein specific autoantibody detection with high accuracy. A drawback is that low platelet counts demand a large blood sample to have sufficient patient platelets available for analysis. Circulating platelet autoantibodies are more difficult to detect by MAIPA; and may demand more sensitive detection platforms, such as those using surface plasmon resonance.

In general, the presence of anti‐GPIIb/IIIa, anti‐GPIb/IX and anti‐GPV platelet autoantibodies is investigated. All these antibody specificities have been found in patients with ITP. In ITP, platelet autoantibody‐mediated destruction via the spleen has been proposed; but also other mechanisms leading to low platelet counts in ITP may play a role. Inhibition of megakaryocytopoiesis by autoantibodies or by T cells has been suggested. In mice, GPIb‐directed antibodies induce loss of platelet‐sugar epitopes, inducing hepatocyte‐medicated platelet destruction. Platelet autoantibodies can cause complement activation, which may contribute to platelet autoantibody‐mediated destruction. Interestingly, we recently found that lack of detectable platelet autoantibodies is correlated with non‐responsiveness to rituximab (CD20 MoAb) treatment in ITP patients. In children with newly diagnosed and often transient ITP, platelet autoantibodies of IgG class or not often found, but of IgM class are present for short duration.

In conclusion, testing for platelet autoantibody characteristics and their pathologic effect may be helpful in establishing the diagnosis of ITP and in choosing the best individualized therapy for ITP patients.

THROMBOPOIETIN RECEPTOR AGONIST (TPO‐RA) TREATMENT RAISES PLATELET COUNTS AND INDUCES IMMUNOMODULATION IN IMMUNE THROMBOCYTOPENIA (ITP)

JW Semple¹, R Aslam², E Speck², J Rebetz¹, R Kapur¹

¹Lund University, Lund, Sweden ²St. Michael's Hospital, Toronto, Canada

Background: ITP is an autoimmune bleeding disorder in which autoantibodies and/or autoreactive T cells target the destruction of platelets and megakaryocytes in the spleen and bone marrow. Several therapeutic options e.g. corticosteroids, intravenous immunoglobulins (IVIg), Rituximab and splenectomy are available for patients but inadequate efficacy, side effects and/or expense can make them undesirable. For the last 10years, TPO‐RA e.g. Romiplostim and Eltrombopag have made a substantial contribution to the treatment of ITP patient's refractory to first‐line treatments. Of interest, approximately 30% of patients that are tapered from TPO‐RA therapy show a sustained response (e.g. a stable higher platelet count than before treatment). The mechanism of how TPO‐RA induce these sustained responses is unknown.

Aims: To analyze the efficacy and immunomodulatory properties of a murine TPO‐RA (AMP4, Amgen) in a well‐established murine model of ITP that demonstrates both antibody‐ and T cell‐mediated thrombocytopenia (Chow L et al., Blood 2010).

Methods: Platelet glycoprotein (GP) IIIa (CD61) knockout (KO) mice were immunized with CD61⁺ platelets and ITP was initiated by the transfer of their splenocytes into mice with severe combined immunodeficiency (SCID). The SCID mice were treated with either placebo or TPO‐RA weekly and platelet counts and serum anti‐platelet antibodies were measured weekly.

Results: In an initial pilot dose escalation study, control naïve SCID mice treated with a single subcutaneous bolus of different concentrations of murine TPO‐RA (1, 10 and 20 ug/kg) had significantly higher platelet counts by 72h post infusion. In addition, compared with untreated mice, bone marrow histology revealed significantly increased numbers of megakaryocytes.Maximal platelet count increases were observed with the highest TPO‐RA dose and this dose was chosen to treat SCID mice suffering from ITP. When SCID mice were treated with weekly injections of TPO‐RA, platelet counts began to increase after 2weeks and were fully rescued to control levels after 3weeks post splenocyte transfer. Of interest, compared with non‐treated ITP mice, serum IgG anti‐platelet antibody production in the TPO‐treated mice was significantly reduced starting from two weeks post splenocyte infusion.

Summary/Conclusions: These results suggest that murine TPO‐RA is not only an efficacious therapy for murine ITP but also induces immunomodulation indicative of immunosuppression. Thus, this model may be able to elucidate the mechanism of how TPO‐RA's induced immunosuppression in patients with ITP.

AUTOANTIBODY MEDIATED CHANGES IN PLATELETS GLYCAN PATTERN: POTENTIAL IMPACT ON PLATELET FUNCTION AND LIFESPAN IN IMMUNE THROMBOCYTOPENIA

J Zlamal¹, I Marini¹, R Jouni¹, F Rigoni¹, T Bakchoul^1,2

¹Transfusion Medicine, Medical Faculty of Tübingen ²Centre for Clinical Transfusion Medicine ZKT GmbH, Tübingen, Germany

Background: Desialylation, the loss of sialic acid content on platelets (PLTs) glycoproteins (GPs) was recently identified to contribute in immune thrombocytopenia (ITP). However, the potential impact of autoantibodies (AAbs) on megakaryocyte sialylation remains unclear.

Aims: To investigate the effect of ITP AAbs on PLTs and megakaryocytes (MKs) sialylation and the subsequent impact on PLT survival.

Methods: AAbs from well‐characterised ITP patients induced GP‐modifications were tested using a lectin binding assay. After incubation of MKs or PLTs with ITP or control sera, glycan changes were analysed by flow cytometry (FC). To investigate the impact of desialylation on PLTs life‐span, the NOD/SCID mouse model was used.

Results: 112 ITP sera were investigated in this study. 35 (31%) sera induced a significant increase in RCA signal on PLT surface compared to control sera from healthy donors (RCA‐mean fold increase (RCA‐FI): 3.23, range: 1.76–13.61, P=0.0001). In addition, 23 (21%) sera caused higher ECL binding to test PLTs (ECL‐FI: 2.31, range: 1.54–5.7, P=0.0001). Injection of desialylating AAbs resulted in accelerated clearance of human PLTs from the circulation of the NOD/SCID mice which was significantly reduced by a specific neuraminidase inhibitor that prevents desialylation on the PLT surface (survival of human PLTs after 5h: 29%, range 22–40% vs. 48%, range 41–53%, P=0.014, respectively). Most interestingly, a subgroup of ITP sera induced desialylation on MKs surface. In particular, 8 out of 13 (62%) induced a significant increase in RCA signal on MK surface (mean RCA‐FI: 2.19, range: 1.15–3.66, P=0.01); and 10 out of 13 (77%) sera with PLT‐desialylating AAbs increase ECL binding (mean ECL‐FI: 1.29, range: 1.01–1.7, P=0.005).

Summary/Conclusions: Our findings suggest that ITP AAbs of different specificities are able to induce desialylation on PLTs as well as on MKs which seems to result in an impairment of function and contribute to PLT destruction in vivo.

AUTOIMMUNE HEMOLYTIC ANEMIA: SEROLOGICAL CHARACTERISTICS AND TRANSFUSION CHALLENGES

M Raos¹, D Pulanic^2,3, M Lukic¹, M Vinkovic¹, P Kilic³, G Tomac¹, I Vidovic¹, BG Cepulic¹

¹Clinical Department of Transfusion Medicine and Transplantation Biology ²Department of Internal Medicine, Division of Hematology, Clinical Hospital Centre Zagreb ³Medical School University of Zagreb, Zagreb, Croatia

Background: Autoimmune hemolytic anemia (AIHA) is a rare autoimmune disease characterised by hemolysis associated with the presence of immunoglobulins (IgG, IgM, or IgA) and/or components of complement system on red blood cells (RBCs), which is usually demonstrated by a positive direct antiglobulin test (DAT). Depending on the presence of an underlying disorder, AIHA can be subdivided into primary and secondary and, by the temperature at which autoantibodies bind optimally to RBCs, into warm antibody AIHA (wAIHA), mixed AIHA (including both warm IgG and cold IgM antibodies), cold agglutinin disease (CAD), Paroxysmal Cold Hemoglobinuria (PCH) and DAT negative AIHA. A frequent finding in immunohematology is the presence autoantibodies on RBCs without clinical symptoms of hemolysis that may later develop.

Aims: The aim of this study was to analyse serologic findings and transfusion support in patients with AIHA and also to analyse DAT positive patients without clinical symptoms.

Methods: We included data for all adult patients with AIHA and DAT positive patients without clinical symptoms diagnosed and/or treated at the University Hospital Centre (UHC) Zagreb, Croatia in the period between 2014 and 2018. The diagnosis of AIHA was defined by anemia with features of hemolysis (elevated bilirubin and/or elevated lactate dehydrogenase and/or low haptoglobin level) and a positive DAT.

Results: The data from 64 patients (52% women) meeting the inclusion criteria was analysed. The mean age at the time of AIHA was 63years (range 22–89years). The mean Hg level at diagnosis was 68.60g/L. DAT results were positive mostly with IgG+C3d (59%) or IgG (31%). Most patients had warm AIHA (70%). Other types of AIHA diagnosed were mixed AIHA (15%), CAD (11%), PCH (1.5%) and DAT negative AIHA (1.5%). In 6 cases alloantibodies were detected with autoantibodies in the patient's plasma. Patients were treated with corticosteroids as 1st line therapy and some with intravenous immunoglobulins (IvIG). In severe or refractory patients rituximab and/or splenectomy was applied. A total of 80% of patients were transfused at a mean hemoglobin level of 67.88g/L. During this period we detected 136 DAT positive patients without clinical symptoms.

Summary/Conclusions: Most patients from our study were diagnosed with warm type of AIHA, followed by mixed type AIHA and CAD. On the other hand, PCH and DAT negative AIHA were very rare, which is in concordance with relevant literature. Most patients were transfused despite therapy used, which is not desirable in patients with AIHA and should be better controlled, especially in moderate cases of anemias, where this is rarely necessary. A significant number of patients that were DAT positive without clinical symptoms may later develop AIHA and should be closely monitored.

AUTOIMMUNE HAEMOLYTIC ANAEMIA: A SURVEY OF DIAGNOSTIC AND MANAGEMENT PRACTICE IN ENGLAND

M Horan¹, A Charlton¹, T Bullock², E Massey³, S Allard⁴, A Hill⁵, S Stanworth⁶, Q Hill⁵

¹Haematology, NHS Blood and Transplant, Newcastle upon Tyne ²Red Cell Immunohaematology ³Haematology, NHS Blood and Transplant, Bristol ⁴Haematology, NHS Blood and Transplant, London ⁵Haematology, The Leeds Teaching Hospitals NHS Trust, Leeds ⁶Haematology, NHS Blood and Transplant, Oxford, United Kingdom

Background: Autoimmune haemolytic anaemia (AIHA) is a decompensated acquired haemolysis caused by the host's immune system acting against its own red cell antigens. AIHA is a rare disorder and although British Society of Haematology (BSH) guidelines for diagnosis and treatment were published in February 2017, there is little evidence for clinical practice in the United Kingdom.

Aims: To investigate the approach to diagnosis, investigation and management of patients with AIHA in English NHS Trusts.

Methods: A survey of diagnostic and management practice was designed, piloted and disseminated to clinical transfusion leads in all English acute NHS trusts from November 2017 to March 2018. Completion was by a consultant haematologist treating patients with AIHA but a response that represented a departmental consensus was encouraged.

Results: Responses represented 42% (58/137) of English acute trusts. Median number of adults with AIHA diagnosed annually was 4–6. In the preceding 5years, 31% (18/58) recalled at least one patient who had died due to AIHA. Although 7% (4/57) undertook a bone marrow biopsy in all patients, 93% required additional features, mainly: neoplasia, age over 60 or being treatment‐refractory.For patients with suspected drug‐induced immune haemolysis, 59% (34/58) would not organise confirmatory tests, either because it was considered unnecessary (29/34), or because clinicians were unsure how to access tests (5/34). When determining AIHA subtype, 29% (17/58) indicated there were no circumstances in which they would undertake cold antibody testing (antibody titre and/or thermal amplitude), with 12 considering this unnecessary and 5 unsure how to access tests.

In 4 clinical scenarios of patients with AIHA and DAT positive to C3d ± IgG ± cold associated symptoms, up to 87% (47/54) of respondents would not test for cold antibodies. For first line treatment of primary warm AIHA, mean duration of prednisolone 1mg/kg given before judging the patient refractory and reducing the dose was 3.5weeks (SD 1.70, range 1–19weeks). Second line treatment of choice was rituximab for 82% (45/55) of respondents and splenectomy for 5%. Intravenous immunoglobulin and splenectomy were the most cited rescue therapies. For primary cold haemagglutinin disease (CHAD), first line treatment was rituximab‐based for 88% (49/56) but single agent steroid for 9%.

We also explored the potential for future audit and research. 64% (37/58) of respondents were able to identify patients with AIHA who previously required transfusion. 96% (55/57) of respondents would consider supporting a registry of patients with AIHA requiring transfusion. The key questions that respondents thought a registry should address were: morbidity and mortality, treatment response, and differences in the diagnosis and treatment of AIHA subtypes. There was uncertainty over access to cold and drug‐induced antibody tests and clinicians do not always conduct BSH‐recommended cold antibody tests for AIHA with C3d positive DAT. Initial treatment of primary warm AIHA and CHAD broadly matched BSH guidelines although 44% (25/57) would continue prednisolone at 1mg/kg beyond the recommended 21days before starting a taper, with greater toxicity risk.

Summary/Conclusions: The findings support the need for a range of research initiatives, including creation of an AIHA registry.

AN UPDATE ON PATIENT BLOOD MANAGEMENT

K Pavenski

Laboratory Medicine, St. Michael's Hospital, Toronto, Canada

Preoperative anemia is common and is associated with adverse outcomes in the peri‐operative period. Preoperative anemia also increases the risk of allogeneic blood transfusions, which may lead to increased perioperative mortality, increased hospital length of stay and infections. Diagnosis and treatment of anemia is one of the tenets of patient blood management (PBM), along with reduction in unnecessary transfusions and diagnostic phlebotomy, as well as use of hemostatic agents to reduce bleeding among many others. Effective PBM is multi‐disciplinary, multi‐modal, timely, individualized and patient‐centered. Early referral to PBM and multi‐modal PBM interventions are associated with greater improvement in pre‐operative hemoglobin. PBM has been shown to reduce transfusions and cost, while system‐wide, multi‐modal programs may also be associated with improvement in mortality. Using examples from our local research and practice, I will discuss three aspects of PBM. Iron and erythropoiesis stimulating agents (ESA) are effective, safe and used extensively in management of pre‐operative anemia. Previous studies have questioned whether ESA leads to increased risk of thrombosis, however, recent systematic reviews do not support these concerns. Another PBM approach is to reduce bleeding during surgery by using hemostatic agents such as tranexamic acid (TXA). TXA reduces transfusion requirements in knee and hip arthroplasty, and is safe, widely available and relatively cheap. TXA is effective in both anemic and non‐anemic patients, making it an attractive universal PBM strategy. Finally, recommendations and evidence‐based guidelines on PBM exist, including the most recent international guidelines developed by the PBM International Consensus Conference. However, knowledge translation in PBM has been a problem and a number of barriers to its implementation have been identified. These include perceived or actual lack of expertise, time, and resources, as well as lack of physician and patient engagement. One way to address patient engagement is education through character driven animation and we are currently trying this approach.

LOW VS. HIGH HEMOGLOBIN TRIGGER FOR TRANSFUSION IN VASCULAR SURGERY (TV): A RANDOMIZED CLINICAL FEASIBILITY TRIAL (THE TV TRIAL)

A Møller¹, H Nielsen², J Wetterslev³, O Pedersen⁴, D Hellemann⁵, P Winkel³, K Marcussen⁵, B Ramsing⁶, A Mortensen⁵, J Jacobsen³, S Shahidi⁷

¹Anaesthesia and Intensive Care, Slagelse Hospital, Slagelse ²Sanos Clinic, Herlev ³Copenhagen Trial Unit, Rigshospitalet, Copenhagen ⁴Clinical Immunology, Naestved Hospital, Naestved ⁵Anesthesia and Intensive Care ⁶Anesthesia and Intensive Care ⁷General and Vascular Surgery, Slagelse Hospital, Slagelse, Denmark

Background: Current guidelines advocate to limit red‐cell transfusion during surgery, but the feasibility and safety of such strategy remains unclear as the majority of evidence is based on postoperative stable patients.

Aims: We assessed the effects of a protocol aiming to restrict red‐cell transfusion during elective vascular surgery.

Methods: Fifty‐eight patients scheduled for lower limb‐bypass or open surgery of abdominal aortic aneurysm were randomized to a low‐trigger (hemoglobin<8.0g/dl, 5mmol/L) vs. high‐trigger (hemoglobin<9.7g/dl, 6mmol/L) for red‐cell transfusion throughout hospitalization. Intraoperative change in cerebral‐ and muscle tissue oxygenation was assessed by near‐infrared spectroscopy. We used a nationwide registry to collect data on death and major cardiovascular events, which encompassed (1) severe adverse transfusion reaction, (2) acute myocardial infarction, (3) stroke, (4) new‐onset renal replacement therapy, (5) vascular reoperation, and (6) amputation of the lower limb.

Results: The primary outcome, mean hemoglobin within 15days of surgery, was significantly lower in the low‐trigger group: 9.46g/dl vs. 10.33g/dl in the high‐trigger group (mean difference 0.87g/dl; P=0.022, longitudinal analysis) as were units of red‐cells transfused (median [interquartile range (IQR)] 1 [0–2] vs. 3 [2–6]; P=0.0015). While the cerebral desaturation load increased in the low‐trigger group (median [IQR] 421min*% [42–888] vs. 127 [11–331], P=0.0036), muscle oxygenation was similar. The low‐trigger associated to a higher rate of death or major cardiovascular events: 19/29 vs. 8/29 (hazard ratio 3.18; P=0.006) and fewer days alive outside hospital within 90days (median [IQR] 76 [67–82] vs. 82 [76–84] days, P=0.049).

Summary/Conclusions: A perioperative protocol restricting red‐cell transfusion successfully separated hemoglobin levels, red‐cell units transfused and intraoperative cerebral tissue oxygenation. Exploratory outcomes suggested potential harm with the low‐trigger and warrants further trials in vascular surgery before such strategy is universally adopted.

REDUCTION OF RED CELL UTILIZATION BY A HB‐TRIGGERED SINGLE‐UNIT TRANSFUSION POLICY IN AN INPATIENT HEMATO‐ONCOLOGICAL PATIENT POPULATION: A RETROSPECTIVE SINGLE‐CENTRE ANALYSIS

H de Lil¹, J Oomen¹, C Eijsink², N Blijlevens¹, M Hoeks¹, D Evers¹

¹Hematology ²Laboratory medicine, Radboudumc laboratory for diagnostics, section transfusion medicine, Radboudumc, Nijmegen, Netherlands

Background: Controlled non‐hemato‐oncological studies have consistently demonstrated a single‐unit red blood cell (RBC) transfusion policy as well as a stringent hemoglobin (Hb) RBC transfusion threshold to be safe and reduce blood product utilization. Yet, it is unclear whether these conclusions also apply to the hemato‐oncological patient population.

Aims: To quantify reduction of RBC blood product utilization by the introduction of a restrictive single‐unit Hb‐triggered RBC transfusion policy among the inpatient hemato‐oncological population.

Methods: Under the liberal transfusion protocol, applied up till November 1, 2017, standard double‐unit RBC transfusion was indicated with a Hb threshold≤8.1g/dl and/or anemia‐related symptoms. Following this date, the restrictive transfusion protocol was introduced involving a lowering of threshold to 7.3g/dl and single‐unit transfusion. For patients with an ASA‐score of II‐III and IV, a Hb threshold of respectively≤8.1g/dl and≤9.7g/dl applied.

We evaluated RBC blood product utilization over a 6month period starting December 1, 2016 (liberal protocol) and December 1, 2017 (restrictive protocol) in all hemato‐oncological patients admitted for chemotherapeutic treatment including hematopoietic stem cell transplantation (HSCT) with an expected duration of neutropenia of≥7days.

Analysis of categorical and continuous data was performed using the chi‐square and Mann‐Whitney test, respectively.

Results: During both observational periods, 137 patients were admitted who in total received 164 therapy cycles, including 56 acute myeloid leukemia (AML) induction cycles and 69 autologous HSCTs.

Distribution of indications of admittance, median age, duration of hospitalization and duration of neutropenia did not differ between both periods.

During the restrictive period, in 303/402 (75.4%) of transfusions the assigned Hb trigger was adhered to. The percentage of single‐unit transfusion episodes increased from 29/112 (29.0%) to 81/111(73.0%) with the introduction of the restrictive protocol.

Overall, RBC blood product utilization per admittance did not reduce under the restrictive protocol (cumulative number of transfused RBC units 4.0 (interquartile range (IQR) 2.0–8.0) during the liberal versus 2.5 (IQR 0.0–9.0) during the restrictive period (P=0.36)). However, RBC blood product utilization per neutropenic day demonstrated a trend towards reduction: 0.25 (IQR 0.11–0.33) versus 0.15 (IQR 0.00–0.32) units per day during the liberal versus restrictive period, respectively (P=0.06). This reduction was mainly attributed to autologous HSCTs during which RBC blood product utilization decreased from 2.0 (IQR 0.0–2.0) to 0.0 (IQR 0.0–1.5) units (P=0.06), corresponding to a reduction from 0.13 (IQR 0.00–0.21) to 0.0 (IQR 0.0–0.13) (P=0.01) units per neutropenic day. Moreover, 14/38 (36.8%) patients during the liberal versus 21/31 (67.7%) during the restrictive period did not require RBC transfusion during admittance. Consequently, stringent Hb thresholds as compared to single‐unit transfusions seem to more strongly impact RBC blood product utilization.

Summary/Conclusions: A Hb‐triggered single‐unit transfusion policy results in a strong reduction of RBC blood product utilization in the setting of autologous HSCT. No utilization reduction was observed among other hemato‐oncological inpatient populations receiving intensive chemotherapy. Further improvement of protocol adherence rates could potentially increase the benefit of this blood saving strategy.

ASSESSMENT OF HB CONTENT OF PACKED RED CELLS (PRBC): IS IT TIME TO LABEL EACH UNIT WITH HB CONTENT?

R Jain¹, N Marwaha², S Sachdev²

¹Transfusion Medicine, Aiims, New Delhi ²Transfusion Medicine, Pgimer, Chandigarh, India

Background: In the current era of evidence based medicine and individualized care of patients, RBC transfusion continues to be administered on the basis of conventional wisdom and the notion of an average benefit per unit. The existing blood transfusion practice based on the “number of units transfused” ignores the fact that the total Hb varies markedly among the individual RBC units.

Aims: The present study was aimed at estimating the Hb content in packed red cell unit prepared by three different protocols from 350ml and 450ml whole blood collection in three types of blood donors: replacement blood donor (RD), first time voluntary donor (FTVD) and regular voluntary blood donor (RTVD).

Methods: A total of 900 prospective blood donors were included in this study. Three hundred whole blood collections were performed in each of the three groups of donors (RD, FTVD, and RTVD). Within each group 100 collections were done in Double 350ml, Triple 450ml and Quadruple 450ml blood bags respectively. A pre‐donation venous sample was drawn from sample collection pouch for analysis in hematology analyzer as reference method for Hb concentration of donor. The Hb content of packed red cell units were estimated after collection of representative sample from the blood unit. Volume of PRBC unit was estimated by the formula of weight of blood in PRBC divided by specific gravity. The Hb content in unit was estimated by the formula: Hb content in unit=Hb value of the PRBC unit (g/dl)×volume of PRBC unit (dl).

Results: In this study the Hb concentration (g/dl) was comparable among three types of blood donors except that RTVD had lower Hb values when compared to RD (P=0.007). Hemoglobin concentration of PRBC ranged from 14.2–29.6g/dl; mean Hb was 21.02±2.90g/dl. Net Hb content of PRBC bag was lower in PRBC prepared from RD as compared to FTVD (P=0.0001) and RTVD (P=0.008). The Hb content of PRBC units prepared from 450ml collection ranged from 35.19–87.36g and from 350ml collection ranged from 30.77–65.78g. We observed a wide range of net Hb content in the PRBC units and the correlation coefficient showed the strongest association of net Hb content of the PRBC unit with the overall volume of PRBC (r=0.730, P=0.0001).Higher volume PRBCs have more Hb content. Volume of PRBC bags in the study ranged from 155ml to 370ml (including both 350 and 450ml collections).

Summary/Conclusions: The present study shows that labelling Hb content of the PRBC unit help in better inventory management for patients. The Hb content may help in decision making for release of units for paediatric/low weight versus adults/higher weight patients. Adopting a policy of optimizing dosage of RBC transfusion could have the potential to significantly improve RBC utilization and decrease patient exposure to allogenic blood. This would help further in the clinical transfusion practices based on evidence.

EVALUATION OF CLINICAL PRACTICES OF RED BLOOD CELL TRANSFUSIONS IN A TERTIARY CARE ONCOLOGY CENTRE

NV More, P Desai, S Rajadhyaksha, A Navkudkar, N Deshpande

Transfusion Medicine, Tata Memorial Centre, Homi Bhabha National Institute, Mumbai, India

Background: Red Blood Cell (RBC) transfusion is an important medical therapy benefiting the patient in a wide spectrum of clinical setting. Critically ill Intensive Care Unit patients in particular, as well as medical and hemato‐oncology patients, are among the largest group of the user of RBC. Periodic review of blood components usage is essential to assess the blood utilization pattern in any hospital or health care set up. Our institute is a 639 bedded tertiary care oncology centre with approximately 18,000 to 20,000 RBC transfusions annually. These transfusions are required in various stages of patient treatment like chemotherapy, radiotherapy, surgical and palliative care and there are established guidelines by the institute to be followed by clinicians.

Aims: To study clinical practices of RBC transfusions based on indications and to evaluate appropriateness of RBC utilization practices at the institute.

Methods: This was a prospective observational study, started after approval from Institutional Ethics Committee. Total of 4413 RBC transfusion events in adult patients over a period of four months were included and analyzed as per institutional guidelines for their appropriateness. Details of transfusion events in form of pre transfusion hemoglobin, indication of transfusion, type of request, number of unit requested and issued, time of issue, site of transfusion and adverse reactions, etc were obtained from Department of Transfusion Medicine records. Overall statistical analysis was descriptive using SPSS software. Chi‐square test in cross tables was applied to see the relationship between different variables and considered significant if p‐value was<0.05.

Results: Total 4413 RBC transfusion events for 2012 patients were analyzed. There were 1877 (43%) events in 628 patients of medical oncology and 2536 (57%) in 1384 patients of surgical oncology. Maximum transfusions were received by patients in age group of 41 to 60years (47%). Total 83% of transfusion events were appropriate as per institutional guidelines. All transfusions administered in operation theatre were found to be appropriate with p value<0.05. Inappropriateness was more 53%(396/735) and significant in daycare setup (P<0.05). Anemia was the most common indication of RBC transfusion observed in 90% of events (3971/4413). Total 62% RBC transfusions were given as planned and 38% as urgent transfusions. Most common adverse transfusion event observed was allergic reaction in 0.3% of total transfusion reactions.

Summary/Conclusions: Clinical practice of RBC transfusions in our hospital was largely found to be appropriate and rational with adherence to institutional guidelines. Blood utilization audits should be conducted regularly by transfusion services and results should be discussed with clinician for ensuring judicious use of the scarce resource. The concept of Transfusion Safety Officer (TSO) can be introduced for better coordination between clinicians and blood transfusion services to improve practices.

APPROACHES TO CONTROL INFECTIONS IN VARIOUS SETTINGS

CM Nuebling

Paul‐Ehrlich‐Institut, Langen, Germany

On a global scale, blood services are quite diverse in regard to aspects like organisational structure, regulatory background, donor populations, donation rates or pathogen epidemiology. The World Health Organization (WHO) recognizes blood and blood products as essential medicines and provides guidance to member states for various aspects like blood regulation, best practices in blood collection and transfusion, or screening parameters. More recently a WHO guideline on residual risk of transfusion associated infections has been established which may facilitate decision‐making for the most appropriate screening algorithms. It emphasizes the need for regional evaluation of screening assays and regulatory control of blood‐associated IVDs.

DETECTION OF BABESIA IN US BLOOD DONORS

J Sunga¹, K Chugh¹, S Bakkour², J Cruz³, Y Erickson⁴, J Gottschall⁵, M Janzen⁶, B Sachais⁷, T Straus⁸, J Thebo⁹, L Pate¹⁰

¹Roche Molecular Systems, Pleasanton ²Vitalant Research Institute, San Francisco ³Versiti Indiana, Indianapolis ⁴Mississippi Valley Regional Blood Center, Davenport ⁵Versiti Wisconsin, Milwaukee ⁶Innovative Blood Resources, St. Paul ⁷Blood Bank of Delmarva, Newark ⁸Community Blood Center, Appleton ⁹Central Pennsylvania Alliance Laboratory, York ¹⁰Medical and Scientific Affairs, Roche Molecular Systems, Pleasanton, United States

Background: Babesia, a protozoan parasite that infects red blood cells, is a leading infectious cause of mortality in U.S. transfusion recipients. Babesia is usually transmitted through the bite of an infected tick but may be transfusion transmitted (TT) or transmitted from mother to child during pregnancy. Babesiosis is a world‐wide disease; the ticks that carry Babesia have a global distribution. Babesiosis has been reported throughout Europe and in Canada, Korea, India, and Japan. Prospective testing of blood donations in endemic areas of the U.S. revealed 0.38% of donors were positive for Babesia DNA or antibodies (Moritz, NEJM, 2016)

Aims: To report results of ongoing Babesia clinical trial

‐ To explain significance of Babesia as a TT infection

Methods: In cobas^® Babesia for use on the cobas^® 6800/8800 Systems, is a qualitative polymerase chain reaction nucleic acid amplification test, developed to detect in whole blood (WB) donor samples the 4 Babesia species that cause human disease: B. microti, B. duncani, B. divergens, and B. venatorum. Testing began in October 2017 under a U.S. FDA‐approved Investigational New Drug Application. WB was collected into a proprietary medium that lysed red blood cells and stabilized Babesia RNA and DNA. Donations were collected in states with high, low, and no Babesia endemicity and screened as individual blood donor (IDT) samples. Reactive index donations were retested in simulated minipools of 6 (MP6), plus 3 IDT replicates with cobas^® Babesia. Reactive index donations were also tested with 2 validated alternate Babesia NAT and for B. microti IgM and IgG antibodies. Donors with reactive results were invited to enroll in a follow‐up study to test for additional evidence of infection.

Results: To date, 256,802 valid donations have been screened with cobas^® Babesia, and 15 (0.006%) were reactive. 13 of 15 (87%) initially‐reactive donations were confirmed to be positive for Babesia with a positive alternate NAT or serology result. 1 of 13 (8%) confirmed‐positive donations was collected in a state with low Babesia endemicity (Pennsylvania), and 1 (8%) was collected in a state where Babesia is not considered endemic (Iowa). 11 of 13 (85%) confirmed positive donations were collected in states with high endemicity. 8 of 13 (62%) confirmed Babesia‐positive donations were detected in late fall or winter. All 13 (100%) confirmed Babesia‐positive donations were reactive in MP6. Serology results are available for 12 of 13 confirmed‐positive donations: At index, 6 of 12 confirmed Babesia‐positive donations were only IgG‐positive, while none were only IgM‐positive; 3 were positive for both IgG and IgM. 3 of the confirmed‐Babesia positive donations were negative for both IgG and IgM antibodies. cobas^® Babesia showed an overall specificity of 99.999% (256,787/256,789; 95% Exact CI: 99.997%>100%).

Summary/Conclusions: The cobas^® Babesia test successfully identified 13 Babesia‐positive donations, including 3 confirmed‐positive donations with no IgM or IgG reactivity. 2 donations were collected in states considered low‐or non‐endemic for Babesia. 8 confirmed‐positive donations were collected outside of the summer Babesia season, when most clinical cases occur. Screening with cobas^® Babesia continues in several laboratories. cobas^® Babesia is not FDA licensed or available commercially.

PREVALENCE OF BABESIA MICROTI IN CANADIAN BLOOD DONORS

S O'Brien¹, S Stramer², M Proctor², L Tonnetti², V Bres³, J Linnen³, F Bernier⁴, G Delage⁴, T Gaziano⁵, Y Gregoire⁴, J Labrie⁴, M Bigham⁵, G Hawes⁵, V Scalia⁵, M Fearon⁵, S Drews⁶

¹Epidemiology and Surveillance, Canadian Blood Services, Ottawa, Canada ²American Red Cross, Gaithersburg ³Grifols Diagnostic Solutions Inc, San Diego, United States ⁴Hema‐Quebec, Saint‐Laurent ⁵Canadian Blood Services, Ottawa, Canada ⁶Medical Microbiology, Canadian Blood Services, Ottawa, Canada

Background: Babesiosis in humans is caused by the erythrocytic protozoan parasite, Babesia microti which is transmitted by tick bites, but is also transfusion transmitted. Although frequently asymptomatic or presenting with flu‐like symptoms in a normal host, if immunocompromised infection can lead to severe complications and death. B. microti is endemic in the North Eastern/Upper Midwest United States where partial testing of donations has been implemented. In Canada, a 2013 study of ˜14,000 donors did not identify any B. microti antibody‐positive samples, suggesting low risk at that time, but risk should be monitored.

Aims: To evaluate the prevalence of B. microti‐positive donations in potentially at‐risk areas in Canada.

Methods: Between July and November 2018, 50,752 blood donor samples were selected from sites near the US border. Minipools were tested for B. microti nucleic acid by Transcription Mediated Amplification (TMA) using the Procleix^®Babesia Assay on the Panther^® system with individual testing on reactive pools. Reactive donations were also tested by B. microti‐specific: American Red Cross (ARC) IgG immunofluorescence assay [IFA] and IMUGEN IFA/PCR. A subset of 14,758 TMA‐negative samples, primarily from the province of Manitoba and eastwards to Nova Scotia, were tested for B. microti antibody using the ARC IFA and if positive, the IMUGEN IFA/PCR. Donor age, sex, donation status, residential location and collection site location were recorded. Donors who tested reactive/positive were informed, deferred and asked about risk factors (possible tick exposure and travel within Canada, the USA and elsewhere, history of symptoms) and a follow‐up sample was requested for supplemental testing (TMA, ARC IFA). Reactive donations were removed from inventory.

Results: The 50,752 donor samples were proportional to collections in target geographic regions. Age group, sex and donation status were also similar to the donor base in the collection areas. One sample from Winnipeg, Manitoba was TMA reactive and antibody positive on supplementary testing. The donor did not remember symptoms or spending time in wooded areas. He visited the city of Fargo, North Dakota, USA. The subset of 14,758 samples tested for antibody were also proportional to collections in the targeted areas. Four antibody‐positive samples were identified from mid‐September to October 1, all in south western Ontario near Lake Erie. None were TMA reactive. Three were interviewed and none remembered any symptoms, any likely tick exposure, or relevant travel within Canada or the USA.

Summary/Conclusions: This is the largest B. microti prevalence study in Canada. The results indicate very low prevalence with only 1 TMA‐confirmed‐positive donation of 50,752 tested. The donor was from the only region in Canada where one autochthonous human case has been reported and active tick surveillance identified B. microti positive tick populations. Seropositive donations in south western Ontario may suggest low prevalence in that region, but interpretation is less certain due to lack of corroborating supplementary results or case history. Given the close proximity to the US border, forgotten US travel should not be ruled out.

SEROEPIDEMIOLOGY OF TOXOPLASMA GONDII IN BLOOD DONORS IN PORTUGAL

F Teixeira Rodrigues¹, A Sousa², M Escoval², J Condenço³, I Pires⁴, J Dubey⁵, L Cardoso^1,6, A Lopes^1,6

¹Animal and Veterinary Research Centre (CECAV), University of Trás‐os‐Montes and Alto Douro (UTAD), Vila Real ²Blood and Transplantation Centre of Lisbon, Portuguese Institute of Blood and Transplantation (IPST), Lisbon ³Blood and Transplantation Centre of Oporto, IPST, Oporto ⁴Blood and Transplantation Centre of Coimbra, IPST, Coimbra, Portugal ⁵Animal Parasitic Diseases Laboratory, Beltsville Agricultural Research Centre, Agricultural Research Service, United States Department of Agriculture, Beltsville, MD, United States ⁶Department of Veterinary Sciences, UTAD, Vila Real, Portugal

Background: The protozoan parasite Toxoplasma gondii is prevalent in animals and humans worldwide. Wild and domestic felids are the definitive hosts, and homoeothermic animals serve as the intermediate ones. After primary infection, the parasite persists lifelong within latent tissue cysts. Transmission is by ingestion of undercooked or raw meat infected with cysts, by ingestion of food or water contaminated with oocysts, or transplacentally. However, it can also be acquired by blood transfusion and organ transplantation. Toxoplasmosis can be a severe disease in immunosuppressed people and neonates whose mothers have acquired primary infection during pregnancy.

Aims: There is no information about the specific epidemiology of T. gondii infection in blood donors in Portugal. Therefore, we sought to determine the seroprevalence of T. gondii and associated risk factors in the population of blood donors in Portugal.

Methods: Between September 2015 and July 2017, 520 blood donors who attended the Portuguese Blood and Transplantation Institute blood banks located in Oporto, Coimbra and Lisbon, and also at regional blood collection meetings, were invited to participate in the study. A written informed consent was obtained and a questionnaire about socio‐demographic and behavioural variables was answered. Sera were assessed for IgG antibodies to T. gondii by a modified agglutination test (MAT) commercial kit (Toxo‐Screen DA^® bioMérieux, Lyon, France).

Results: Of the 520 blood donors (mean age 38.55±11.14; range 18–65years old), 38.1% were positive for antibodies to T. gondii. When questioned about toxoplasmosis, almost half the blood donors did not have any knowledge about the disease. The Centre of Portugal had the highest seroprevalence (55.1%) followed by the North (37.2%) and the South (25.3%). Blood donors living in rural areas had a significantly higher seroprevalence (P=0.001) than those living in urban areas. Seroprevalence increased with age, with the highest seroprevalence (60.2%) found in the age group of 46–55years old (multiple logistic regression [MLR]: OR=7.68; CI: 4.08–14.46; P<0.001), and decreased with educational level (P<0.001). Engaging in soil‐related activities (gardening or agriculture) was significantly related to T. gondii seropositivity (P=0.02). Regarding water consumption, untreated sources (even though including mineral and tap water) was confirmed as a risk factor (MLR: OR=2.72; CI: 1.27–5.86; P=0.001). Other behavioural and eating characteristics, including cats in the household, eating raw or undercooked meat, processed pork products, or not washing raw fruit and vegetables before eating, were not associated with T. gondii infection.

Summary/Conclusions: The risk of T. gondii transmission through blood transfusion is low, and serologic testing of antibodies, with exclusion of blood donors, appears not to be feasible. Immunosuppressed individuals, organ transplant patients and pregnant women, should receive T. gondii antibody‐negative blood components for transfusion. This study explored the epidemiology of T. gondii in Portugal thus providing useful information on the seroprevalence and potential risk factors for T. gondii transmission. Information regarding toxoplasmosis and its prevention could be promoted by medical and public health authorities among blood donors, and also the general population, when addressing policies, and designing screening programs, for monitoring and controlling infection and disease in Portugal.

WHO IS SYPHILIS TESTING EXCLUDING?

C Reynolds¹, C Pearson², K Davison², S Brailsford¹

¹NHSBT/PHE Epidemiology Unit, NHS Blood and Transplant ²NHSBT/PHE Epidemiology Unit, Public Health England, London, United Kingdom

Background: Screening for treponemal antibodies to detect syphilis in blood donors has been in place in England since the 1940s. There have been no reported syphilis transfusion transmissions in England since records began in part due to sensitivity of the organism to cold storage. Since we have specific tests in place for other sexually transmitted infections such as HIV and hepatitis B virus (HBV), the utility of syphilis screening is often questioned. However, it may be a useful proxy for higher risk behaviours particularly following shortening of deferrals for higher risk sexual behaviours from 12 to 3months in November 2017 and against a background of increasing infectious syphilis in the general population.

Aims: Here we describe the epidemiology of recently‐acquired syphilis in blood donors in England compared with HIV and acute HBV infection between 2009 and 2018.

Methods: Monthly donation testing results are collected from the NHS Blood and Transplant (NHSBT) screening centres and reference laboratory. The demographics, possible sources of infection, and compliance to donor selection in confirmed positive donors are collected by proforma at post‐test discussion with the NHSBT clinical team. Recent syphilis is classified as IgM positive and/or recent history including a negative donation within 12months for regular donors.

Results: Between 2009 and 2018 there were 153 recent syphilis cases, 121 HIV and 32 acute HBV infections identified by donation screening. Recent syphilis rates per 100,000 donations increased from 0.66 to 1.64 whereas HIV decreased from 1.04 to 0.19 with less than 5 positive donations in 2018. Acute HBV rates rose slightly from 0.28 to 0.38 in 2018. Males outweighed females accounting for 71.9%, 63.6% and 59.4% of cases of recent syphilis, HIV and acute HBV respectively. Nearly a quarter of cases of recent syphilis and HIV were seen in donors below 25years old. Of the male donors with recent syphilis, 58.2% reported sex between men and women (SBMW), 19.1% sex between men (SBM) and 22.7% did not report a risk. This contrasted with HIV where 45.5% of male donors reported SBM, just 2.6% not reporting a risk. Overall 19, 32 and 3 males with recent syphilis, HIV or acute HBV respectively were non‐compliant to the SBM deferral in place at the time of donation. In 2018, 26 donors with recent syphilis aged 23–65years (median 36years) were excluded from the donor pool, including 3 non‐compliant to the SBM deferral. There were fewer than 5 HIV cases identified in 2018, all over 40years old, all compliant, reporting SBMW. Of the 6 HBV acute cases in 2018, 5 were male, all but one in the 45 and over age‐group.

Summary/Conclusions: Over the 10year period demographics of recent syphilis cases appeared similar to HIV with highest rates in young males, albeit lower proportions reporting SBM. Following the switch to a 3month deferral, HIV case detection continued at low level, while syphilis screening continued to exclude higher numbers spanning all age‐groups, potentially at risk of other sexually transmitted infections, including non‐compliant donors.

CURRENT OPINION IN DONOR HEALTH – PERSONALISED DONOR CARE

C Erikstrup¹, O Pedersen², H Ullum³

¹Department of Clinical Immunology, Aarhus University Hospital, Aarhus ²Department of Clinical Immunology, Naestved Hospital, Naestved ³Department of Clinical Immunology, Copenhagen University Hospital, Copenhagen, Denmark

Background: Globally, an estimated 113 million blood donations are given annually. In the blood service we are obliged to monitor donor health and ensure that blood donation is safe. In recent years, large‐scale blood donor cohort studies in several countries have increased our knowledge on health effects of blood donation. Health concerns relate both to immediate side effects like fainting and to possible long‐term health issues related to repeated blood or plasma donation. The studies have provided us with data that can now help us introduce an evidence‐based individualised donor care ‐ a parallel to personalised medicine.

Individualised donor care in the management of iron depletion

Studies have shown that a large percentage of our frequent whole blood donors, especially young women, are iron depleted. Iron depletion is a strong predictor of deferral for low haemoglobin but has also been associated with e.g. restless legs syndrome and lower birth weight in children of frequent donors. The risk of iron deficiency can be mitigated by ferritin‐guided prolongation of interdonation intervals or by iron supplementation. Prolongation of interdonation intervals can challenge our inventories. Iron supplementation, on the other hand, may give gastrointestinal side effects and other effects have been proposed as well, e.g. the masking of malignant disease and increased iron availability with subsequent risk of infection. In a large study we found that iron supplementation is not associated with increased risk of infection. What is the optimal balance between iron supplementation and prolongation of interdonation intervals?

A growing number of blood services have implemented various flavours of iron management regimens generating more results. Moreover, genetic studies in e.g. the UK, US, Holland, and Denmark can help us to find donors at high risk of iron depletion or low haemoglobin. We can use all these data in a Big Data approach in the pursuit of an individualised risk assessment model.

Other risks for blood donors

The presentation will also cover other risks associated with donation. New studies identify predictors of fainting after blood donation and also new interventions to prevent fainting. The global demand for plasma derived medicinal products has increased severalfold the last 10years. Plasma donors are bled up to 104 times per year in the US. Very little is known about the health effects of frequent plasma donation. We know that immunoglobulin levels decrease with frequent donation but how does this affect health?

Summary/Conclusions: The precautionary principle mitigates risk through early intervention prior to evidence. We tolerate next to no risk of transfusion‐transmitted infectious diseases. The health of the blood donors, however, has not been protected similarly. We owe to our whole blood and plasma donors to investigate health effects of blood donation and ensure their safety. While the first attempts may not be perfect, we now have the tools to construct models for individualised donor care.

PILOT OF DONOR ADVERSE EVENT SEVERITY GRADING TOOL IN A LARGE BLOOD COLLECTION CENTER

M Townsend, M Bravo, H Kamel

Medical Affairs, National Office, Vitalant, Scottsdale, United States

Background: In 2014, the ISBT, AABB, IHN and EBA jointly issued the Standard for Surveillance of Complications Related to Blood Donation which categorized Donor Adverse Events (DAE) into 6 categories (16 subcategories) defined by specific criteria. Severity and Imputability were briefly described but were optional. Subsequent validation of these categories demonstrated consistency in categorizing reactions, but wide variation in assignment of Severity. In 2018, with international input, the AABB Donor Biovigilance Committee developed a Severity Grading Tool (SGT) using a recognized medical adverse event grading system in which neutral Grades replace subjective terms (mild, moderate, severe).

Aims: A large US blood collection establishment (BCE) applied the draft SGT to assess its use in real cases of DAE.

Methods: We performed retrospective analysis of all allogeneic and apheresis needle‐in donations between 1/1/2017 to 9/30/2018. Severity grading was assigned based on criteria defined by the SGT. Database review of DAE was performed, and each event was assigned a Grade based on the type of outside medical care (OMC), and on specific key search terms. Search terms for OMC included Emergency Room, Emergency Medical Response, Urgent Care, Healthcare Professional, and Hospital Admission. Additional specific key search terms included fracture, concussion, laceration, dental injury, surgery, and hospitalization. Since duration and activities of daily living (ADL) limitations were not captured in our DAE database, cases in our DAE claims’ database were reviewed. Case files of events classified as Grade 2 or higher were individually evaluated by a physician for grading accuracy.

Results: In 1,511,758 needle‐in collections, 31,320 DAE were graded for severity. The majority (16,143, 51.5%) were vasovagal reactions (VVR), followed by 8,570 apheresis‐related (27.4%), 6,572 needle‐related (21.0%) and 35 allergic (0.1%) events. The majority of DAE were Grade 1 accounting for 98.6% of all DAE, followed by Grade 2 (1.2%), and Grade 3 (0.1%). There were 2 Grade 4 and no Grade 5 DAE.

Among the VVR, 98.1%, 1.6%, 0.2% and 0.01% were Grade 1, 2, 3, and 4 respectively. Grade 3 VVRs included 14 concussions, 11 fractures, 1 dental injury, and 2 pre‐faint and 8 fainting events requiring hospitalization for work‐up. Two Grade 4 VVRs involved falls resulting in intracranial hemorrhage requiring immediate medical intervention. For allergic and apheresis DAE, there were only 6 and 5 Grade 2 reactions respectively, and no Grade 3 or 4 events. Needle‐related DAE included 98.3% Grade 1, 1.6% Grade 2, 0.1% Grade 3, and no Grade 4 events. Of the six Grade 3 Needle‐related DAE, 4 were nerve irritations lasting>6months, and 2 were DVTs requiring hospitalization.

Summary/Conclusions: The SGT provided consistent assignment of severity for the majority of DAE, based on Outside Medical Care and Specific Key search terms. Assignment of severity based on Impact on Activities of Daily Living or on Duration of injury/condition requires tracking over time making such assignments more difficult; modification of our DAE tracking database and claims database to capture ADL and Duration should improve severity assignment for such cases.

COMPLICATIONS OF BLOOD DONATION: ELEVEN YEARS OF INTERNATIONAL DATA

JC Wiersum¹, P Constantina², C Richardson³, P Renaudier⁴, N Goto⁵, E Grouzi⁶, L Kevin⁷

¹TRIP and Sanquin, Leiden, Netherlands ²Hellenic Center for Disease Control ³Panteion University, Athens, Greece ⁴Hôpital Pierre Zobda‐Quitman, Fort de France, Martinique ⁵Japanese Red Cross, Tokyo, Japan ⁶“St Savvas” Oncology Hospital, Athens, Greece ⁷Vitalant, San Antonio, United States

Background: The International Haemovigilance Network (IHN) has collected aggregate data on complications of whole blood and apheresis donations from member national haemovigilance systems (HVS) since 2006.

Aims: We analysed the data collected in 2006–2016 in order to learn from the data and consider future improvement of data collection.

Methods: National HVS entered annual data on donor complications in the password‐protected “ISTARE” (International Surveillance of Transfusion Adverse Reactions and Events) online database. From 2008 the donor complication spreadsheet allowed entry of separate data for whole blood donation (WBD) and apheresis, but also provided an option for entering data for all donation types. Annual numbers of whole blood and apheresis donations were also collected. The harmonised international standard definitions were implemented in 2015. Reactions were captured according to severity level (mild, moderate, severe) but without distinction between donor sex or first time vs repeat donation. Extracted data were used to calculate national and aggregate donor complication rates (generally per 1000 donations).

Results: Twenty‐four HVS provided figures for donations and donor complications for one or more years (median years per country was 7, IQR 2–8). The total number of country years (CY) was 138, covering 155 million donations. The overall complication rate was 6.3/1000 donations and the median country rate was 3.2 complications/1000 donations (IQR 1.1–10.1).

Rates were generally consistent within a HVS from year to year but showed considerable variation between HVS; this was also the case for reactions classed as severe. Not all countries differentiated between mild and moderate reactions and some reported all reactions under a single severity level.

Vasovagal reactions were the most commonly reported complication: overall 4.6/1000 donations, median country rate 3.1/1000 donations (IQR 0.6–7.7). Rare and apheresis‐related types of complications such as generalized allergic reaction (0.10 per 100,000, 40 CY), and major blood vessel injury (category available since 2015; overall 0.12 per 100,000, 6 CY) were only reported occasionally.

Eighteen of the HVS provided separate data for complications of whole blood and apheresis donations in one or more years (total 89 CY, 101.6 million WBD and 26.3 million aphereses, total 128 million donations). For these HVS the median rate of vasovagal reactions was 3.4/1000 WBD (IQR 1.0–9.1) and 1.5/1000 apheresis procedures (1.0–4.2). Reported haematoma rates were higher for apheresis than for WBD: the median per HVS was 0.39/1000 WBD (IQR 0.31–1.2) vs 4.2/1000 aphereses (0.69–5.6); rates of arm pain and/or nerve injury (not separated in 2006–2013) also tended to be higher: median 0.09/1000 with WBD, IQR 0.03–0.34, vs 0.39/1000 with apheresis, 0.05–0.57.

Summary/Conclusions: International reporting allows HVS to study rates of blood donation complications, to distinguish between WBD and apheresis complication rates and capture information about very rare events. Variability of reporting and of severity assessment between countries impairs the feasibility of comparisons between HVS. Work is needed to improve harmonisation of classification of donation complications and severity assessment for data comparison and research.

IMPLEMENTING FERRITIN TESTING IN STOCKHOLM – FIRST REPORT

MK Kvist, F Boström, E Watz, J Berg Thorsson, B Aspevall Diedrich

Clin Immunology and Transfusion Medicine, Karolinska University Hospital, Stockholm, Sweden

Background: To prevent iron related Hb loss, screening with ferritin testing was implemented in Stockholm county (approx. 52000 registered blood donors) during a two‐year roll‐out. Iron supplementation is offered to blood donors but has not prevented Hb deferrals resulting in 8000 control visits per year. Ferritin testing is hypothesized to increase iron compliance.

Aims: Implementation of ferritin testing for surveillance of iron levels for the entire blood donor population with specific attention to new donors, women returning after pregnancy, donors with low Hb and at return visit after low Hb. Yearly testing of plasma and platelet donors.

Methods: Ferritin testing, following a staff education program, was implemented for applicant donors, donors with low Hb, women after pregnancy, apheresis donors, followed by screening of registered blood donors per donation site. After initial screening, donors will be tested at each 4^th (women) or 8^th (men) donation, and with yearly testing of young adult blood donors below 25years. Six nurses were educated to process ferritin and blood count results. Donors with aberrant ferritin were contacted by letter.

Results: Establishment of cut‐off levels and algorithms for ferritin testing and iron treatment was evidence based but met practical limitations such as number of analyses and results that could be processed per week,limitations in LISS set‐up,blood demand contra preferred cut‐offs, iron supplementation compliance. For applicant donors, Hb testing show that 20% of female and 9% of male applicants cannot be registered because of low Hb (125 and 135mg/L respectively). Adding ferritin testing, a preferred cut‐off level of 30μg/mL (male reference level), would result in additionally 24% female and 1.3% male applicant donor loss. As this would threat the blood demand, cut‐off was set to 15μg/mL for women, above the female reference 10μg/mL, with an acceptable 6% loss of female applicant donors. For registered blood donors, 4000mg of extra iron tablets were offered at low ferritin (10–29μg/mL). This was sometimes combined with prolonged intervals and often repeated before ferritin was restored above 30μg/mL. Donors with ferritin below 10μg/mL (in 0.6% applicant donors, 0.8% registered donors) or above 600μg/mL (0.5% applicant donors, 0.1% registered donors) were deferred and recommended to see their physician. For Hb deferral, the interval was prolonged from 3 to 5months, irrespective of ferritin levels. This, together with iron supplementation, resulted in an increase from 50% to 70% approved Hb at return. The team of nurses processing ferritin and blood count results (1½ nurse fulltime weekdays) reacted to approximately 40 donor results daily, representing 20% of test results.

Summary/Conclusions: Many female donor applicants have suboptimal ferritin levels although they meet required Hb for donation. Iron treatment was added to retain donors with low ferritin as only prolonged intervals may danger the blood supply. For implementation of ferritin testing, it is necessary to have a well‐functioning and agile organization to create and apply algorithms for testing, extension of intervals and iron treatment.

FERRITIN SCREENING IN THE NETHERLANDS: FIRST RESULTS FROM A NATION WIDE DONOR DEFERRAL POLICY

M Janssen, M Vinkenoog

Donor Medicine Research, Sanquin, Amsterdam, Netherlands

Background: Since November 2017 a new donor screening regime is introduced in the Netherlands where serum ferritin levels in whole blood donors are measured periodically to further control potential iron deficiency in donors.Donor deferral thresholds are set at 30 and 15ng/mL, and donors are deferred for six and twelve months respectively if ferritin levels are below these values. As limited information is available on ferritin recovery in whole‐blood donors, the policy is introduced in parts such that adaptations to the implementation may be considered based on intermediate results and the impact of the measure on donor well‐being can be evaluated.

Aims: To assess the effect of donor deferral on donor ferritin levels.

Methods: Ferritin levels are measured in new donors and at every fifth donation in repeat donors. Donors with ferritin levels below the indicated thresholds are deferred and ferritin is re‐evaluated at their return for donation after six or twelve months. The policy allows estimating long term trends in ferritin levels post donation in repeat donors. As ferritin levels are measured in all new donors a reference distribution of ferritin levels in healthy individuals is obtained as well.

Results: Among repeat donors 46% (44% of 16,433 male donors, and 48% of 13,525 female donors) have ferritin levels below 30ng/mL and are deferred for their next donation. Furthermore, the distributions of ferritin levels in repeat male and female donors are similar and each has an average ferritin level of 43ng/mL. In contrast, we found that only 27% of new female donors (n=13,283) and 1.7% of new male donors (n=6,334) have a ferritin levels below 30ng/mL. The average ferritin level in new donors was 148ng/mL for males and 60ng/mL for females. Comparing the ferritin levels in new and repeat donors, a reduction in average ferritin levels between 1.5 and 2.0 was observed in female donors and between 3.2 and 4.4 in male donors. Both ratios increased with donor age. At the end of December 2018 2884 donors with low ferritin levels returned for donation after six or twelve months deferral. Repeat ferritin measurements show that on average the ferritin levels in female donors increased by 13ng/mL per year whereas average ferritin levels in male donors increased by 34ng/mL per year.

Summary/Conclusions: In line with earlier findings in literature our results show that repeat donations substantially reduce ferritin levels in repeat donors. These range from 1.5 to 2.0 in female and from 3.2 to 4.4 in male donors, who generally have higher ferritin levels. Deferral of donors with low ferritin levels seems to be effective in increasing ferritin levels in donors, however, further monitoring of follow‐up in repeat donors is warranted to see whether the proposed scheme allows for sufficient donor recovery over time.

CURRENT OPINION IN DONOR HEALTH – PERSONALISED DONOR CARE

W. H. Ouwehand^1,2,3

¹Department of Haematology, University of Cambridge ²Wellcome Sanger Institute ³NHS Blood and Transplant, Cambridge, United Kingdom*On behalf of the NIHR BioResource, 100,000 Genomes Project and Blood transfusion Genomics Consortium

There are ˜7000 different rare diseases and the genes for half have been identified. Approximately 3.5 million UK citizens experience premature ill‐health because of a rare disease. A conclusive diagnosis is generally not reached and on average the diagnostic odyssey lasts 2.2years. The main aims of the 100,000 Genomes Project are to reduce the diagnostic delay by embedding whole genome sequencing (WGS) to accredited standards in the care path of patients with undiagnosed rare diseases. The project started in 2013 and DNA samples from 100,000 NHS patients and their close relatives have been analysed by WGS. Here we review the results from the NIHR BioResource pilot study for the 100,000 Genomes Project comprising phenotype and genotype data from 13,037 individuals recruited at 83 hospitals using approved eligibility criteria for 15 rare disease domains.

We determined the population structure including ethnicity and relatedness estimation, high level phenotypes collected using Human Phenotype Ontology (HPO) terms and quality control and summary metrics for samples and variants. The sequence resource contains over 165 million unique variants in the 10,258 genetically independent samples, with 47% of variants previously unobserved in other large scale publicly available genome datasets. We summarise the curation of gene lists and pertinent findings in 2,000 unique diagnostic‐grade genes for the 15 domains. Over 1,000 reports assigning pathogenic or likely pathogenic causal variants have been issued with diagnostic yields varying between domains from 0.5% to 55%, while the proportion of novel causal variants ranged between 25% and 73%. We show the power of the Bayesian association test, BeviMed, to recapitulate decades of clinical genetics discoveries and by identifying >30 novel genes and novel disease‐causing variants in the non‐coding space of the genome.

We show how typing data for all red cell, HPA and HLA class antigens can be extracted from WGS data. We mined the data from the 100,000 Genomes Project and similar sequence resources to re‐version the probe content of the UK Biobank Axiom array. We genotyped 7588 donors from England and the Netherlands with this new array and observed a 99.92% concordance when comparing 92,387 blood centre‐determined antigen typing results with genotype‐determined ones. For the 48 red cell and HPA antigens that were available for 7,473 donors, the array typing provided a 3.6‐fold increase in typing results per donor (13.2 vs 47.9) and 38 rare donors were identified. Using the genotyping data we identified 2.6 times more compatible units among this cohort of donors when blood demand was modelled using referral data from 3,146 English patients with more than three red cell alloantibodies.

In conclusion the 100,000 Genomes Project has shown the feasibility of using WGS across a universal healthcare system to deliver a diagnosis for patients with rare diseases. Based on these results the NHS has commissioned the analysis of another 500,000 DNA samples from patients with cancer and rare disease. With analysis of DNA by WGS and arrays becoming part of routine clinical care, blood services must develop competencies to extract transfusion and transplant relevant information from clinical‐grade genotyping data.

MOLECULAR GENETICS TO GENOMICS

F Yamamoto^1,2

¹Immunohematology and Glycobiology, Josep Carreras Leukaemia Research Institute (IJC) ²Program for Predictive and Personalized Medicine of Cancer (PMPPC), Institut d'Investigació en Ciències de la Salut Germans Trias i Pujol (IGTP), Badalona, Spain

Next‐generation sequencing (NGS) enables the sequencing of thousands of genes, the exomes, and even entire genomes by single experiments at a reasonable price. There have also been advances in cytometry: Use of antibodies with different fluorescence tags enables simultaneous monitoring of the expression of dozens of antigens. However, immunological methods cannot detect every variant discovered by NGS. Genome sequencing reveals not only the exome but also the regulatory elements of transcription/translation, such as promoters and enhancers. RNA sequencing determines which genes and spliced transcripts are expressed. It is amazing to realize how much this novel technology has been contributing to the better understanding of various biological phenomena.

Since the initial cloning of the human blood group A transferase cDNAs in the early 1990s, we have been studying the ABO genes, A and B glycosyltransferases, and A and B oligosaccharide antigens. Various scientific disciplines including genetics, immunohematology, biochemistry, enzymology, and glycobiology have been applied to their study. We have made several important scientific contributions. We demonstrated the central dogma of ABO: the A and B alleles at the ABO genetic locus encode A and B transferases, which synthesize A and B antigens, respectively. We elucidated the allelic basis of the ABO system. We found 4 amino acid substitutions between A and B transferases and inactivating mutations in O alleles. We became the first who succeeded in the ABO genotyping, discriminating the AA and AO genotypes, as well as the BB and BO, which was impossible by the immunological approach.

We have taken a simple experimental strategy: preparation of eukaryotic expression constructs of A/B transferases and their derivatives, DNA transfection to human HeLa cells or their sublines, and immunological detection of the A/B antigen and/or biochemical examination of the enzymatic activity. We used this to show that the codons 266 and 268 are crucial in determining the sugar specificities of GalNAc/galactose of A/B transferases. We also identified mutations in several subgroup alleles causing restricted substrate use and diminished transferase activity. We also showed that cis‐AB and B(A) alleles specifying the expression of both A and B antigens by single alleles encode A‐B transferase chimeras. Since then, other scientists have characterized more than 250 ABO alleles. Recent human genome sequencings have identified many more single nucleotide polymorphism variations. The genome sequences of many species are also available. Taking advantage of those sequences and associated information, we have expanded our research to include evolutionarily related α1,3‐Gal(NAc) transferases and their genes and scaled it up from the genetic to genomic level.

In this talk, I would like to present the followings. 1: Our elucidation of the molecular genetic basis of the ABO blood group system (as requested by the organizer); 2: Identification of novel ABO alleles by others; 3: More SNP data from genome sequences and potential problems for ABO genotyping; 4: Findings obtained from analysis of ABO genes from other species; bacteria, vertebrates, to primates; 5: α1,3‐Gal(NAc) transferases and their genes and the crosstalk between A transferase and Forssman glycolipid synthase (FS); and 6: the potential causes of generation of ABO polymorphism and of species variations of the GBGT1 gene specifying the FORS polymorphism.

METHODOLOGICAL CONSIDERATIONS FOR BIG DATA APPROACHES TO TRANSFUSION MEDICINE RESEARCH

N Roubinian^1,2,3, S Kleinman⁴, E Murphy^2,3, S Glynn⁵, G Edgren⁶

¹Kaiser Permanente Division of Research, Oakland ²Vitalant Research Institute ³Laboratory Medicine, University of California, San Francisco, United States ⁴University of British Columbia, Vancouver, Canada ⁵National Heart, Lung, and Blood Institute, Bethesda, United States ⁶Karolinska Institutet, Stockholm, Sweden

In recent years, there has been a concerted effort to improve our understanding of the quality and effectiveness of transfused blood components. The expanding use of large datasets built from electronic health records allows the investigation ofpotential benefitsor adverse outcomes associated with transfusion therapy. Together with data collected on blood donors and components, these datasets permit an evaluation of the effect of donor and blood component factors on transfusion recipient outcomes. Large linked donor‐component recipient datasets provide the power to study exposures relevant to transfusion efficacy and safety, many of which may not otherwise be amenable to study for practicality or sample size reasons. Analysis of these large blood banking‐transfusion medicine datasets allow for characterization of the populations under study and provide an evidence base for future clinical studies. Knowledge generated from linked analyses has the potential to change the way donors are selected and how components are processed, stored and allocated. However, unrecognized confounding and biased statistical methods continue to be limitations inthe studyof transfusion exposures and patient outcomes. Results of observational studies of blood donor demographics, storage age, and transfusion practice have been conflicting. This review will summarize statistical and methodological challenges in the analysis of linked blood donor, component, and transfusion recipient outcomes.

A LARGE DELETION SPANNING XG AND GYG2 CONSTITUTES A GENETIC BASIS OF THE XGNULL PHENOTYPE, UNDERLYING ANTI‐Xg^a PRODUCTION

Y Lee¹, J Storry¹, V Karamatic Crew², G Halverson³, N Thornton², M Olsson¹

¹Department of Laboratory Medicine, Lund University, Lund, Sweden ²International Blood Group Reference Laboratory, NHS Blood and Transplant, Bristol, United Kingdom ³LifeSouth Community Blood Centers, Atlanta, Georgia, United States

Background: The XG blood group system comprises the homologous antigens Xg^a and CD99. The CD99 gene resides within pseudoautosomal region 1 on the short arms of the sex chromosomes and thus mimics autosomal inheritance. XG, on the other hand, is X‐linked and straddles the pseudoautosomal boundary; a truncated pseudogene composed of only the first 3 exons remains on the Y chromosome and therefore males carry a sole full‐length copy of XG. This phenomenon manifests as asymmetric frequencies of the Xg(a+) phenotype between the sexes: roughly 11% of women and 34% of men are Xg(a−). Also, whilst Xg^a immunization is rare, the vast majority of all anti‐Xg^a makers reported are men. Recently, we reported that the rs311103C variant disrupts a GATA motif between XG and CD99. This abolishes erythroid Xg^a expression and causes the common Xg(a–) red cell phenotype. However, rare individuals who produce anti‐Xg^a cannot be accounted for by this finding. We hypothesized that a structural defect in the XG coding region causes the true Xgnull phenotype underlying anti‐Xg^a production.

Aims: We undertook to determine a genetic explanation for anti‐Xg^a production.

Methods: Genomic DNA (gDNA) was extracted from two whole blood samples and cell‐free DNA (cfDNA) from 13 archived plasma samples from donors producing anti‐Xg^a; one cfDNA sample was from a female donor and the rest from males. Polymerase chain reaction (PCR) experiments, Sanger sequencing, and database searches were performed to identify and confirm the deletion. Aliquots of gDNA from four males reported to carry a similar deletion in the 1000 Genomes Project were also tested.

Results: In one gDNA sample, exon‐specific PCR identified a deletion involving part of XG and the downstream gene GYG2. Database searches indicated that the most likely deletion was the infrequent genomic structural variant esv2662319 reported in the 1000 Genomes Project. Further analyses with a short (714bp) and a long (3555bp) PCR amplicon across the suspected breakpoint determined that this deletion was approximately 114kb and corresponded well with esv2662319. This finding was confirmed in the second gDNA sample. Given the rarity of anti‐Xg^a producers, we decided to test for the same deletion in cfDNA extracted from old archived plasma samples. Of the 13 cfDNA samples, poor quality in four samples prevented amplification even from control reactions and one was contaminated with bacterial DNA. In the remaining nine samples, eight could be amplified for the deletion‐specific 714‐bp short amplicon while one was negative for the deletion. Sanger sequencing of the amplicons revealed a heterogeneous repetitive DNA element, LTR6B, hinting at a previously‐reported recombination event. This deletion was not detected in the samples from the 1000 Genomes Project which reiterates the previously identified deficiency in data interpretation and reporting for deletions.

Summary/Conclusions: A large deletion disrupting the XG and GYG2 genes accounts for the Xgnull phenotype underlying the majority (10 of 11) of anti‐Xg^a makers. One sample remained unexplained, indicating further heterogeneity to be explored. Our data help to explain why anti‐Xg^a production is rare and has primarily been reported in men.

GYPB*S WITH TWO SILENCING CHANGES COMMONLY INHERITED INDEPENDENTLY

JE Aeschlimann¹, S Vege¹, C Lomas‐Francis¹, C Westhoff¹, G Denomme²

¹Immunohematology and Genomics, NYBC, Long Island City ²Blood Research Institute, Versiti, Milwaukee, United States

Background: S and s antigens encoded by GYPB differ by one nucleotide (nt), c.143C>T, p.Thr48Met. Two different genetic backgrounds are associated with silencing of S antigen and a U+^w phenotype. These include the nt change c.230C>T (p.Thr77Met) causing partial exon skipping and designated GYBP*03N.01 (GYPB*NY) and c.270+5G>T, an intron change causing complete skipping of exon 5, designated GYPB*03N.03 (GYPB*P2).

Aims: Samples from three individuals, a previously transfused African American sickle cell patient (P1), a blood donor of unknown ethnicity (P2), and an African American patient (P3) (Lapadat R. 2014 AABB abstract) were investigated for discrepant serologic and molecular results when determining S and s phenotype.

Methods: Standard methods were used for RBC typing with licensed S and s reagents and RBCs from donor P2 were also tested with monoclonal and polyclonal anti‐S and anti‐s. DNA was isolated from WBCs and HEA PreciseType performed on P1 and P2. P1 was also tested by GYPB*S/s AS‐PCR, exon 5 PCR‐RFLP for c.270+5G>T and AS‐PCR for c.230C>T. P3 was tested for GYPB*S/s and c.230 C>T and c.270+5G>T changes by a real‐time PCR‐fluorogenic 5’ nuclease TaqMan chemistry. For all, GYPB exons 1–6 were amplified and Sanger sequenced and aligned to consensus using Clustal X.

Results: RBCs of all three probands typed S– and strongly s+ while DNA testing indicated c.143T/C (GYPB*S/s). Assay for the two common GYPB*S silenced alleles, c.230C>T and c.270+5G>T, indicated all three samples had both silencing mutations previously reported to be independently associated with a S−U+^w phenotype. HEA PreciseType could not interpret this novel allele combination and indicated GYPB*s as PV (possible variant). Samples were confirmed to be heterozygous for c.143C/T, c.230C/T and c.270+5G/T by exon specific sequencing and AS‐PCR, PCR‐RFLP and real‐time PCR. By long range sequencing of GYPB, all three were heterozygous c.59T/G and c.60A/G (p.20Leu/Trp), c.67A/T (p.23Thr/Ser), c.71A/G and c.72G/T (p.24Glu/Gly), c.143C/T (s/S), c.208G/T (p.70Val/Leu), c.230C/T (p.77Thr/Met), and c.270+5G/T. All samples were also c.251G/G (p.84Ser) and heterozygous for several previously recognized silent changes in exon 2, c.87T/C, c.96T/C and c.102A/G.

Summary/Conclusions: We report a novel silenced GYPB*S allele that can confound GYPB genotyping interpretation. The allele was found in three probands associated with a S−s+ phenotype. In these samples, two changes previously reported to be inherited independently and both associated with silencing of S antigen are carried on the same allele. DNA‐based testing could not rule out that c.230T or c.270+5T are separate and that GYPB*s was also silenced. Robust s+ RBC typing indicates both changes are on GYPB*S. Gene sequencing confirms the c.270+5T change is on a GYPB*03N.02 [GYP*He(NY)] background. c.230C>T (rs79492560) and c.270+5G>T (rs139511876) have a frequency of 0.0053 respectively 0.032 in the African population (ExAC). Although we identified 3 samples, the frequency of this novel allele is unknown.

A LUTHERAN RELATED ANTIBODY DETECTED IN A PATIENT WITH A HOMOZYGOUS MISSENSE BCAM MUTATION INDICATING A NOVEL ANTIGEN OF THE SYSTEM

L Yosephi¹, V Karamatic Crew², E Shinar¹, O Zelig³, V Yahalom^1,4, J Benjamin², P Walser⁵, N Thornton², L Muncher¹

¹Immunohematology, Magen David Adom, Ramat‐Gan, Israel ²International Blood Group Reference Laboratory, NHS Blood and Transplant, Bristol, United Kingdom ³Blood Bank, Hadassah Medical Center, Jerusalem ⁴Blood Services & Apheresis Institute, Rabin Medical Center, Petah Tikva, Israel ⁵Clinical Biotechnology Centre, NHS Blood and Transplant, Bristol, United Kingdom

Background: The Lutheran blood group system currently consists of 25 antigens. These antigens are of low immunogenicity and may cause mild‐to‐moderate transfusion reactions and hemolytic disease of the fetus and newborn. The activation of Lu‐glycoprotein/BCAM on red blood cells (RBCs) and its interaction with laminin‐5α is thought to play a role in vaso‐occlusion in sickle cell disease and other hematological disorders. The two glycoprotein isoforms Lu‐glycoprotein and BCAM are encoded by the BCAM gene which consists of 15 exons located on chromosome 19q13.2. A number of rare Lutheran phenotypes have been previously recorded in Israel, including LU:‐6,9 observed among Iranian Jews, LU:‐20 in one thalassemia patient and one case of LU:‐21. In this report, a previously transfused pregnant Arab patient with β‐thalassemia intermedia was investigated because she presented with an antibody to an unknown high frequency antigen (HFA), potentially related to the Lutheran system.

Aims: To characterize a novel Lutheran antigen through serological and molecular investigation of a patient with a Lutheran related antibody.

Methods: Initially, the red cell phenotype and the presence of a Lutheran related antibody in the serum of the patient were detected by standard serological techniques, utilizing enzyme treated and chemically modified cells and rare cells and sera from the NBGRL collection. Further serological investigations were carried out using standard IAT (LISS tube and Bio‐Rad gel) technique. Plasma inhibition studies were performed using soluble recombinant Lu protein (srLu). Eluates were prepared using acid elution method (Gamma Elu‐kit II). Genomic DNA was isolated from whole blood and all exons of the BCAM gene were amplified by PCR and directly sequenced by Sanger sequencing. The impact of the identified mutation on Lutheran glycoprotein structure was studied by molecular dynamics calculations.

Results: The patient's plasma reacted with all cells tested, except for three examples of In(Lu) cells and cells treated with 2‐aminoethylisothiouronium bromide, trypsin and α‐chymotrypsin. Inhibition studies with srLu protein showed complete inhibition of the antibody, thereby confirming the antibody to be directed toward an epitope on the Lu‐glycoprotein. In addition, testing of inhibited plasma revealed the presence of underlying anti‐E and anti‐Fy^a. An eluate was prepared to isolate the patient's Lu‐related antibody and this eluate was found to be incompatible with examples of LU:‐5, LU:‐6, LU:‐8, LU:‐13, LU:‐21, LU:‐22, and LU:‐23 cells, whereas In(Lu) were compatible. Results of serological typing of the patient's cells, for Lu system HFAs, could not be conclusively determined due to the patient having been recently transfused. However, results suggested (through absence of mixed field reactivity) the patient's cells to be LU:‐1,2,8,12,13,‐19,21. BCAM sequence analysis confirmed the patient to be LU*02, LU*18 and revealed a novel homozygous mutation c.1351A>C in exon 11, encoding p.Lys451Gln in the Lutheran glycoprotein.

Summary/Conclusions: A novel homozygous mutation c.1351A>C (p.Lys451Gln) in exon 11 of BCAM was identified in a patient with an antibody to a Lutheran HFA. Serological and genetic evidence presented here indicates discovery of a novel antigen of the Lutheran blood group system, which we propose to name LURA.

A NOVEL HIGH FREQUENCY ANTIGEN IN THE LUTHERAN BLOOD GROUP SYSTEM (LUNU)

V Karamatic Crew¹, B Mayer², L Baglow¹, S Yürek², T Bartolmäs², P Walser³, N Thornton¹

¹International Blood Group Reference Laboratory, NHS Blood and Transplant, Bristol, United Kingdom ²Institute of Transfusion Medicine, Charité‐Universitätsmedizin Berlin, corporate member of Freie Universität Berlin, Humboldt‐Universität zu Berlin, and Berlin Institute of Health, Berlin, Germany ³Clinical Biotechnology Centre, NHS Blood and Transplant, Bristol, United Kingdom

Background: Lutheran glycoprotein and basal cell adhesion molecule antigen B‐CAM are two isoforms of a type I membrane glycoprotein residing on red cell surfaces. Both isoforms are adhesion molecules with the main function of laminin binding, and both carry antigens of the Lutheran blood group system (LU). The system currently comprises 25 antigens, all encoded by mutations in the alternatively spliced single gene BCAM located on chromosome 19. Currently, ISBT lists 20 high incidence antigens in the system.

Aims: We report a case study of an individual with an unidentified alloantibody to high incidence antigen present in her plasma. Samples from the patient and her family were investigated. We provide here serological and molecular evidence for a novel high incidence antigen of the Lutheran blood group system.

Methods: Serological investigations were performed by standard IAT (LISS tube and Bio‐Rad gel) technique. Plasma inhibition studies were completed with soluble recombinant Lu (srLu) protein. Genomic DNA was isolated from whole blood of the patient and her family members; all the exons of the BCAM gene were amplified by PCR and analysed by direct Sanger sequencing. The impact of the identified mutation on Lutheran glycoprotein structure was studied by molecular dynamics calculations.

Results: Presence of a Lu‐related antibody in the patient's plasma was confirmed, reacting moderate strength by LISS IAT with untreated and papain treated cells. Cells from the patient's mother, father and two siblings were all incompatible with her plasma, though weaker than panel cells, reflecting dosage. Only In(Lu) cells were compatible with patient's plasma. The antibody was successfully inhibited with srLu protein, thereby confirming the epitope recognised by the antibody resides on the Lutheran glycoprotein. The patient's cells were found to be LU:‐1, 2, 3, 5, 6, 8, 13, 20, 21. BCAM sequencing revealed a novel homozygous mutation c.121G>A in exon 2, encoding p.Val41Met in the Lu glycoprotein. The c.121G>A change appears to be an extremely rare mutation, listed in gnomAD database with a frequency of 3.98×10^‐6 and with no known homozygous examples. Homology model of the novel Lutheran glycoprotein was subjected to all‐atom molecular dynamics calculations to analyse potential conformational changes.

Summary/Conclusions: We report serological and genetic evidence for a novel antigen of the Lutheran system, which we propose to name LUNU. The evidence will be submitted to the ISBT Red Cell Immunogenetics and Blood Group Terminology Working Party for consideration for allocation of antigen status. The absence of this high incidence antigen arises from a rare single amino acid change p.Val41Met in the Lutheran glycoprotein and the presence of anti‐LUNU in the patients’ plasma was presumed to have been made in response to previous pregnancy.

CHARACTERIZATION OF A NOVEL HIGH‐PREVALENCE ANTIGEN IN THE JMH BLOOD GROUP SYSTEM

C Vrignaud^1,2,3, S Ramelet¹, A Herb¹, A Raneri¹, M Khalloufi⁴, G Laiguillon¹, J Babinet¹, S Azouzi^1,2,3, T Peyrard^1,2,3

¹Centre National de Référence pour les Groupes Sanguins, Institut National de la Transfusion Sanguine ²UMR_S1134 Inserm Université Paris Diderot ³Laboratoire d'Excellence GR‐Ex, Paris, ⁴Site d'Avicenne, Etablissement Français du Sang Ile de France, Bobigny, France

Background: The JMH blood group system includes six high‐prevalence antigens carried by semaphorin 7A (Sema7A/CD108), a GPI‐linked glycoprotein. SEMA7A consists of 14 exons and encodes a preproprotein of 666 amino acids, further proteolytically cleaved to generate a mature 604 amino acid protein. JMH:‐1 is mostly found in elderly people and considered an acquired phenotype, whereas JMH:‐2 to JMH:‐6 originate from homozygous missense mutations in SEMA7. JMH antibodies are considered poorly clinically significant, most of them being IgG4, but they can mask alloantibodies of clinical relevance.

Aims: We describe here a novel high‐prevalence antigen in the JMH blood group system.

Methods: Blood samples were referred for investigation of a pan‐agglutination. Antibody identification (gel‐test IAT, polyspecific and IgG, Bio‐Rad) was performed on native, papain‐treated (Diagast) and trypsin‐treated (Sigma) RBCs. Genomic DNA was extracted from peripheral blood cells by an automated method, amplified by SEMA7A exon‐specific primers and sequenced.

Results: The proband was a 32‐year‐old female patient of Moroccan origin, group A, D+C+E‐c+e+, K‐, without transfusion history. She was hospitalized at 27weeks gestation for a blighted ovum requiring a manual vacuum aspiration, with a significant hemorrhage risk. A RBC antibody screening was performed by a first laboratory. The antibody reacted 1+by IAT on all native reagent RBCs, with negative autocontrols, but was nonreactive on papain‐ and trypsin‐treated cells. An anti‐Ge2 was initially suspected, due to the pattern of reactivity and ethnic background. New blood samples were referred to our national immunohematology reference laboratory. The antibody showed the same profile. Anti‐Ge2 and anti‐Ch1 could be ruled out. The serum was nonreactive with two JMH:‐1 and positive with two JMH:‐3 samples. The patient was found to be JMH1 positive. In addition, a soluble recombinant JMH protein (JMH Imusyn/Inno‐train) fully abolished the reactivity of the pan‐agglutinating antibody. The antibody was an IgG4. Overall, these results were consistent with a probable JMH variant and prompted us to perform SEMA7A sequencing. Three nucleotide changes were found, in homozygous state: a rare non‐synonymous change in exon 7, c.709G>A (p.Asp237Asn, rs140707085, MAF<0.01, SIFT score=1); a common synonymous change in exon 12, c.1545A>G (p.Gln515Gln, rs741761, MAF=0.5); a rare non‐synonymous change in exon 14, c.1865G>A (p.Arg622His, rs140128092, MAF<0.01, SIFT score=0.36). The analysis of surface accessibility of Asp²³⁷ and Arg⁶²² using the 3D structure of Sema7A (RCSB PDB‐3NVQ https://www.rcsb.org/structure/3nvq) showed that only Arg⁶²² was predicted to be an exposed‐epitope. Interestingly, all other reported JMH variant phenotypes correspond to an arginine substitution. Of note, we retrospectively found another individual of Algerian ancestry (pregnant woman) with a pan‐agglutinating IgG4 antibody showing a similar pattern of reactivity, and with the same three changes in SEMA7A. We unfortunately could not perform a cross‐compatibility testing with the proband (no material left and unsuccessful contact).

Summary/Conclusions: Serological and molecular studies allowed us to provide evidence for a novel high‐prevalence antigen in the JMH blood group system, very likely encoded by the p.Arg622His substitution in Sema7A. We propose to provisionally assign the name JMH7 for this antigen. Interestingly, our two unrelated JMH:‐7 individuals were from North African ancestry.

β1,3GalNAc‐T1‐DEPENDENT EXTENSION OF THE HUMAN BLOOD GROUP B ANTIGEN RESULTS IN A NOVEL ABO‐RELATED GLYCOLIPID STRUCTURE ON ERYTHROCYTES

J Ricci Hagman^1,2, A Barone³, J Westman⁴, J Storry^1,2, C Jin³, A Hult^1,2, S Teneberg³, ML Olsson^1,2

¹Dept. of Laboratory Medicine, Lund University ²Dept. of Clinical Immunology and Transfusion Medicine, University Laboratories, Region Skåne, Lund ³Department of Medical Biochemistry and Cell biology at Institute of Biomedicine, University of Gothenburg, Gothenburg, Sweden ⁴Sanford Burnham Prebys Medical Discovery Institute, Center for Nanomedicine, University of California Santa Barbara, Santa Barbara, CA, United States

Background: The ABO system was discovered almost 120years ago and the underlying structures later elucidated as carbohydrates carried by glycoproteins and glycolipids. The terminal trisaccharides GalNAcα3(Fucα2)Gal and Galα3(Fucα2)Gal constitute the clinically important A and B epitopes, respectively. Clausen et al. (PNAS, 1985) showed that the A antigen could be extended to a repetitive glycolipid A epitope, GalNAcα3(Fucα2)Galβ3GalNAcα3(Fucα2)Galβ4GlcNAc‐R. However, extended forms of B antigen have not been described. We encountered two related situations with unexplained serological reactivity. Firstly, enzyme‐conversion to group O treatment of group B (B‐ECO) red blood cells (RBCs) with α3‐specific GH110 family exogalactosidase (Bzyme) abolishes B antigens as detected by hemagglutination and flow cytometry with all monoclonal anti‐B tested. Despite this, 40% of group O plasmas have been reported to give positive crossmatch results with B‐ECO RBCs. Secondly, plasmas from AB and B individuals of the globoside‐deficient P^k phenotype contain anti‐P and anti‐PX2 but react stronger with Bpp‐RBC than with App/Opp‐RBC. Based on these findings, we hypothesized the presence of a Bzyme‐resistant, B‐related glycolipid.

Aims: To identify the molecular basis of the enigmatic serological observations outlined above.

Methods: Plasma and eluates from an A1B individual with the P1^k phenotype were investigated by hemagglutination and flow cytometry, as were eluates from B P1^k and O plasma. RBC membrane glycolipids were extracted from two batches of pooled, expired group B‐RBC units (frozen 500‐litre reference preparation and confirmatory preparation from 24 freshly collected units). Native or enzyme‐treated glycolipid fractions were analysed by liquid chromatography electrospray ionization‐mass spectrometry (LC‐ESI/MS) and immunostaining of thin layer chromatography (TLC) plates. Antigen expression in the H+B− human erythroleukemia (HEL) cell line was analysed by flow cytometry following overexpression of selected glycosyltransferases.

Results: Anti‐P‐depleted eluates made from A1B P1^k plasma contained anti‐PX2 and antibodies of unknown specificity that reacted stronger with native or papain‐treated Bpp‐RBCs compared to App/Opp‐RBCs. Anti‐PX2 was removed by adsorption onto Opp‐RBCs but reactivity (here designated anti‐ExtB) remained against B/Bpp/B‐ECO RBCs. LC‐ESI/MS of glycolipid fractions from group B units revealed an unknown HexNAc‐Hex‐(Fuc‐)Hex‐4HexNAc‐Hex‐4Hex heptasaccharide. Upon β‐N‐acetylhexosaminidase treatment of this candidate structure, a group B type 2 hexasaccharide was produced, demonstrating that the terminal HexNAc of the HexNAc‐Galα3(Fucα2)Galβ4GlcNAcβ3Galβ4Glc heptasaccharide was β‐linked. Immunostaining of TLC fractions with GalNAc‐specific W. floribunda lectin gave distinct binding as did anti‐ExtB. Bzyme converted B‐hexasaccharide to H‐pentasaccharide whilst the candidate heptasaccharide persisted. Co‐transfection of HEL cells with wild‐type ABO (GTB) and B3GALNT1 induced expression of GalNAcβ3Galα3(Fucα2)Galβ4GlcNAcβ3Galβ4Glc, as demonstrated by staining of HEL cells by anti‐ExtB. Co‐transfections involving mutated GTB (c.261delG) or B3GALTN1 (c.202C>T) completely ablated reactivity.

Summary/Conclusions: We demonstrate for the first time the presence of extended blood group B glycolipids with a terminal β3‐linked GalNAc and propose this previously undescribed structure to be designated ExtB. Plasma from the globoside‐deficient AB or B individuals tested contain anti‐ExtB, as do many group O plasmas. We identified the B‐elongating glycosyltransferase as P synthase (β1,3GalNAc‐T1). Consequently, ExtB fulfils all blood group criteria and is hereby proposed as an emerging GLOB system antigen. Our findings may explain Bzyme‐resistant crossmatch reactions, which have hampered attempts to make ABO‐universal RBCs for transfusion.

An update on Anti‐platelet agents

D Cox

Royal College of Surgeons in Ireland, Dublin, Ireland

Since the discovery of the anti‐platelet effects of aspirin platelets have been a major therapeutic target for pharmaceutical companies and also a very profitable target. However, the effectiveness of aspirin has also been a challenge as it is an inexpensive drug and any new agent needs to show clear benefit over aspirin. Furthermore the risk of bleeding from anti‐platelet agents, especially cerebral bleeds, has also presented challenges. In the 1990's orally active GPIIb/IIIa antagonists were considered to be the `super aspirin’ but clinical trials showed increased mortality and ultimately this class of drugs was relegated to iv use only in high‐risk patients. GPIb/IX/V antagonists were also a promising drug target but no agent made it to market. The real breakthrough was the discovery of the P2Y12 antagonist clopidogrel which, in conjunction with aspirin, proved to be very effective at preventing thrombotic events and as a result it became the biggest selling drug in the world at the time. With clopidogrel now off‐patent the combination of aspirin and clopidogrel is a formidable challenge to any new agent both in efficacy terms and pharmacoeonomic terms. So is there a future for new anti‐platelet agents? With the growing awareness of the role of platelets in inflammation and an understanding of how the immune activation of platelets differs from the classical haemostatic activation of platelets it is now possible to develop novel anti‐platelet agents that target inflammation without compromising haemostasis. It is here that we should look for the next generation of anti‐platelet agent.

PLATELET FUNCTION MEASUREMENTS IN TRANSFUSION MEDICINE

P Fontana

University Hospitals of Geneva, Geneva, Switzerland

Platelet function defects, either congenital or acquired, are associated with increased bleeding risk, particularly in a perioperative setting. The use of platelet function assays is therefore tempting in order to tailor transfusion and limit platelet transfusion to those bleeding patients with impaired platelet function, as assessed by those assays.

However, the current guidelines provide only weak recommendations supporting the routine use of these assays. Indeed, there are numerous platelet function assays on the market that differ in their method of evaluation of platelet function and agreement between their results is at best moderate. The threshold values beyond which procedure‐associated bleeding risk becomes worrisome is not standardized. Moreover, observational studies addressing the predictive value of platelet function testing in perioperative or spontaneous bleeding are not consistent. Finally, management trials with randomized patients assessing the benefit of platelet function testing are scarce. More recent data identified selected situations where platelet function testing may be useful though.

I will review the different platelet function assays as well as selected clinical studies addressing the impact of platelet function testing to improve bleeding and transfusion‐related outcomes. The latest recommendation will be addressed too.

PLATELET COMPATIBILITY AND PLATELET ANTIBODIES DETECTION: A STEP TOWARDS RESOLVING DILEMMA IN MANAGEMENT OF PLATELET REFRACTORINESS IN ONCOLOGY PATIENTS

S Mangwana, A Kacker, N Simon

Transfusion Medicine and Immunohematology, Sri Balaji Action Medical Institute, New Delhi, India

Background: Platelet refractoriness complicates the provision of platelet transfusions in management of thrombocytopenia in Oncology patients. Platelet refractoriness poses challenge due to alloimmunization to HLA and human platelet antigens and is associated with adverse clinical outcomes.

Aims: A prospective study was undertaken to analyse result of platelet compatibility with post‐transfusion platelet count increment and to ascertain presence of platelet antibodies as causative factor in platelet refractory oncology patients.

Methods: Eighty oncology patients having thrombocytopenia in a tertiary care centre; both solid organ and hematological malignancies, requiring platelet transfusion were included. ABO‐compatible, leucoreduced, random donor platelets with less than 72hours storage were transfused. Simultaneously platelet cross‐matching was performed with pre transfusion sample. Blood samples were collected after 1hour of completing transfusion for post‐transfusion platelet count and Corrected Count Increment calculation. In case of platelet refractoriness and presence of platelet incompatibility, platelet antibody screening test was performed using SPRCA technique. Mann‐Whitney test for Quantitative variables and Fisher's exact test for Qualitative variables were used. A p‐value<0.05 was considered statistically significant.

Results: Study population was 18–85years with maximum number of cases (31.3 %) in 60–70years age group with equal number of case of both genders. 64 cases (80 %) cases showed platelet cross‐match compatibility, while 16 cases (20%); 9 male and 7 female patients, were platelet cross‐match incompatible. Amongst incompatible platelet cross matches, 14 (87.5%) cases showed presence of platelet alloantibodies (17.5%) and all cases except one showed platelet refractoriness. Out of 64 compatible platelet cross matches cases, 38 cases (59.37%) showed platelet refractoriness which is statistically significant than cases showing adequate CCI. None of these cases showed platelet alloantibodies. Platelet yield in compatible platelet cross match was higher than in patients with incompatible platelet cross match (p‐value<0.001).Previous history of Pregnancy in female patients (61% cases) and transfusion (54% cases) played a vital role in platelet refractoriness and development of platelet alloantibodies. Solid organ malignancies showed more refractoriness (67.4%) than in hematological malignancies (63.15%). Patients treated with chemotherapy (78.8%) had high risk to platelet refractoriness. Leukocytosis had statistically significant correlation with platelet refractoriness. Age and Gender did not affect platelet increment. Body Surface Area had negative correlation with platelet increment.

Summary/Conclusions: To manage platelet refractoriness in medical oncological patients, platelet cross matching; similar to red blood cell cross match, using SPRCA method is an effective, useful tool and rapid, first‐line approach for selecting compatible platelets from the local inventory as compared to HLA‐matched platelets in the treatment of thrombocytopenic cancer patients. Platelet cross‐match along with testing for anti‐platelet antibodies is an important, less time‐consuming and cost‐effective component than the molecular testing in the management of oncology patients.Blood services must be aware of the measures to prevent alloimmunisation and correct identification of refractoriness to provide adequate transfusion support for oncology patients.

ASSESSMENT OF CORRELATION OF PLATELET CROSSMATCH RESULTS BY SOLID PHASE RED CELL ADHERENCE ASSAY (SPRCA) WITH POST‐TRANSFUSION PLATELET COUNT INCREMENT IN ADULT HEMATO‐ONCOLOGY PATIENTS

PB Sontakke, P Desai, A Navkudkar, S Rajadhyaksha

Transfusion Medicine, Tata Memorial Centre, Mumbai, India

Background: Platelet transfusions are an important aspect of supportive transfusion therapy for patients of hematological malignancies and hypoproliferative thrombocytopenia. A less than expected increase in platelet count occurs in about 20% to 70% of multiply transfused patients with thrombocytopenia. Transfusing HLA matched platelets is the best strategy to overcome this clinical condition caused by alloimmune mechanisms. The disadvantage of HLA matched platelet transfusion is that it is logistically challenging. Alternate approach is to transfuse crossmatched platelets in these patients which is convenient and feasible. In this study, we tried to assess the correlation of platelet crossmatch result by SPRCA with the post‐transfusion platelet count increment among adult hemato‐oncology patients in the form of one hour corrected count increment (CCI).

Aims: To assess platelet crossmatch result by SPRCA and find its correlation with post‐transfusion platelet count increment among adult hemato‐oncology patients

Methods: This was a pilot study approved by institutional ethics committee in which 50 adult patients of hematological malignancies having history of two or more platelet transfusions and absence of nonimmune causes for inadequate response to platelet transfusion were included after obtaining prior informed consent. They were transfused one unit of ABO identical single donor platelet of which the platelet samples were preserved. Ten minutes to 1hour post‐transfusion CCI was calculated and documented. CCI>7500 was considered as adequate response for each patient. Crossmatching by SPRCA was performed by using preserved platelet samples from donor units with the serum of the respective patient. Statistical analysis of the correlation between platelet crossmatch results and CCI was done by using appropriate statistical tools.

Results: Out of 50 crossmatches, 78% (39/50) samples showed compatible and 22% (11/50) showed incompatible results. Among 39 compatible results, 87.2% (34/39) showed adequate CCI and 12.8% (5/39) showed inadequate CCI. Among 11 incompatible results, 18.2% (2/11) had adequate CCI and 81.8% (9/11) had inadequate CCI. The difference between response to platelet transfusion in terms of CCI to compatible and incompatible crossmatches was found to be statistically significant (P<0.05). Other variables like gender, age, body surface area, number of previous transfusions and underlying clinical condition of the patient were not found to have any effect on compatibility of crossmatch (P>0.05).

Summary/Conclusions: In present study, most of the patients who showed adequate CCI had compatible crossmatch. To select the compatible unit from the available inventory, the SPRCA is a rapid, effective and feasible method in an oncology set up which demands multiple platelet transfusions to patients of hematological malignancies. Thus, transfusion of crossmatched platelets from the available inventory to multiply transfused patients of hematological malignancies can be a better option to transfusing randomly selected platelets.

UPDATE ON TRALI/TACO AND TAD (AABB PRE MEETING)

A Vlaar

Intensive Care Medicine, Amsterdam UMC, Amsterdam, Netherlands

Pulmonary complication after blood transfusion is the leading cause of transfusion–related morbidity and mortality, with an incidence reported between 0.05 – 15% of all transfused patients. The most important transfusion related pulmonary complications are transfusion associated circulatory overload (TACO), transfusion related acute lung injury (TRALI) and transfusion associated dyspnea (TAD). In this presentation the recent changes in the international definitions will be presented and discussed. Furthermore, insights in the different underlying pathophysiologic mechanisms will be highlighted. In the past decades only for TRALI prevention strategies have successfully been designed and implemented. Currently no evidence‐based treatment strategy is available for any of these life‐threatening syndromes. Insight in the pathogenesis of pulmonary complications after transfusion should pave the way for future prevention and treatment studies.

IRON OVERLOAD FROM TRANSFUSION AND IMPACT ON TRANSPLANT OUTCOMES

E Angelucci¹, F Pilo²

¹Hematology and Oncology, IRCCS Ospedale Policlinico San Martino, Genova ²Internal Medicine and Oncology, Businco Cancer Center, Cagliari, Italy

The issue of the impact of iron overload / toxicity on the hematopoietic stem transplantation (HCT) outcome has been firstly addressed in the field of transfusion dependent thalassemia. Today the concept has been extended to other diseases characterized by periods of variable duration of transfusion dependence such as Myelodysplastic Syndrome (MDS) and myeloproliferative diseases. Patients requiring regular blood transfusions certainly develop iron overload leading to tissues and organ damage.

Iron burden before transplant significantly impacts outcome and long‐life post‐transplant.

It is well known that iron overload is deleterious for organs such as liver, heart and endocrine glands and it has been postulated could also increases the risk of infections and severe Graft versus Host Disease early after HCT.

Recent preclinical data has shown how increased production of reactive oxygen species (ROS) resulting under iron overload condition, could impair the stem cells clonality capacity, proliferation and maturation. Also, microenvironment cells could be affected through this mechanism. For this reason, iron overload is becoming an important issue also in the engraftment period early post‐transplant.

High baseline Ferritin levels before HCT have been shown to negatively influence clinical outcome, but nowadays, ferritin is considered a steady and not biologically active form of iron, while free iron forms as non ‐transferrin bound iron (NTBI) and labile plasma iron (LPI) are considered the main trigger of cell damage more representative of the dynamic tissue damage. The scientific community is moving the iron disease from a “Bulky” disease, such as classically in thalassemia (based on quantitative iron parameters as ferritin, red blood cell transfusion number, MRI) to a “toxic” disease (based on active and dynamic biological markers as NTBI/LPI).

At this time in all the studies published on HCT setting, only the correlation between direct or indirect estimates of iron overload (mainly serum ferritin) and outcome parameters has been explored, while the duration of exposure to toxic iron species has not been taken into account.

The first study that explored the LPI role in relationship with outcome was published by Wermke and colleagues in malignancies. They investigated the predictive value of both stored (MRI‐derived liver iron content) and non‐transferrin‐bound‐iron, defined as enhanced labile plasma iron (eLPI) on post‐transplantation outcomes in patients with acute myeloid leukemia or MDS. Their prospective, observational ALLIVE study showed that patients who had raised eLPI concentration at baseline, also had significantly increased incidence of non‐relapse mortality at day 100 (33%) compared with those who had normal eLPI at baseline (7%) (P=0.00034).

Reinterpreting transplant predictive factors in the light of the current advances in understanding iron homeostasis further supports the concept that the key to successful transplantation is regular and life‐long chelation therapy to consistently suppress tissue reactive iron species and prevent tissue damage in the years before HCT.

IMPACT OF BLOOD DONOR SEX ON RECIPIENT OUTCOMES

R Middelburg^1,2

¹Center for Clinical Transfusion Research, Sanquin Research ²Clinical Epidemiology, Leiden University Medical Center, Leiden, Netherlands

In transfusion medicine, the role of donor sex was long considered to be limited to the increased risk of TRALI observed after transfusions from female donors. This risk has been shown to be limited to female donors with a history of pregnancy and to plasma rich products (i.e. excluding red blood cell products, typically containing<50mL plasma).

Until, in 2011, we found that sex‐mismatched red blood cell transfusions were associated with increased recipient mortality. Since then, several other studies have confirmed these findings, but some studies also did not find an association. All of these studies relied on the analyses of routinely collected health care data, which was not primarily intended to be used for research. As a result, analyses are complex and often difficult to properly appraise based on published descriptions. Therefore, the discussion about possible reasons for these discordant findings has largely focused on the methodological approaches of the different studies. Other potential explanations include differences in donor or patient populations, production methods, or storage time of blood products. The different potential explanations are expected to be associated with different underlying biological mechanisms. Therefore, further delineating which donor, patient, and product characteristics modify the observed association could provide more insight into the underlying mechanism.

In 2017, we observed that only transfusions from female donors with previous pregnancies were associated with increased mortality and only in male recipients under 50years. This leads us to postulate that pregnancy induced long term changes in the female immune system are transferred during red cell transfusion, with negative consequences for young male recipients. The low amount of plasma present in red cell products further lead us to assume a cellular component, like passenger leukocytes, to be involved. It has been shown that micro‐chimerism of passenger leukocytes can persist for decades after transfusion, even of leuko‐reduced blood products, suggesting long term immune‐modulation could play a role.

We hypothesized that passenger leukocytes would die during storage of blood products and the negative effect of ever‐pregnant female donors, on the survival of young male red cell recipients, would therefore be attenuated by increased storage time. However, our data seem to indicate the opposite. The risk of death was increased over three‐fold for young male recipients of old (>24days storage) red cells from ever‐pregnant donors, compared to for young male recipients of fresh (<10days storage) red cells from ever‐pregnant donors (3‐year cumulative incidence of death 15.4% versus 4.8%). The negative control group (i.e. young male recipients of red cells from male donors) showed a much weaker association of mortality with storage time (i.e. 7.2% versus 4.7%).

These findings seem to falsify our hypothesis that mortality could be caused by passenger leukocytes, establishing long term immune‐modulatory effects.

Another potential mechanism that has been suggested could be the presence of cell‐free DNA in transfused blood products. This cell‐free DNA increases during storage. However, more research is needed both to establish if cell‐free DNA can also be linked to previously pregnant blood donors and by which mechanism it could negatively affect young male transfusion recipients.

CHALLENGES OF BLOOD COMPONENT THERAPY IN SUB‐SAHARA AFRICA

JO Mulenga

Zambia National Blood Transfusion Service, Lusaka, Zambia

The World Health Organization (WHO) defines Blood Components as the constituents of blood which are prepared through physical means at controlled speed, Time and Temperature.

The four blood components are: Red cell Concentrates (Rcc), Fresh Frozen Plasma (FFP) Platelets and Cryoprecipitate. The value of the Blood components therapy includes: provision of optimal usage of products, patients receiving the appropriate portion of blood for the condition to be treated or managed.

According to the WHO AFRO regional status report on Blood Component production on the continent of Africa, Blood component preparation Red cell concentrates were being prepared in 32 countries in 2006 compared to 29 countries in 2004, while 25 countries reported that they could prepare platelet concentrates in 2006 as against 29 in 2004. Fresh frozen plasma was reported as being prepared in the same number of countries in both 2004 and 2006, while the number of countries preparing cryoprecipitate sharply declined from 29 in 2004 to 15 in 2006

CHALLENGES

• ‐

  Lack of centrally coordinated Blood system:

• ‐

  Inadequate and /or Poor blood component production equipment

• ‐

  Lack of adequate budget for equipment maintenance.

• ‐

  Lack of sufficient funding for operations.

• ‐

  Lack of Apheresis equipment

• ‐

  Lack of Gamma Irradiated products

• ‐

  Lack leucodepleted products

• ‐

  Lack of capacity to detect rare antibodies

• ‐

  Dependence on first time Voluntary non- remunerated Blood Donors

• ‐

  Dependence on family replacement donors

• ‐

  Lack of capacity in recruiting next generation of donor using ICT solutions.

• ‐

  Escalating cost of reaching blood donors as most collections are done on long outreach trips

• ‐

  Lack of evidence based donor selection criteria;

• ‐

  Lack of capacity to investigate effects of long term donations.

• ‐

  High incidence of TTIs in donor and general populations.

• ‐

  Lack of specialized continued education for clinicians in appropriate use of blood and blood components

• ‐

  Lack of Hospital Transfusion Committees.

CHALLENGES OF CLINICAL TRIALS IN RESOURCE LIMITED SETTINGS: PERSPECTIVES FROM UGANDA

A Kambugu

Infectious Diseases Institute, Makerere University, Kampala, Uganda

Clinical trials (CTs), the gems in clinical research for generating robust evidence in medicine and public health, are costly and complicated undertakings. In resource limited setting like sub‐Saharan Africa (SSA) where the health systems are sub‐optimal and where capacity for research is limited, the conducting of CTs can be a daunting challenge.

The challenges of undertaking CTs in RLs may be categorized based on the occurrence of the bottleneck(s) in relation to the ethics and regulatory approval process:

Pre‐approval

Protocol Development: In order to develop a context‐specific protocol which is subsequently subjected to an ethics and regulatory approval process, investigators need to review and ensure that the protocol is pragmatic and feasible with respect to implementation. This results into a time‐consuming reiterative process of reality‐checking the protocol.

Site Selection: In light of the limited research infrastructure, investigators in RLS and their developed world partners spend considerable time reviewing and selecting suitable sites for participation in the anticipated protocol for the CTs. Suitable sites are usually very few and with competing on‐going studies.

Approval: Institutional Review Board (IRB) approval: The IRB approval process can be quite lengthy (6‐9months) with considerable unpredictability in the periods between the initial and subsequent IRB reviews.

National regulatory approval: The requirements by national regulators are unusually innumerable with limited flexibility to accommodate specific CTs.

Post‐approval

The key post‐approval challenges for CTs implementation in RLS are attaining appropriate participant enrolment and maintaining high retention rates. Specifically, for participant enrollment, the challenge may be unforeseen competing CTs targeting the same participant pool or community perspectives that may discourage participants from getting screened for the CTs. Retention may also be a challenge particularly where participants view enrollment as a chance to access healthcare services may therefore not have any incentive to keep in a study after the initial study visits.

In conclusion, CTS are complex undertakings wherever they are conducted but are doubly challenging in RLS like sub‐Saharan Africa. The bottlenecks at the pre‐approval, approval and post‐approval stages are considerable. Nevertheless, it is rewarding to perform CTUs in RLS given that the data generated therein is highly valued by national regulators and may hasten the registration process for medical products.

Toward an Appropriate Pathogen Reduction Technology for Whole Blood in Resource–Limited Settings

S Amar¹, R Schwabe¹, A Grzesiczek¹, M Lanteri², N Mufti², J Pitman², C Tayou Tagny^3,4

¹Transfusion Service, Transfusion CRS Suisse, Bern, Switzerland ²Cerus Corporation, Concord, CA, United States ³Hematology and Blood Transfusion Service, Yaounde University Teaching Hospital ⁴Faculty of Medicine and Biomedical Sciences, University of Yaounde I, Yaounde, Cameroon

Background: Interest in an appropriate and effective whole blood (WB) pathogen reduction technology (PRT) is growing, especially in sub‐Saharan Africa where the residual risk of transfusion‐transmitted infections (TTIs) remains unacceptably high and WB is still frequently used. Cerus Corporation, manufacturer of the INTERCEPT™ Blood System, and Swiss Transfusion SRC are collaborating on a clinical development program to adapt INTERCEPT PRT using amustaline (S‐303) and glutathione (GSH) for red blood cells (RBCs) into an appropriate PRT for WB in resource‐limited settings in Africa. Treatment with amustaline/GSH has been shown to inactivate a broad spectrum of transfusion‐transmissible pathogens in RBCs. Studies with amustaline/GSH in WB have shown effectiveness against a duck hepatitis B virus (>5.3 log reduction) and Plasmodium falciparum (>7.5 log reduction), with future studies planned. A WB PRT system with amustaline/GSH also has the potential benefit of minimal electricity requirements.

Aims: To describe the safety and clinical objectives for a Phase 1 clinical trial using the amustaline/GSH PRT system for WB in Africa, and describe research and development efforts to adapt the INTERCEPT PRT system for RBCs into a robust and appropriate WB system for settings with high burdens of TTI and limited resources.

Methods: The protocol for a Phase 1 clinical trial using pathogen‐reduced WB treated with amustaline/GSH in an African country is presented, as are current research and development activities related to the development of a PRT system for WB.

Results: In the planned Phase 1 clinical trial in Africa, 20 clinically stable patients with anemia who require WB transfusion will be randomized into two study arms at a large medical center in a sub‐Saharan African country. Enrolled patients will receive one unit of non‐leucocyte‐reduced WB treated with amustaline/GSH, or a unit of untreated control WB or RBCs. The primary safety endpoint will be the incidence of high‐imputability transfusion reactions (Swissmedic ≥Grade 2) within the first 24hours of transfusion. Data will also be collected on all adverse events and transfusion reactions (all grades) and the development of treatment‐emergent antibodies to pathogen‐reduced WB or auto‐antibodies within 59 (±3) days of the study transfusion. Clinical efficacy will be characterized by hemoglobin increment 24hours after transfusion adjusted to hemoglobin dose and body weight.

Summary/Conclusions: A PRT system for WB is being developed based on the INTERCEPT PRT for RBCs that is in advanced development in Europe and the United States. INTERCEPT‐treated RBCs have met efficacy and safety endpoints in Phase 3 clinical trials. The amustaline/GSH PRT system used to treat INTERCEPT RBCs has demonstrated effective inactivation against a broad spectrum of agents that may result in TTIs. A Phase 1 clinical trial using an adapted PRT system for WB in Africa is the first step in a clinical development program that includes additional pathogen inactivation efficacy studies and improvements to the WB PRT implementation process. Together, these developments and evaluations represent progress toward a realistic and appropriate PRT for WB in Africa and other resource‐limited settings.

BRINGING FIRST‐TIME PLASMA DONORS BACK: FINDINGS OF AN INTERVENTION STUDY

R Thorpe¹, T Davison¹, B Masser², L Nguyen¹

¹Clinical Services and Research, Australian Red Cross Blood Service, Melbourne ²School of Psychology, University of Queensland, Brisbane, Australia

Background: In Australia, demand for plasma‐derived products has increased dramatically, and there is a need to increase plasma collections. First‐time donor retention, including the rate at which first‐time donors return, is a pressing issue. A quick return is optimal as this increases the overall plasma yield and is associated with long‐term retention. However, we lack evidence of effective interventions to encourage first‐time donors, particularly those donating plasma, to return and to establish a higher frequency donation routine.

Working from Schultz's (2014) framework, this intervention study was based upon insights from interviews with first‐time plasmapheresis donors. Participants identified barriers such as time and lack of knowledge about plasmapheresis. Facilitators included being able to help more people and to donate more frequently than allowed with whole blood. Participants generally favoured donating at a frequency of every 4weeks.

Aims: The aim of this study was to test the effectiveness of three intervention conditions compared with the business‐as‐usual (BAU) procedure on the proportion of donors returning to donate plasma and the number of plasma donations. We report on the data from 2months post‐donation.

Methods: Donors were randomly assigned to one of four study conditions. In Conditions 1 and 2, donors received an email one day after their initial donation. In the first condition, donors received the BAU ‘thankyou’ email. Donors in the second condition received an alternative email with content derived from the interview study. Donors in the remaining conditions received either the BAU email (Condition 3) or the revised email (Condition 4) coupled with a telephone call. The phone call was scripted to provide additional information about plasma, including how often plasma can be donated, a suggestion to donate every 4weeks, and a prompt to forward‐book appointments.

Results: The final sample (N=6788) comprised 3859 women (57%) and 2929 men (43%) aged 18–70 (mean=32). After two months 37.2% of donors returned to donate plasma at least once. After controlling for gender, age, and blood group, donors in each of the intervention conditions were more likely to return to donate plasma than were donors in the BAU condition. The greatest effect was found between donors randomized to Condition 4 (revised email + phone call), OR=1.305, CI=1.128–1.510, and BAU. Donors assigned to the two telephone conditions (Condition 3 and 4) donated plasma at a higher frequency than BAU.

Summary/Conclusions: This study tested the effectiveness of interventions designed to encourage first‐time plasma donors to return to donate plasma and to establish a routine of donation. Early indicators suggest that the evidence‐based email and phone call elements are more effective than BAU in bringing donors back to donate plasma, and the revised email combined with a phone call had the greatest positive effect on short‐term plasma yield.

DOUBLE‐ERYTHROCYTE APHERESIS VERSUS CONVENTIONAL WHOLE BLOOD PHLEBOTOMY AS IRON DEPLETION TREATMENT IN HEALTHY CARRIERS OF HFE MUTATIONS WITH HYPERFERRITINEMIA

L Infanti^1,2, G Leitner³, A Plattner¹, V Pehlic¹, A Holbro^1,2, N Worel³, A Buser^1,2

¹Blood Donation Center Swiss Red Cross Basel ²Division of hematology, University Hospital, Basel, Switzerland ³Clinic for Blood Group Serology and Transfusion Medicine, Medical University, Vienna, Austria

Background: Healthy individuals with hereditary hemochromatosis (HH defined as hyperferritinemia and homozygous p.C282Y mutation), but also carriers of other HFE mutations (p.C282Y/p.H63D or homozygous H63D) with elevated serum ferritin (SF) are accepted as blood donors, if allowed by local regulations and if eligibility is fulfilled. Generally, blood components are released for transfusion at normal SF levels (< 200ng/mL in females, < 300ng/mL in males).

Aims: Prospective, two‐center, randomized study comparing theefficacy and tolerability of double‐erythrocyte apheresis (2RBCaph) and whole blood phlebotomy (WBph) for iron depletion in asymptomatic subjects with HH or hyperferritinemia and other HFE mutations in the setting of routine blood donation.

Methods: Eligibility criteria included age≥18–60years, total blood volume≥5L, BMI<35kg/m², Hb≥140g/L, elevated SF levels and no end organ damage due to iron overload. 2RBCaph (360mL RBC) were scheduled every 14days and WBph (450mL) every 7days until SF was<100ng/mL. A complete blood count and SF were measured at baseline, at every visit and at follow up 8weeks after completion of the study. Adverse events were systematically recorded. The treatment effect was tested by Poisson regression, with gender, HFE mutation, BMI and baseline SF as covariates.

Results: 30 subjects (5 females; mean age 47years) were randomized to WBph (n=16; 1 female) or 2RBCaph (n=14; 4 females). HFE mutations were p.C282/p.C282Y in 17 subjects, p.C282Y/p.H63D in 9, and p.H63D/p.H63D in 4. At baseline, mean Hb was 149g/L (SD 7.8) and median SF was 504ng/mL (IQR 406–620ng/mL). 222 procedures (WBph n=146, 2RBCaph n=76) were completed; 9 were interrupted (local hematoma, insufficient flow); 35 (16 WBph, 19 2RBCaph) were postponed because of low Hb and 15 for non medical reasons. There were 2 drop‐outs in the WBph arm due to depression and poor compliance, respectively. Anemia (Hb<130g/L in males, <120g/L in females) occurred after 15 visits in 8 WBph subjects and after 5 visits in 3 2RBCaph subjects. Fatigue was reported after 37 phlebotomies and 31 aphereses. Only 5 participants (17%) completed the study per protocol. 136 blood components (94 RBC concentrates and 42 plasma units) for transfusion were obtained. Overall, a median of 7.5 WBph (IQR 6.2–9.8) was needed to reach SF<100ng/mL, corresponding to 1.8 times of 2RBCaph (median 4.0, IQR 3.0–5.8) (P=0.0001). Analyzing separately p.C282/p.C282Y and p.C282Y/p.H63D carriers, the relation WBph to 2RBCaph was 1.6 and 1.8, respectively. Treatment arm and HFE mutation were the covariates with significant effect on the primary endpoint (P=0.0001 and 0.007, respectively).

Summary/Conclusions: 2RBCaph is more efficient than WBph for iron depletion in healthy subjects with HH or other HFE mutations and moderate hyperferritinemia. Intensive treatment schedules, generally recommended for HH, are difficult to keep because of Hb drop and compliance. Less intensive treatment in asymptomatic individuals with HH and their inclusion in blood donation would avoid negative effects on quality of life and benefit blood collection centers in the long term.

LONG‐ TERM COURSE OF HEMOGLOBIN AND FERRITIN VALUES IN HIGH‐FREQUENCY BLOOD DONORS DONATING WHOLE BLOOD OR DOUBLE ERYTHROCYTE APHERESIS

V Pehlic¹, T Volken², A Holbro^1,3, Z Jirout¹, B Drexler^1,3, A Buser^1,3, L Infanti^1,3

¹Blood Donation Centre Basel Switzerland, Basel ²School of Health Professions, Zurich University of Applied Sciences, Winterthur ³Division of hematology, University Hospital Basel, Basel, Switzerland

Background: Serum ferritin (SF) measurements in whole blood (WB) donors demonstrated that female sex and intensity of donation are major risk factors for iron deficiency. Approximately 200mL red blood cells (RBC) and 200–240mg iron are lost with WB donation. Double unit RBC (2RBC) collections of 360mL (ca. 40mL less than the RBC amount of two WB donations) lead to a loss of about 400mg iron. In Switzerland, the maximal allowed donation frequency for male donors is once every 6months for 2RBC and once every 3months for WB donation.

Aims: To describe and compare the course of hemoglobin (Hb) and SF in male subjects donating WB and 2RBC at our institution.

Methods: We included 294 WB and 151 2RBC donors (n=445) who donated with the maximal allowed donation frequency over 48months between 2008 and 2013, yielding 4,704 WB and 1,208 2RBC donations. We excluded subjects with hyperferritinemia and known HFE mutations. Hb limits were 135g/L for WB and 140g/L for 2RBC donation. With 2RBC apheresis 360mL RBC were collected. SF was measured on a predonation serum sample; Hb was determined from finger prick samples. The donors received no iron substitution. We used generalized estimating equation models for Hb and SF trajectories.

Results: Mean age at the first blood donation was 53 (WB) and 48years (2RBC), respectively.

At the first donation, mean Hb was 153g/L (SD 13) in WB and 159g/L (SD 8) in 2RBC donors; mean SF was 44 (SD 52) and 73μg/L (SD 56), respectively. On average, Hb and SF were higher in 2RBC donors (5.1g/L and 26μg/L, respectively; P<0.001). There were 137 subjects with SF<30μg/L in WB and 19 in 2RBC group, and 85 with SF<50μg/L (but>30μg/L) and 40, respectively.

In 2RBC donors, between the first and the last donation, mean Hb declined from 159g/L to 157g/L (P<0.05) and mean SF from 73μg/L to 66μg/L (ns). In WB donors, mean Hb dropped from 153g/L to 152g/L (P<0.05) and SF from 44μg/L to 35μg/L (P<0.001). Similar results were found when adjusting for age and season.

Hb values dropped from baseline until the 11^th donation for WB donors and until the 4^th donation for 2RBC donors with an upward trend thereafter. In both groups, no Hb value below the limits of blood donation and no anemia were observed.

SF reached a nadir at the 4^th donation in both WB and 2RBC donors (37μg/L and 60μg/L) and increased thereafter in 2RBC donors. In WB donors, SF followed a parabolic trend that peaked at the 10^th donation, and then declined until the last donation.

Summary/Conclusions: The maximal allowed blood donation frequency for WB and 2RBC male donors in Switzerland is not only protective for the development of anemia, but also for deferral of blood donors because of low Hb. This was observed even in subjects with low SF at baseline.

IMPACT OF HYDROXYETHYL STARCH AND MODIFIED FLUID GELATIN ON GRANULOCYTE PHENOTYPE AND FUNCTION

N Doblinger^1,2, A Bredthauer², M Mohrez¹, V Hähnel¹, B Graf², M Gruber², N Ahrens¹

¹Institute of Clinical Chemistry and Laboratory Medicine, Transfusion Medicine ²Department of Anesthesiology, University Hospital Regensburg, Regensburg, Germany

Background: Granulocyte concentrate transfusion is a potentially lifesaving option for patients without functional neutrophils. However, recent studies have failed to demonstrate the anticipated clinical effectiveness of this procedure. Granulocyte concentrates are manufactured using sedimentation agents to separate granulocytes from red blood cells and enhance granulocyte collection efficiency. High‐molecular‐weight hydroxyethyl starch (HES) is most commonly used for this. However, authorities recently restricted the use of HES due to its unfavorable risk‐benefit‐profile. Modified fluid gelatin (MFG) is an already used alternative sedimentation agent. As the granulocyte product contains these substances, any impact of the sedimentation agent on granulocyte function may affect the clinical effectiveness of granulocyte transfusion.

Aims: We tested the hypothesis that MFG is not inferior to HES in terms of the functionality and viability of granulocytes.

Methods: Granulocytes from ten healthy donors were isolated, aliquoted and incubated in parallel for 2hours with either 0% (control), 7.5%, 15% or 30% MFG (Gelafundin 4%, B. Braun Melsungen AG) or HES (Hespan 6%/450/0.7, B. Braun Medical Inc.), respectively, and granulocyte migration, chemotaxis, reactive oxygen species (ROS) production, neutrophil extracellular trap formation (NETosis), antigen expression of CD11b, CD62L and CD66b, and viability were subsequently investigated in vitro.

Testing was performed using live cell imaging of the cells embedded into a collagen I matrix for parallel testing of migration, ROS production and NETosis. In addition, flow cytometric (FACS) analysis was utilized for surface marker expression, viability and respiratory burst measurement.

Results: Granulocyte migration decreased in a dose‐dependent manner in response to HES and MFG. Relative to the controls, all three concentrations of HES lowered migration distances (P<0.001 respectively), whereas only the higher concentrations (15% and 30%) of MFG showed lower relative migration distances (P<0.001 respectively). Track straightness was reduced with both sedimentation agents at 15% and 30% to the same extent (P<0.001 respectively).

HES resulted in lower CD11b expression (P=0.028) and higher CD62L expression (P=0.007) compared to the controls, whereas the differences for CD66b did not reach statistical significance. MFG did not affect the expression of any investigated surface antigen mediating endothelial adhesion and transmigration in comparison to the controls.

No significant differences in the timing of ROS production or NETosis, or in neutrophil viability or respiratory burst were observed.

Summary/Conclusions: These results indicate that MFG is not inferior to HES in terms of granulocyte phenotype and function in vitro when used at equal concentrations, and that potential impairment of granulocyte function can occur with HES.

PLATELETPHERESIS DONATIONS AND DONOR RISK OF SUSTAINED THROMBOPENIA

J Py¹, M Barnoux²

¹Directeur Médical ²Responsable Prélèvement, EFS Centre Pays de la Loire, Saint Jean de la Ruelle, France

Background: Plateletpheresis donation leads to a well‐known transient decrease of donor's platelets. The question of long‐term effects raised with the development of regular donations by some donors in order to satisfy a growing demand. A seminal work (Lazarus, Transfusion, 2001) stated that there is a sustained thrombopenia in frequent plateletpheresis donors, correlated with the total number of donations.

Aims: French regulation authorizes up to 12 plateletpheresis donations per year, with a minimum 4weeks interval between them. We tried to evaluate the risk of sustained thrombopenia under these conditions.

Methods: We retrieved all plateletpheresis donations occurring between 01/01/2014 and 08/31/2018 from the French civilian blood donors’ base and then selected a cohort of donors with at least 24 donations during that period. In order to minimize measurement errors, platelet counts analysed were means of three consecutive donations, i.e. 8 measures for each donor.

Results: The cohort includes 2,186 donors (384 women and 1,802 men). Mean platelet counts fluctuate between 276.4 and 278.6 platelets/mL. Analysis of variance does not show any statistically significant difference (F=0.462), even taking donor's sex or age in consideration. There is no difference if we consider the total duration of the 24 donations, either. Donors with the lowest first counts show a significant rise in subsequent measures and donors with the highest counts show a decrease trend, exhibiting a classical regression toward the mean.

Summary/Conclusions: Plateletpheresis French regulation does not seem to be at risk of sustained donor thrombopenia. This conclusion is in agreement with recent literature data.

HLA IN BLOOD TRANSFUSION

Z Grubic

Tissue Typing Centre, University Hospital Centre Zagreb, Zagreb, Croatia

The primary biological role of the Human Leukocyte Antigen (HLA) system is the regulation of the immune response to foreign antigens. Because of this role, HLA genes and molecules have an important role in transplantation, etiology of many autoimmune, non‐autoimmune and infection diseases, but also in transfusion medicine.

An increasing probability of an HLA non‐compatible blood products, tissues or organs exists due to the extremely high polymorphism of HLA genes, with more than 20,000 described alleles to date, and their different frequency distribution in various worldwide populations.

The HLA system, originally discovered as a result of a transfusion reaction in the 1950s, can cause detrimental immune reactions in transfusion therapy. HLA antibodies present in the patient are responsible for some of these reactions, while in other cases HLA antibodies or HLA reactive cells present in the transfused product are accountable for the immunoreactivity. HLA antibodies form as a result of exposure to foreign HLA antigens during pregnancy, transplantations and blood transfusions and can cause platelet immune refractoriness, febrile transfusion reaction, transfusion‐related acute lung injury, and transfusion associated graft versus host disease. In order to avoid or reduce the development of these transfusion‐related events, HLA antibody negative or compatible products should be used.

Almost all existing methods presently used for molecular typing of HLA polymorphisms are based on polymerase chain reaction, but with different resolution levels (low resolution ‐ two digits or high resolution ‐ four digits). In addition to providing a more precise detection of polymorphisms at HLA classical loci (e.g. HLA‐A, ‐B, ‐C, ‐DRB1, ‐DQB1), molecular methods can also determine polymorphisms at HLA loci which previously could not be typed by serology (e.g. HLA‐DRB3, ‐DRB4, ‐DRB5, ‐DQA1, ‐DPA1).The most commonly used method for the detection of HLA antibodies was until recently complement‐dependent cytotoxicity (CDC) technique, but it is increasingly being replaced by a more sensitive, solid phase based method (Luminex technology). In conclusion, an accurate and precise determination of both HLA gene polymorphism and HLA antibodies presence is essential for the safe and efficient administration of transfusion products.

SURFACE PLASMON RESONANCE‐BASED DETECTION OF ANTI‐HPA 1A ANTIBODY GLYCOSYLATION, AN ASSAY TO PREDICT DISEASE SEVERITY IN ALLOIMMUNE CYTOPENIAS

Z Szittner¹, R Temming¹, D Schimdt¹, A Bentlage¹, R Visser¹, S Lissenberg‐Thunnissen¹, J Mok², W van Esch², M Sonneveld¹, E de Graaf¹, M de Haas^3,4, M Wuhrer⁵, E van der Schoot¹, G Vidarsson¹

¹Dept. Experimental Immunohematology ²Sanquin Reagents ³Department of Immunohematology Diagnostics, Sanquin Research, and Landsteiner Laboratory, Amsterdam ⁴Centre for Clinical Transfusion Research, Sanquin Research and Department of Immunohematology and Blood Transfusion of Leiden University Medical Centre ⁵Center for Proteomics and Metabolomics, Leiden University Medical Center, Leiden, Netherlands

Background: In only a minority of pregnancies complicated with anti‐HPA1a antibodies serious fetal/neonatal disease develops. The difficulty in predicting which mothers should be treated with IVIg hampers implementation of FNAIT screening. We found that Fc‐core fucosylation and galactosylation are highly variable in anti‐HPA1a IgG, and that these glycan features strongly affect binding to FcγRIIIa receptor. The level of Fc‐core fucosylation of anti‐HPA 1a alloantibodies was found to correlate with platelet count and outcome of the newborn, suggesting that antibody‐specific fucosylation might serve as a biomarker in FNAIT screening. However, at present the Fc‐glycosylation pattern can only be determined by complicated methods involving purification of the antigen‐specific IgG, and analyzing trypticly released ‐IgG‐derived‐ glycopeptides by tandem liquid chromatography‐mass‐spectrometry (MS) techniques. These methods, although powerful, are not yet suited for high throughput clinical screening.

Aims: Our aim was to provide a simplified method to quantify the biological activity of anti‐HPA‐1a antibodies, and possibly other alloantibodies against blood cells.

Methods: Here we explored if cellular surface plasmon resonance (SPR) imaging can replace MS, resulting in less complicated handling of patient sera and donor‐antigen‐bearing cells. The strength of the binding of platelets to FcγR on SPR sensor was monitored under flow. The SPR sensor was equipped with both WT FcγRIIIa (sensitive to Fc‐glycosylation status) and mutant FcγRIIIa‐N162A (insensitive to Fc‐glycosylation status). In addition, the biosensor was prepared with anti‐platelet CD61 (C17) and anti‐IgG to calibrate the number of injected platelet as well as to quantify IgG‐opsonization. The quality of the anti‐HPA 1a glycosylation was monitored as the ratio of the binding of opsonized platelets to the WT and the mutant N162A‐FcγRIIIa. Platelets opsonized with recombinant glycoengineered anti‐platelet antibodies with different levels of Fc‐fucosylation were used as standards. For validation, 166 plasma samples with anti‐ HPA1a antibodies, already analyzed by mass spectrometry and with known clinical outcome were tested (Sonneveld, BJH, 2016).

Results: We found that the ratio between the binding to the WT FcγRIIIa and to the mutant N162A‐FcγRIIIa correlated with the level of fucosylation of the HPA1a antibodies, as measured by mass‐spectrometry (r=−0.524; P<0.0001). Overall, a similar predictive value for disease severity was obtained as we previously reported for this retrospective cohort. In addition, quantitative information on antibody concentration can also be extracted using the FcγRIIIa‐N162A receptor as sensor on the chip, while anti‐IgG gave aspecific signals, presumably because it recognized cytophilic platelet‐FcγRIIa‐bound antibodies as well.

Summary/Conclusions: In conclusion, the combined use of WT and mutant FcγRIIIa in a label free SPR assay provides both quantitative and qualitative information of platelet bound anti‐HPA 1a antibodies, which circumvents the need for purification of specific antibodies and laborious mass spectrometric analysis. This approach might be generally applicable to determine the biological activity of cell bound antibodies not only for anti‐HPA1a in FNAIT, but also for anti‐RhD alloantibodies in HDFN or anti‐platelet antibodies in ITP.

RESULTS OF HPA‐1A SCREENING PROGRAM FOR IDENTIFICATION OF PREGNANT WOMEN AT RISK OF FOETAL/NEONATAL ALLOIMMUNE THROMBOCYTOPENIA (FNAIT)

M Uhrynowska¹, K Guz¹, A Orzińska¹, A Gierszon¹, S Purchla‐Szepioła¹, M Dębska², P Łopacz¹, I Kopeć³, K Maślanka¹, H Tiller⁴, A Husebekk⁵, R Dębski⁶, E Brojer¹

¹Department of Immunohematology and Transfusion Medicine, Institute of Hematology and Transfusion Medicine ²2nd Department of Obstetrics and Gynaecology, Medical Centre of Postgraduate Education ³Department of Hematology, Institute of Hematology and Transfusion Medicine, Warsaw, Poland ⁴Institute of Medical Biology; University of Tromsø The Arctic University of Norway ⁵Institute of Medical Biology, University of Tromsø The Arctic University of Norway, Department of Obstetrics and Gynecology, University Hospital North Norway, Tromso, Norway ⁶2nd Department of Obstetrics and Gynaecology, Medical Centre of Postgraduate Education, Warsaw, Poland

Background: Immunization against the human platelet HPA‐1a alloantigen is the most common cause of severe fetal and neonatal alloimmune thrombocytopenia (FNAIT) in otherwise healthy term newborns. The screening for HPA‐1a antigen in pregnant women is an important tool for identification of pregnantwomen at risk of having a fetus/neonate with FNAIT. Any targeted intervention depends on efficient screening methods as well as sensitiveand specific methods for detection of anti‐HPA‐1a. Within the framework of the Polish‐Norwegian project (PREVFNAIT) we have performed HPA‐1a screening program in Poland.

Aims: Our aim was to assess the frequency anti‐HPA‐1a antibody detection and the clinical outcome of newborns identified through the study. Women who joined the program due to the FNAIT in the previous child or in the current newborn are not analyzed in this study.

Methods: HPA‐1a screening of 24244 pregnant women in 8–40 gestational weeks was performed by FACS phenotyping or RQ‐PCR genotyping at IHTM in Warsaw. HPA‐1a negative/HPA‐1B/1B women were tested for HLA DRB3*01:01 and for anti‐HPA‐1a antibodies by MAIPA (followed up at week 17–20, 28, 32, 38–40 and 6weeks after delivery). If anti‐HPA‐1a were detected, quantitative MAIPA was performed.All HPA‐1a negative women were contacted for information concerning the newborn. If the baby had thrombocytopenia and anti‐HPA‐1a were not detected by MAIPA, the look back samples were tested retrospectively by PAKLx test (Immucor).

Results: 554 HPA‐1a negative women were identified (2.3%). Anti‐HPA‐1a was antibodies were detected by MAIPA in 44 women (two delivered tweens). Inaddition, anti‐HPA‐1a antibodies were later detected by PAKLx in further 3 women who delivered baby with severe thrombocytopenia and/or ICH. Total number of immunized mothers was 47 (8.5%). They delivered 49 babies; 35 were boys. Three women were treated by IVIg: two by 4 and 8 injections since 33^th and 26^th gw respectively. The anti‐HPA‐1a concentration in the 1^st one was 0.4; 0.1; 5.88IU/ml in 17, 28, 32 gw respectively and in the 2^nd <0.05IU/ml in all examined samples. The decision on treatment was based on the low PLT count ˜50G/L in the fetus in cordocentesis. Their newborns (one delivered tweens) were healthy. The 3^rd treated woman entered the program in 35 gw (anti‐HPA‐1a concentration was high 22.64IU/ml). She obtained one injection of IVIg. Her baby was born with mild thrombocytopenia with no ICH. Severe FNAIT occurred in 5/49 newborns: in 3 with anti‐HPA‐1a detected in PAKLx only and in 2 with antibody concentration in MAIPA ‐ 1^st: 6.86/12.91/23.36 at 28/32/38^th gw respectively; 2^nd: 8.85/17.58 at 32/38^th gw respectively. ICH was observed in all of them; PLT count was<50x10⁹ in four, 56×10⁹/ in one.

Summary/Conclusions: 1/ the severe thrombocytopenia due to anti‐HPA‐1a alloimmunisation in our prospective study occurred in 2/10000 pregnancies 2/ The PAKLx could improve anti‐HPA detection in the screening program and should be considered as an additional diagnostic test, if MAIPA result is negative 3/ The HPA‐1a alloimmunisation frequency is higher in pregnancies withmale than female fetus.

PLATELET ALLOIMMUNIZATION AND CLINICAL MANIFESTATIONS IN THE MATERNO‐FOETAL CONTEXT

C Martageix¹, F Bianchi¹, C Casale¹, C Chenet¹, N Ferré¹, N Francelle¹, J Quesne¹, T Granchon‐Riolzir¹, L Gilles¹, Y Mammasse¹, V Jallu¹, R Petermann^1,2

¹Platelet Immunology, INTS, F‐75015 Paris ²Centre de Recherche des Cordeliers, INSERM, USPC, Université Paris Descartes, Université Paris Diderot, F‐75006 Paris, France

Background: Foeto‐maternal platelet alloimmunization (FMPAI) is mainly characterized by foetal and / or neonatal thrombocytopenia (FNAIT), sometimes revealed by intracranial hemorrhage (ICH) or even by foetal death in utero (FDIU). The experience of the PNIL Milwaukee (USA) reported in 2014 that the diagnosis of alloimmunization was carried in only 33% of neonatal thrombocytopenia cases with a clinical symptomatology highly suggestive of an alloimmune etiology.

Aims: The aim of this two‐year study was i) to determine the frequency of platelet incompatibilities in FNAIT, ICH and FDIU and ii) to evaluate the frequency of detectable platelet alloantibodies (alloAb) and their specificity in cases of incompatibility.

Methods: Platelet genotyping was performed by HPA Beadchip Genotyping kit (BioArray Solutions, Immucor, Warren, NJ). Serology investigation was carried out by 3 different methods: Complete MAIPA Kit (apDia bvba, Turnhout, Belgium), Pack LxTM Assay (Immucor GTI Diagnostics, Waukesha, WI) and « in house» MAIPA. All 2017 and 2018 data were collected using the Laboratory Information Management System.

Results: 544 patient files were analyzed. No incompatibility is demonstrated in HPA‐1 to ‐9, ‐11 and ‐15 systems in 19.3% (n=105). HPA‐1 and / or 3 and / or 5 incompatibilities were found in 271 cases (49.8%), HPA‐2 and / or 15 in 87 cases (16%). Platelet alloimmunization was globally confirmed in only 10.6% of the cases. 58 platelet alloAbs were identified regardless of clinical manifestations: 24 anti‐HPA‐1a (41.4%), 20 anti‐HPA‐5b (34.5%), 8 anti‐CD109 (13.8%), 2 anti‐HPA‐5a and anti‐HPA‐15b (3.4% respectively) and 1 anti‐HPA‐1b and anti‐CD36 (1.7% respectively). 40 alloAbs were found in the context of neonatal thrombocytopenia, 6 in ICH and 10 in FDIU, and 1 in a follow‐up of pregnancy. Even if no anti‐HPA‐3 alloAb could be identified, the incompatibility in this system was highly associated with FNAIT, ICH and FDIU (n=55, n=32 and n=17 on 114 cases).

Summary/Conclusions: This study strongly confirmed the known immunogenicity of some HPA systems and highlighted overall the severity of HPA‐3 and HPA‐5 incompatibilities. The definite diagnosis of FMPAI is difficult to make due to the present technical difficulties in the detection of antibodies against the HPA‐3 and HPA‐15 systems. However, our results suggest that special attention should be paid to the management of pregnancies with these incompatibilities due to the frequency of severe foetal/neonatal adverse events.

FAST AND LOW‐COST MATERNAL HPA‐1A TYPING ELISA FOR HIGH‐THROUGHPUT SCREENING OF WOMEN AT RISK FOR FNAIT

D Winkelhorst^1,2, L Porcelijn³, E Muizelaar³, G Oldert³, E Huiskes¹, E van der Schoot¹

¹Immunohematology, Sanquin, Amsterdam ²Obstetrics, LUMC, Leiden ³Immunohaematology, Sanquin, Amsterdam, Netherlands

Background: Fetal and neonatal alloimmune thrombocytopenia (FNAIT) is a potentially life threatening disease caused by maternal alloantibody formation against fetal human platelet antigens (HPAs), of which anti‐HPA‐1a is accountable for the fast majority of the cases. Population‐based screening for FNAIT has been topic of debate for over decades. Logistically as well as financially, the major challenge of such a screening is the typing of pregnant women to recognize the 2% HPA‐1a negative women. At present, HPA‐1a typing is mostly done by genotyping. For cost‐effective implementation of anti‐HPA‐1a screening there is need for a high‐throughput, quick and low‐cost phenotyping assay.

Aims: The aim was to develop ahigh‐throughput, quick and low‐cost phenotyping assay in order to identify HPA‐1a negative pregnant women.

Methods: An automated sandwich ELISA was developed to perform HPA‐1a phenotyping using a murine monoclonal anti‐GPIIIa as coating antibody and horseradish‐peroxidase‐conjugated recombinant IgG1 anti‐HPA‐1a as detecting antibody. To ensure the applicability for high‐throughput testing in a potential screening setting, 20μl of the uppermost plasma of 3 – 6days‐old stored EDTA anticoagulated blood tubes was used, without first swirling or spinning them. In two phases, samples of pregnant women were tested and compared to an allelic discrimination polymerase chain reaction assay as golden standard. In the first phase, samples from unselected consecutive pregnant women were tested. The second phase was part of a prospective screening study in pregnancy and confirmatory genotyping was restricted to samples with an arbitrary set OD<0.500 in the HPA‐1a ELISA.

Results: The developed ELISA was optimized to require no additional handling (swirling or spinning) of stored tubes. During phase I, 506 consecutive samples were tested. In phase II, the HPA‐1a ELISA was performed in another 62,171 consecutive samples, with confirmatory Q‐PCR in 1,825. The two phases combined, samples from in total 1,585 HPA‐1a negative and 823 HPA‐1a positive pregnant women were genotyped. The assay reached a 100% sensitivity with a cut‐off OD between 0.075 and 0.200, leading to a specificity of 99.6%.

Summary/Conclusions: A quick, low‐cost and reliable assay for HPA‐1a phenotyping was developed that can be used in a population‐based screening setting to select samples that has to be tested for the presence of anti‐HPA1a antibodies. Because plasma from non‐mixed or spinned tubes of three to six day‐old samples can be used, this assay is applicable to settings with suboptimal conditions.

EVALUATING AN AUTOMATED CMV SCREENING ASSAY AT THE IBTS‐ A CHALLENGING PROCESS FOR BLOOD SCREENING LABORATORIES

D Coyne, D Butler, P Williams, N O'Flaherty

Virology, Irish Blood Transfusion Service, Dublin, Ireland

Background: Cytomegalovirus (CMV) sero‐prevalence in Ireland is lower than that which is reported in many other European countries. A study of 1047 pregnant women in 2002 found that 30.4% of Irish women were CMV seropositive in comparison to 56% from Western Europe and 92% Eastern Europe and 97% from Africa. An internal study carried out by the Irish Blood Transfusion Service (IBTS) in 2010 indicated the rate of CMV seropositivity in Irish Blood donors was 21.97%. Therefore a significant proportion of the Irish donor and recipient population are susceptible to primary CMV. This is of particular concern for patients for certain at‐risk groups such as very‐low birthweight CMV seronegative neonates, CMV seronegative patients undergoing transplantation and other CMV seronegative immunocompromised patients. This results in a demand for the provision of CMV sero‐negative blood components. In 2018 the IBTS evaluated the Abbott Alinity s CMV IgG assay as a replacement for the CMV Mastazyme EIA (Total AB EIA).

Aims: To assess the performance of the Abbott Alinity s CMV IgG screening assay in comparison to the CMV Mastazyme EIA (Total AB EIA).

Methods: Diagnostic sensitivity was determined by testing 48 confirmed CMV IgG positive donors from an external laboratory. Sensitivity was assessed using three seroconversion panels (n=54). Analytical sensitivity was calculated using linear regression analysis of the WHO first international standard for anti‐CMV IgG. Diagnostic specificity was determined by testing 6127 donors. Further evaluation of discordant results was carried out using the Architect anti‐CMV IgG and IgM assays and VIDAS anti‐CMV IgG and IgM assays.

Results: The diagnostic sensitivity of the Alinity s anti‐CMV IgG assay was determined to be 100%. The seroconversion sensitivity reported 42 out of 54 samples reactive. The analytical sensitivity of the Alinity s CMV IgG assay was determined to be 1.12IU/ml. The validation reported 65 discordant results from 6127 donor samples tested with both the Alinity s CMV IgG assay and the current Mastazyme total assay. 60 discordant results were observed (Alinity s anti‐CMV IgG positive/Mastazyme total negative). Further testing of these samples classified 27 discordant results as positive, 12 as negative and 21 as indeterminate. 5 discordant results were observed (Alinity s anti‐CMV IgG negative/Mastazyme total positive). Further testing classified these samples as negative. Overall thediagnostic specificity was determined to be 99.80%.

Summary/Conclusions: Both the seroconversion and analytical sensitivities are comparable between the Alinity s CMV IgG assay, the CMV Mastazyme Total AB assay, the Architect CMV IgG assay and the Vidas IgG assay. The slight variations can be attributed to the individual assay cut‐off definitions, which can vary greatly between CMV assays. It must be noted that the determination of the diagnostic specificity (99.80%) does not include indeterminate discordant results. Further testing will be carried out to try to characterize all discordant samples in collaboration with Abbott. This evaluation did not identify any donors with isolated confirmed CMV IgM antibodies in a pool of 6127 donors. Based on this evaluation the Abbott Alinity s CMV IgG assay is a suitable replacement to the Mastazyme total AB assay for blood donor screening.

DRIED PLASMA SPOT TESTING – THE ANSWER FOR MAKING BLOOD TRANSFUSION TESTING SAFER IN AFRICA?

C Pistorius

Virology, Western Cape Blood Service, Cape Town, South Africa

Background: Africa has a unique set of challenges regarding safe blood transfusion. Two of the largest contributing factors are: 1) The most common disease states in Sub‐Saharan Africa (SSA) require large amounts of blood as lifesaving interventions e.g. malaria, 2) the highest burden of infectious diseases transmissible through transfusion (Tapko, Toure, & Sambo, 2014) is found in SSA. This has often led to the binary donor base that exists in SSA, consisting of Voluntary Non‐remunerated Blood donors (VNBD) and family or replacement donors (FRD) as transfusion centres are unable to supply the demand when relying only on VNBD.

Voluntary non‐remunerated donors are the safest blood donors as they have no incentive (other than altruistic motives) and are not under social pressure to donate, both factors that may induce individuals knowing or suspecting themselves to be infected with a blood‐borne agent to donate blood.

Nucleic Acid Testing (NAT) in conjunction with serological testing is the gold standard for testing, however, the vast distances and high temperatures of Africa makes transport of traditional plasma samples a logistical challenge. Many publications evaluating the stability, suitability, and ease of use of dried blood spots (DBS) for NAT have been published. Generally, results have been shown to be comparable to traditional plasma samples. DBS is being used successfully in the early infant diagnosis (EID) programs for HIV by means of PCR testing, especially in Africa.

Aims:

1. To demonstrate that DBS and/or dried plasma spot (DPS) testing is suitable for blood donor screening and can make NAT testing more widely available in Africa

2. To determine the diagnostic sensitivity and specificity of testing DPS and DBS samples, in comparison to testing of plasma samples.

Methods: 900 negative new donor samples and 100 confirmed positive donor samples, as defined by routine blood safety screening done at Western Cape Blood Service, were screened using a dried blood spot kit. After routine testing was completed, one DBS sample and one DPS sample for each blood donor were prepared and analysed with the Ultrio Elite Assay on the Panther analyser.

Results: Invalid rate: 5 DBS invalids 0.56%

Summary/Conclusions: DBS/DPS can be used as a sample for screening blood donors as the invalid rate was 0.56%, and only found on DBS samples. Logistically DBS/DPS is well suited for the resource‐poor countries as samples are:

• ‐

  Easy to obtain (fingerpick samples could be used.)

• ‐

  Transport is simplified as samples will not leak or haemolyse due to high temperatures.

• ‐

  Samples can be stored at room temperature

DBS/DPS demonstrated acceptable specificity. The Ultrio Elite performed well with regards to HIV and HCV sensitivity. Sensitivity with regard to HBV was not as high but this could be due to very low and erratic viral loads.

EXPERIENCE WITH THE ALINITY S ASSAYS AFTER 5 MONTHS OF ROUTINE USE IN HIGH‐VOLUME BLOOD DONATION SCREENING

A Van Weert¹, M Koot², E Bakker¹

¹National Screening laboratory Sanquin ²Sanquin Diagnostics Virology and MAT Services, Sanquin Bloodsupply, Amsterdam, Netherlands

Background: Sanquin Blood Supply is responsible for the blood transfusion services in The Netherlands. At the National Screening laboratory Sanquin (NSS) annually more than 750.000 blood and plasma donations are tested, on average 3.000 samples per day. For more than 10years, infection serology testing was performed using the PRISM (Abbott Diagnostics), but since mid of July 2018, serological testing for the HBsAg, HIV Ag/Ab, Anti‐HCV and Anti‐HBc is done with Abbott's Alinity s system.

Aims: To compare the numbers of initially and repeatedly reactive results of whole blood and plasma donation samples and the rate of non‐specific results leading to deferral of donations and donors for PRISM and Alinity s assays using data from 6months before and 5months after implementation of the Alinity s systems at NSS.

Methods: Initial and repeat reactive rate of the assays run by either PRISM (HBsAg, HIV O Plus, HCV) or Alinity s (HBsAg, HIV Ag/Ab Combo, Anti‐HCV,) were calculated for January to June 2018 (PRISM) and August to December 2018 (Alinity s). Due to the lack of a true confirmatory method for Anti‐HBc, we only compared the rate of repeatedly reactive results for PRISM HBc and Alinity s Anti‐HBc.

Results: The rate of repeat reactive results for PRISM (P) and Alinity s (A) assays were as follows: 1) HBsAg P 0.01 % (42/390.736) versus A 0.03 % (87/322.394); 2) HIV P 0.07 % (262/390.735) versus A 0.06 % (201/322.470); 3) Anti‐HCV P 0.11 % (427/390.737) versus A 0.03 % (109/322.476). The rate of anti‐HBc reactive samples was not significantly different between PRISM (0.39 %) and Alinity s (0.42%). Over the study period, the rate of initially reactive samples for the three main screening assays (HBsAg, HIV, HCV) was also comparable between Alinity s (0.39 %) and PRISM assays (0.34 %), mainly attributable to a rather high number of initially reactive Alinity s HIV Ag/Ab results. This was due to initial issues with blood collection tubes that were resolved. As a result in December, the rate of initially reactive samples decreased to 0.16 %, which was significantly lower than for the three PRISM assays (0.33%).

Summary/Conclusions: The introduction of the Alinity s assays lead to a decrease of the average repeat reactive test results (HBsAg, HIV, HCV) by 0.17 % as compared to the PRISM, mainly due to a lower false reactive rate of the Alinity s Anti‐HCV assay. This will be further investigated for first time and multiple time donors. With the implementation of the Alinity s at Sanquin we aimed to improve not only the operational efficiency but also to further minimize unjustified disapproval of donors. These first data show that the low initial and repeat reactive rates of the Alinity s assays indeed have a positive impact on unnecessary deferrals of donations and donors.

EVALUATION OF THE USABILITY OF THE NEWLY LAUNCHED ALINITY S AND THE SPECIFICITY OF HIV COMBO, ANTI‐HCV, HBSAG AND SYPHILIS IN A BLOOD DONOR SCREENING SETTING

C Tinguely, M Hotz, C Niederhauser

Infektmarker, Interregionale Blutspende SRK AG, Bern, Switzerland

Background: In blood banks, testing all blood donations for markers of infectious diseases plays an important role in maintaining the safety of blood transfusions. Mandatory serological testing in Switzerland is performed for anti‐HCV, HIV Ag/Ab, HBsAg and Syphilis. Highly specific and sensitive tests with corresponding automation are essential for this purpose.

Aims: A comparative study was carried out to evaluate the usability of the newly launched Alinity s System (Abbott) and the specificity of the infectious disease parameters HBsAg, Anti‐HCV, HIV Combo and Syphilis (Abbott) with the currently used ELISA methods on the Quadriga BeFree System (all DiaSorin, formerly Siemens Healthcare Diagnostics).

Methods: The study took place at the Interregional Blood Transfusion Service in Berne, Switzerland. The specificity of the parameters was studied on 2,748 blood donor sera from both first time and repeat donors. The samples were tested first on the Quadriga Be Free System with Enzygnost HBsAg 6.0, Enzygnost Anti‐HCV 4.0, and Enzygnost HIV Integral 4 assays and on the PK7300 with the newbio‐pk TPHA assay (Newmarket Biomedical). All samples were retested on the same day with HBsAg, Anti‐HCV, HIV Combo and Syphilis on the Alinity s. Initial reactive samples were repeated in duplicate. Discriminatory tests were carried out for repeatedly reactive samples using alternative screening tests and neutralisation (for HBsAg) on an Abbott Architect i1000 system and immunoblots (HIV‐, HCV‐, Syphilis‐ INNO‐LIA, Fujirebio). For all samples, results from our routine individual donation nucleic acid testing (HCV, HIV, HBV, Roche cobas 8800 system) were available.

Results: Based on the results from testing 2,748 blood donations, the observed specificities of Alinity s assays (A) and Enzygnost assays (E) are comparable: % specificity / 95 % confidence interval: HBsAg 99.96 / 99.80 – 100.00 (A), 99.67 / 99.38 – 99.85 (E), HCV 100.00 / 99.87 – 100.00 (A), 100.00 / 99.87 – 100 (E), HIV 99.89 / 99.68 – 99.98 (A), 99.89 / 99.68 – 99.98 (E), Syphilis 99.89 / 99.68 – 99.98 (A) and 100.00 / 99.87 – 100.00 for TPHA. The initial reactive rates (IRR %) were also comparable, % IR: HBsAg 0.15 (A), 0.44 (E), HCV 0.00 (A), 0.00 (E), HIV 0.18 (A), 0.11 (E), Syphilis 0.11 (A), 0.00 (TPHA).

Summary/Conclusions: The Alinity s System was easy to use by the operators after a very short introductory training and provides good operational efficiency such as high throughput even when selective testing for samples is needed. The observed specificity of Abbott Alinity s versus Siemens Enzygnost assays is comparable in a blood donor screening setting. Unfortunately, we were not able to analyse statistically the specificity data due to the insufficient number of donor samples tested in parallel. It is worth mentioning that around 90% of the samples included in the study derived from repeat donors who had been previously tested with the Enzygnost assays but were “first time donors” for the Alinity s assays. All four assays from both systems exhibit a very good specificity and are highly suitable and practicable for routine blood donor screening.

EVALUATING THE CLINICAL PERFORMANCE OF ELECSYS^® INFECTIOUS DISEASE PARAMETERS ON THE COBAS E 801 ANALYSER FOR ROUTINE FIRST‐TIME BLOOD DONOR SCREENING

C Maugard¹, J Relave¹, M Ahne², M Klinkicht², C Fabra³, F Langen²

¹Etablissement Français du Sang, Occitanie, Montpellier, France ²Roche Diagnostics GmbH, Penzberg, Germany ³Etablissement Français du Sang, Provence Alpes‐Côte d'Azur, Marseille, France

Background: Effective screening for transfusion‐transmissible infections is essential to ensure safe blood transfusions. The World Health Organization recommends mandatory serological testing of blood donations for human immunodeficiency virus (HIV), hepatitis B (HBV)/C (HCV), and syphilis. Due to increasing demands on clinical laboratories, there is a need for reliable and accurate automated blood screening tests. The fully automated cobas e 801 analyser can be used with Elecsys^® infectious disease parameters to screen donor blood samples.

Aims: To compare the performance of Elecsys^® infectious disease parameters on the cobas e 801 analyser (Roche Diagnostics) with other commercially available assays for routine first‐time blood donor screening.

Methods: We provide results from Etablissement Français du Sang (Montpellier), a blood bank which participated in a large, multicentre study of the cobas e 801 analyser. The following infectious disease marker assays were compared: HIV, Elecsys^® HIV Duo versus PRISM HIV O Plus; HCV, Elecsys^® Anti‐HCV II versus PRISM HCV; HBV surface antigen (HBsAg), Elecsys^® HBsAg II versus PRISM HBsAg; HBV core antigen antibodies (anti‐HBc), Elecsys^® Anti‐HBc II versus PRISM HBcore; syphilis, Elecsys^® Syphilis versus newbio pk TPHA assay. Specificity was tested using residual fresh serum samples from unselected first‐time blood donors, and calculated according to assay package inserts and site‐specific cutoffs. Samples were tested using comparator assays, then retested the same day using Elecsys^® assays. Initially reactive samples were repeated in duplicate; confirmatory tests were conducted on repeatedly reactive samples. Confirmatory tests: HIV, nucleic acid testing (NAT), ARCHITECT HIV Ag/Ab and INNO‐LIA^® HIV I/II Score assays; HCV, NAT, ARCHITECT HCV and INNO‐LIA^® HCV Score assays; HBsAg, NAT, ARCHITECT HBsAg and Elecsys^®/PRISM HBsAg Confirmatory assays; anti‐HBc, NAT, HBsAg, Anti‐HBs, and ARCHITECT Anti‐HBc assays; syphilis, ARCHITECT Syphilis TP and INNO‐LIA^® Syphilis Score assays. Sensitivity was tested using 30 preselected, anonymised, positive, citrate‐phosphate‐dextrose‐plasma samples (Plasmatec Laboratory Products) and compared with archived data for comparator assays. Sensitivity was calculated according to the final NAT result.

Results: Across all infectious disease markers, specificity to detect repeatedly reactive samples using Elecsys^® versus comparator assays was similar (99.81–100.00% versus 99.79–99.98%; n≥5195). In specificity analyses, there were 14 discrepant results for HIV testing, 27 for HCV, two for HBsAg, eight for anti‐HBc, and five for syphilis. Sensitivity of the Elecsys^® HIV Duo assay (83.33%; 95% CI 65.28–94.36) was higher than the PRISM HIV O Plus assay (76.67%; 95% CI 57.72–90.07), but the difference was not statistically significant. Sensitivities of Elecsys^® and comparator assays were the same for HCV (85.19%; 95% CI 67.33–95.97), HBsAg (70.00%; 95% CI 50.60–85.27), anti‐HBc (100.00%; 95% CI 85.75–100.00), and syphilis (100.00%; 95% CI 88.43–100.00); three HCV and six anti‐HBc samples were classified negative/indeterminate and excluded from the analyses. In sensitivity analyses, there were two discrepant results for HIV testing, three for HCV, and five for anti‐HBc.

Summary/Conclusions: Elecsys^® infectious disease parameters on the cobas e 801 analyser demonstrate high specificity/sensitivity for screening first‐time blood donor samples, with similar clinical performance to other commercially available assays.

SIDE BY SIDE COMPARISON OF SPECIFICITY OF ALINITY S AND COBAS E801 SEROLOGICAL ASSAYS IN SERUM AND PLASMA SAMPLES

EM Wagner¹, B Gindl¹, R Ilk², R Rolka¹, M Mach³, G Driesel⁴, C Schmitt⁴, D Jahn⁴, B Baumann‐Baretti⁴

¹Plasma Analytics ²Statistics and Six Sigma ³Quality Assurance BioLife Europe, Takeda, Vienna, Austria ⁴Haema AG, Berlin, Germany

Background: Individual plasma and serum specimens from whole blood or plasmapheresis donors are tested for absence of infectious agents by serological assays prior to use for transfusion or production of blood derived therapeutics. The Department of Plasma Analytics (PA), Takeda (Austria), and Haema AG, Grifols (Germany), both labs with high throughput and a high level of automation, were seeking for alternatives to replace their current serological test systems (Abbott PRISM Next).

Aims: To allow a direct comparison of the two final candidate analyzers Alinity s (Abbott) and cobas e801 (Roche Diagnostics GmbH), a side by side evaluation was carried out by the PA and Haema with support from Abbott and Roche (provision of instruments and reagents). The aim was to compare assay specificities as well as handling and performance of the instruments. The outcome should be used to better understand potential specificity differences and practical handling aspects (throughput, etc.) of a next generation serological analyzer.

Methods: The two candidate instruments were installed in the PA. From March to June 2018, close to 10,000 aliquots from routine preselected repeat donors, provided by Haema, were run on both study instruments in parallel. Plasma samples were tested for HBs antigen (Ag), HCV antibody (Ab), HIV Ag/Ab, and partially for Syphilis Ab. Serum samples were additionally tested on HBc Ab. Samples with repeat reactive results (“RR”, two reactive results out of three tests) not confirmed by confirmatory tests were counted as false reactive. The necessary sample size was calculated based on a one‐sided comparison of proportions with the aim to detect potential specificity differences (a=10%) in the size of those specified by the manufacturers’ instructions. Two different lots were tested for the three main assays.

Results: Out of 7,389 plasma and 2,242 serum samples, 41 test results representing 37 individual donations were found RR on one or both instruments. Two samples were confirmed positive (1×HBsAg, 1×HCV), two others were indeterminate. The sample containing low level antibodies against HCV was PCR negative and only detected by the Roche system.

The percentage of false reactive results for the five assays on the two systems were (Alinity s/e801): HBs Ag: 0.04/0.03 % in a total of 9590/9589 samples tested; HCV Ab: 0.01/0.09 % in 9620/9585, P<10%; HIV Ag/Ab: 0.03/0.09% in 9362/9329, P<10%; Syphilis Ab: 0.1/0.07% in 2971/2987; HBc: 0/0% in 1531/1549. No significant difference was found between the calculated specificities in our study and the manufacturers’ data.

A potential influence of sample matrix and kit lots was assessed. A trend towards more false reactive results in serum vs plasma was found for nearly all assays. No clear‐cut statistical difference was seen between lots.

Summary/Conclusions: The study results are in line with the manufacturers’ specificity data, showing that the Alinity s HCV Ab and HIV Ag/Ab assay show a slightly higher specificity in a population of plasma and serum samples from repeat donors prescreened by PRISM. A possible influence on the test specificity by the sample matrix was detected but needs further investigation.

TUMOUR SPECIFIC CELLULAR THERAPIES

H Laubli

No abstract available.

CELLULAR THERAPIES IN THE AGE OF CRISPR‐CAS

T Cathomen

Institute for transfusion medicine and gene therapy, Medical Center ‐ University of Freiburg, Freiburg, Germany

The possibility to edit complex genomes in a targeted fashion has not only revolutionized basic research but biotechnological and therapeutic applications as well. With the rapid development of genome editing tools, in particular zinc‐finger nucleases (ZFNs), transcription activator‐like effector nucleases (TALENs), and the CRISPR‐Cas system, a wide range of therapeutic options have been – and will be – developed at an unprecedented speed. Therapeutic genome editing in hematopoietic cells enable new interventions in the blood and immune system, including novel approaches to treat immunological disorders, infectious diseases, and cancer. We have developed GMP‐compliant protocols to manufacture gene edited CD34+hematopoietic stem and precursor cells (HSPCs) as well as chimeric antigen receptor (CAR) T cells, with the final goal to provide novel cell therapies for patients suffering from primary immunodeficiencies, chronic infection with human immunodeficiency virus type 1 (HIV‐1), and some tumor entities. Despite great success in improving their specificity, engineered designer nucleases can induce genotoxic side effects by introducing mutations or chromosomal aberrations. We have established novel genome‐wide assays that enable us to detect chromosomal aberrations induced not only by off‐target activity but also by on‐target activity, such as micro‐aberrations and translocations, with unparalleled sensitivity. In toto, our developed protocols allow us to achieve genome editing in hematopoietic cells with high efficiency and to assess the genotoxic risk associated with the expression of CRISPR‐Cas nucleases and TALENs in clinically relevant human cells, so forming the basis for planned phase I/II clinical studies.

T CELL CELLULAR THERAPIES

MC Wolkers

Hematopoiesis, Sanquin Research, Amsterdam, Netherlands

Adoptive T cell therapy (ACT) has proven a potent means to treat blood‐borne tumors and solid tumors. Adoptive cell therapies include T cells that are genetically engineered with tumor specific T cell receptors (TCRs), or with chimeric antigen receptors (CARs). In addition, tumor infiltrating cells (TILs) can be isolated from tumor lesions, which are then expanded and reprogrammed in vitro prior to transfusion into the patient.

The anti‐tumoral efficacy of ACT products depends on several parameters, including the capacity of CD8⁺ T cells to produce cytokines, chemokines and granzymes, a feature that is critical for effective anti‐tumoral responses. Here I will discuss our efforts to develop and improve ACT products for future clinical use. I will present pre‐clinical work on developing TIL therapy for non‐small cell lung cancers. In addition, I will show that human CD8⁺ T cells can be divided into different subsets, and that only one of those subsets is highly cytotoxic. This finding may help improve the quality of genetically engineered T cell products, like TCR and CAR T cell products.

MOVING TOWARDS VOLUNTARY NON‐REMUNERATED BLOOD DONATION IN THE BALTIC STATES: A COMMON STARTING POINT DOES NOT GUARANTEE THE SAME RESULTS

R Kullaste¹, E Vilutyt≐², E Pole³

¹North Estonia Medical Centre's Blood Centre, Tallinn, Estonia ²National Blood Centre, Vilnius, Lithuania ³State Blood donor Centre, Riga, Latvia

Background: The Baltic States – Estonia, Latvia and Lithuania have a lot in common. We are located side by side, share the Baltic Sea as a gate to the West, and more importantly, a common history. We were members of the USSR and suffered 50years of Soviet occupation. We held hands in a 600km long human …chain” across the three States to express our mutual support, and later on, even joined the European Union on the very same day – June 1^st, 2004.

The three differ a bit in size, population and more in the languages spoken in each one, but that does not explain why the path towards voluntary unpaid donation varies as it does.

Aim: The aim is to describe the journey towards voluntary non‐remunerated blood donation in the Baltic States after regaining independence from the Soviet Union.

Methods: The information was collected from published and unpublished memories, annual reports and written interviews with Latvian and Lithuanian colleagues.

Results: In Soviet times, all orders came from Moscow and quality control was conducted from the capital city of Latvia, Riga. Donors were mostly paid and given an extra vacation day. Big factories were the best places to collect blood and people were queuing to donate. In 1991, the Soviet Union fell apart and the Baltic States suddenly got the freedom and responsibility to decide.

In Estonia the first edition of “Guidelines for the Preparation, Use and Quality Assurance of Blood Components” was taken as guidance in 1992. A lot of advice came from Finnish colleagues. In 1997, it was decided to move towards non‐paid voluntary donations. The process took 6years. The first couple of years were economically difficult for the reborn state, as money had less value than food. Instead of cash, donors were given rapeseed oil, sugar and pasta, for example. As the situation improved, food items were replaced by small symbolic gifts that carry sentimental value. It has been this way for more than 15years by now.

In Lithuania, the process started later, the first program for developing a framework for voluntary non‐remunerated donations being carried out in 2006–2015. It resulted in 51% of the donations being unpaid. The second program initiated in 2016 is still ongoing, aiming towards 100% non‐remunerated donations by 2020. By the end of 2018, they had reached 99.2%. In the beginning, the main obstacle was a private blood center creating unfair market conditions.

In Latvia, monetary compensation for blood donations still exists, but the younger generation has been encouraged to donate blood for free and some results can already be seen.

Summary/Conclusions: A common starting point does not guarantee the same results, at least not at the exact same time.

Examining the circumstances leading to the different outcomes could benefit countries yet to start moving towards non‐remunerated donations as well as those considering the opposite.

DEMOGRAPHIC AND HEMATOLOGICAL PARAMETERS IN FIRST TIME BLOOD DONORS

K Magnussen¹, S Zykova^1,2

¹Blood Centre and medical biochemistry, Innlandet Hospital Trust, Lillehammer ²Institute of Clinical Medicine, UiT‐The Arctic University of Norway, Tromso, Norway

Background: Little information is available on demographic and hematological parameters in first time donors.

Aims: To describe the distribution of age, gender, hematological parameters and ferritin of first time donors.

Methods: During 01.10.13–31.01.17 16 803 first time donors had routine samples drawn at the Capital Region Blood Centre, Denmark. Full blood count (FBC) was measured using a Sysmex XE‐2100D. Ferritin was measured on Ortho Vitros 3600 or 5600, in the sample also used for testing viral‐markers. Missing values for FBC were 4.7% and for ferritin 1.3%.

Results: Age (years): 23.9% <20 (Group 1; 64.2% females), 50.9% 20–29 (Group 2; 57.9% females); 13.6% 30–39 (Group 3; 50.9% females), 8.3% 40–49 (Group 4; 54.5% females) and 3.3% 50–61 (Group 5; 55.4% females).

Haemoglobin (Hb) was as expected significantly different between women and men (mean±SEM: 13.8±0.01 vs 15.5±0.01g/dl; P<0.000). Percentage of females with low Hb<12.5g/dl were 7.1%, 4.9%, 5.2%, 6.7% and 4.2%, percentage of males with Hb<13.5g/dl were 0.5%, 0.6%, 0.7%, 1.7% and 1.2% for the age groups 1–5 respectively.

Ferritin values were higher in males compared to females (median; 25^th‐75^th %>tile: 51;31–79 vs 157;106–231μg/l; P<0.000) and in older age groups compared to younger age groups (median; range in age groups 1–5 in females: 40;3–702, 52;6–619, 60;6–480, 60;8–725, 98; 10–604 and in males: 99;8–661, 161;18–1080, 206;15–1090, 220;26–1430, 221;33–981 respectively). Percentage of females with ferritin≤15μg/l were 8.6%, 3.8%, 3.7%, 6.7% and 2.0%, while percentage of males with ferritin≤15μg/l were 0.2%, 0.0%, 0.1%, 0.0% and 0.0% for the age groups 1–5 respectively.

White blood cell counts (WBC) were slightly higher in females compared to males (mean±SEM: 7.2±0.02 vs 6.6±0.03; P<0.000). Percentage of females with WBC>12x10⁹/L were 1.8%, 1.3%, 1.7%, 1.4% and 0.3%, while percentage of males with WBC>12x10⁹/L were 1.4%, 0.9%, 0.5%, 1.2% and 0.9% for the age groups 1–5 respectively. None had WBC<2x10⁹/L.

Platelet counts (PLT) were higher in females compared to males (mean±SEM: 257±0.6 vs 222±0.6; P<0.000).Percentage of females with PLT<150x10⁹/L were 1.0%, 1.5%, 2.2%, 2.4% and 1.0%, while percentage of males with PLT<150x10⁹/L were 2.6%, 3.6%, 2.7%, 2.4% and 1.6% for the age groups 1–5 respectively. Among the low PLT counts most were caused by EDTA‐dependent Pseudothrombocytopenia. Extreme deviations from normality were seldom and referred to GPs for further investigations.

Summary/Conclusions: First time donors are young with 75% younger than 30yearsof ageand the female/male ratio was 58/42. Of the16 583 first time donors with data on ferritin available,14% had low ferritin (≤ 30μg/l). The typical male first time donors neither had low Hb nor low ferritin, even with a significantly lower ferritin in younger donors. In female first time donors the prevalence of low Hb (6%<12.5g/dl) and low iron stores (23%≤30μg/l) is high. In all, while all first time donors are highly appreciated, campaigns could target the male population to even out the gender imbalance. Blood Centers must be aware of the higher prevalence of low iron stores in the youngest donors.

POST‐DONATION INFORMATION MANAGEMENT ‐ CONTRIBUTION TO THE SAFETY OF TRANSFUSION TREATMENT

T Vuk, J Ljubičić, J Gulan Harcet, T Očić, I Jukić

Croatian Institute of Transfusion Medicine, Zagreb, Croatia

Background: The aim of assessing suitability of prospective blood donors is protection of their health and the safety of transfused patients. Selection process is not always effective in obtaining all relevant information from blood donors in a timely manner. For several reasons, some risks remain undetected or they are disclosed at a future donation(s). Therefore, recording and management of post‐donation information (PDI) are of great importance for improvement of transfusion safety, donor counselling and education as well as overall improvement of the selection process.

Aims: The aim of the study was to present results of PDI management at Croatian Institute of Transfusion Medicine (CITM) and the effect of education activities on their trends.

Methods: We have analyzed reports on PDI recorded in two‐year period (2017–2018), according to the types of information obtained, age and sex of blood donors, total number of their donations preceding PDI, and the time of receiving the information. The effect of an information leaflet on PDI launched in November 2017 was assessed by comparing results in two study years.

Results: A total of 491 PDI were recorded: 266 in 2017 (1/396 donations) and 225 in 2018 (1/452 donations) with the following distribution: nonsexual risk as tattoo and piercing (17.5%), surgical procedures (17.1%), travel history (16.1%), infections/contact (15.3%), other medical reasons (13.4%), endoscopy/invasive diagnostic procedures (12.2%), malignancy (4.3%), autoimmune diseases (3.7%) and sexual risks (0.4%). Majority (76.4%) were late PDI, revealed on the future donation(s): 58.7% on the first next donation, 28.9% on the second and 12.3% after more than 2 subsequent donations. The mean age of blood donors associated with PDI was 40±12years (median 39years), while the mean age of all donors in 2017/2018 was 38years (median 37years). Of all PDI, 81.1% were related to male donors (84% in total pool of CITM donors). Using Chi‐square test there were no significant difference between female and male donors in total PDI frequency and in their distribution to early and late PDI (P>0.05). The median number of all donations preceding PDI was 6 for female donors and 19 for male donors. Implementation of education leaflet for blood donors resulted in 15.4% reduction of PDI in 2018 compared with 2017 (P>0.05). The effect is more pronounced (P<0.05) when comparing second and first half of 2018 (‐25.6%). Reduction is observed in all types of PDI with the exception of infections/contact (because they are mostly early PDI) and malignant diseases. The share of early PDI increased from 21.1% in 2017 to 26.7% in 2018, which may suggest better awareness of blood donors on the importance to inform blood bank on changes in their health status.

Summary/Conclusions: Our study points to the importance of systematic recording and management of PDI, including education of blood donors about the need of providing all relevant facts related to their health and the safety of donated blood in a timely manner. We are planning further improvements by providing information on this topic on posters and screens on donation sites.

THE DEXTOR PROJECT: DATA EXCHANGE IN TRANSFUSION, OPEN RESOURCE

BI Whitaker¹, K Land², P Ashford³, K Chada¹, R Sylvester⁴, L Lodge⁵, J Wiersum⁶, A Ryder⁷, P Distler⁸

¹Office of Biostatistics and Epidemiology/CBER, U.S. Food & Drug Administration, Silver Spring, MD ²Clinical Services, Vitalant, Tempe, AZ, United States ³ICCBBA, Minster on Sea, United Kingdom ⁴America's Blood Centers, Washington, DC, United States ⁵Scottish National Blood Transfusion Service, Edinburgh, Scotland, United Kingdom ⁶TRIP, Leiden, Netherlands ⁷University of Tennessee Health Science Center, Memphis, TN ⁸ICCBBA, Redlands, CA, United States

Background: Currently, the transfer of data between organizations and/or computer systems is very limited, and where present is typically proprietary. In the absence of a standardized reference format individual organizations and vendors attempting to integrate disparate databases must develop unique solutions. Aggregation of information from multiple sources is complex and costly, constituting a significant barrier to effective analysis of data to improve practice and inform policy.

Aims: To standardize the definitions and facilitate integration of key data items used in blood donation and transfusion. We report here on an initial effort to map internationally harmonized critical steps in the blood collection/donation process in order to test the approach.

Methods: Through a collaborative process of serial conference calls and correspondence, an informal multi‐national consortium of experts across the transfusion industry are attempting to create a vocabulary with sufficiently precise definitions to be usable by automated systems and that can be the foundation of a Blood Collection/Transfusion Medicine Common Data Model (CDM), using the following steps:

• ‐

  Define the scope of activity to be addressed and segment into key processes.

• ‐

  Identify the set of data elements in each segment that are common to all systems.

• ‐

  Review and consider existing standards and definitions for each data element.

• ‐

  Develop draft definitions for each data element.

• ‐

  Release draft to public domain for critical review and refinement with long-term goal of gaining widespread endorsement.

Results: A standardized approach to blood donation was mapped through identification of common pathways and core mappable data elements. Denominator data associated with donor characteristics and blood collection was selected as the first segment to address. A dictionary (or vocabulary) of common terms has been created and will be presented for international comment.

Summary/Conclusions: Developing an international consensus on the core elements and their definitions across the transfusion chain is critical for data integration and automation efforts. The expected benefits of this endeavor include that it allows the establishment of algorithms to automate reporting and thus reduce hands‐on staff time; reduces time and resources needed to integrate new databases; allows systems to continue to use existing concepts and definitions internally while also providing data output in a standardized format; supports the ability to consistently analyse, interpret and present information regardless of the data source; establishes data definitions against which new systems can be developed; helps to improve comparability of results by providing a common data model for researchers and policy makers; improves confidence in data integrity and reliability of the derived information as a basis for rational decision making; and reduces data gathering effort and cost thus improving opportunities for more efficient/complex data analysis.

Standardizing the transfusion medicine dataset is the first step in achieving the automation of data transfer and analysis needed globally to drive patient safety, research innovation, and best business practices. Further steps must address the precise methods of data exchange, identification of responsible entities for maintenance and further development, and engagement of computer system developers.

RECIPIENT FACTORS INFLUENCING RED BLOOD CELL ALLOIMMUNIZATION

J Hendrickson

Laboratory Medicine, Yale University, New Haven, United States

Red blood cell (RBC) alloantibodies develop in a subset of individuals following exposure to non‐self RBCs through transfusion, pregnancy, or other activities; these antibodies can lead to difficulty locating compatible RBCs, acute or delayed hemolytic transfusion reactions, or hemolytic disease of the newborn. Alloimmunization is underestimated due in part to antibody evanescence, the random nature of post‐transfusion antibody screens, fragmented medical care, and the lack of widespread antibody registries. Factors that influence who will develop detectable alloantibodies are not well understood. Transfusion burden is one risk factor for alloimmunization, though many highly transfused individuals never form alloantibodies despite exposure to many RBC units (and many non‐self ABO blood group antigens). Individuals with sickle cell disease (SCD) and myelodysplastic syndrome (MDS) are more likely to form RBC alloantibodies than most other patient populations. Individuals with rheumatologic and other forms of autoimmunity, though not chronically transfused, are also at higher than average risk of forming RBC alloantibodies. Inflammation, in a broad sense, is one common thread among these diagnoses associated with high prevalence rates of RBC alloimmunization. Reductionist murine models support some types of inflammation (including viral‐like stimuli) around the time of RBC exposure as being associated with an increased likelihood of alloantibody formation. Strategies other than transfusion avoidance or extended antigen matching beyond ABO/Rh would be beneficial to prevent new RBC alloantibody formation, especially in patients at highest risk.

FREQUENCY OF RED CELL ALLOIMMUNISATION IN PATIENTS WITH SICKLE CELL DISEASE IN OMAN

H Al Balushi, W Oudeh

Haematology and Blood transfusion, Royal Hospital, Muscat, Oman

Background: The unique genetic makeup of the Omani population makes them rich in the genetic blood disorder. 45 % of Omani populations are −α/−α gene carriers, 44% −α/αα, and 11% of the population are αα/αα. Around 10 % of Omani nationals carry the gene for HbS, and 2 – 3 % carry the gene for β‐thalassaemia. Recent statistics show that there are around 400 patients with thalassaemia major and 3000 with SCD in Oman. The other RBC abnormality that is common in Oman is G6PD deficiency which is found in 28 % of males and 12 % of females. Omanis are known to have the highest frequency of α thalassaemia and G6PD reported so far in any race. Although blood transfusion is one of the supporting treatments of SCD, it can cause some serious complications for the patients. Alloimmunization of red blood cells is one of the consequences of blood transfusion. Alloimmunisation of the RBCS can cause haemolytic transfusion reactions and may trigger hyperhaemolysis, in which transfused and patient's own RBCs are destroyed. Alloantibodies can cause delay in the process of transfusion, it can be costly and time consuming. High number of patients developing alloantibodies may indicate a major difference in the patient and donor population. It may also indicate lack of a controlled, generalised sickle patients management policy. In Oman the decision of transfusing SCD patient is left to physicians attending the patient.

Aims: This study is aimed to highlight the increasing number of alloimmunised sickle cell patients. In the Royal Hospital we get 40 new cases of sickle patient with alloantibodies each year. The acknowledgement of these cases may help in is assessing the current practice of transfusing SCD patients, or will help to define the donor and patient population difference.

Methods: 418 patients were recruited in the Royal Hospital for this study. EDTA blood samples were taken for antibody screening test and in the positive cases antibody was identified, all tests done by capture technique using Immucor Neo machine.

Results: Of the 418 SCD patients, 49% of the patients were male and 51% female, mean age was 17years, in the range of 1–72years. 38 % of the SCD cases were positive for the alloantibodies, 72% were Female and 28% were male, the age range was from 6–68years. 78% of the positive were SCD, 19% S trait and 3% were S/bthal. Most of the patients developed one antibody, however cases of multiples antibodies were also detected. 51% of the patients were with single alloantibody, 30 % of them with two antibodies, 10 % with three antibodies, 7 % with four antibodies and 2 % with five antibodies. The majority of the cases were IgG against Rh antigens anti‐ E is being the majority 23%, followed by anti‐D 13%, anti‐K 17%, anti‐C 14%, anti‐c 8%, anti‐Jk^a 5%, anti‐ Jk^b 5%, anti‐Fy^a 5%, anti‐e 3%, anti‐S 3%, anti‐s 1%, anti‐Kp^a 0.7%, anti‐Fy^b 0.3% and IgM being 2%.

Summary/Conclusions: RBC alloimmunisation rate is high in Oman majority of the patient affected are female. Interestingly sickle trait patients were also transfused and 19% of them developed alloantibodies. The practice of transfusing Rh and Kell matching blood unit is implemented four years ago and still high alloimmunization percentage is achieved.

RED BLOOD CELL ANTIBODIES IN MULTI‐TRANSFUSED PATIENTS WITH SICKLE CELL DISEASE IN GHANA; PREVALENCE, SPECIFICITIES AND RISK FACTORS

LA Boateng^1,2,3, H Schonewille², P Ligthart², A Javadi², Y Dei‐Adomako⁴, A Osei‐Akoto⁵, I Bates¹, C van der Schoot²

¹International Public Health, Liverpool School of Tropical medicine, Liverpool, United Kingdom ²Experimental Immunohaematology, Sanquin, Amsterdam, Netherlands ³Medical Laboratory Technology, Kwame Nkrumah University of Science and Technology ⁴Haematology, Ghana Institute of Clinical Genetics (Adult Sickle Cell Clinic) ⁵Child Health, School of Medical Sciences, Kwame Nkrumah University of Science and Technology, Kumasi, Ghana

Background: In Ghana, routine pre‐transfusion investigations for patients with sickle cell disease (SCD) involve only ABO‐D typing and immediate spin cross‐match, without screening for irregular RBC antibodies

Aims: Determine the prevalence and specificities of and risk factors for RBC alloantibodies in multi‐transfused patients with SCD

Methods: In 2018, a cross‐sectional study in multi‐transfused patients with SCD, from two tertiary hospitals in Ghana was performed. Participants’ data on demography, transfusion and medical history were recorded. Antibody screening and identification tests were done at Sanquin, the Netherlands, with standard serology using LISS as enhancer and with papain treated RBC panel cells (‘enzyme only’). Characterization ofRHD genes was done by Multiplex Ligase Amplification Assay. Logistic regressionwas used to determine the association of patient characteristics, i.e. sex, age at enrollment (continuous), age at first transfusion (categorized as≤1, 2‐5, 6‐9 and ≥10), previous pregnancy, number of transfused units (2, 3‐5 and 6‐10 and>10), and years after last transfusion (<1, 1‐2, 2‐5, >5y) with presence of allo‐antibodies

Results: 226 patients (100 males and 126 females, median age 17years, range 1.7‐66) were included. The median number of transfusions was 3 (range 2‐40). The median years after last transfusion was 2 (range 2weeks‐55.5years). In 56 patients, anti‐RBC antibodies were detected. In 14 of them the antibodies were weakly reactive with enzyme treated cells only or pan‐reactive, possibly some of them representing autoantibodies or antibodies against high frequency antigens. In seven patients enzyme‐only anti‐Le^awas demonstrated, likely naturally occurring antibodies. Thus, in at least 34 patients (15.0%) alloimmunization was demonstrated or suspected; in 11 patients the alloantibodies were ‘enzyme only’. Besides, the 36 alloantibodies of known specificity (8 anti‐D, 2 anti‐D+C, 16 anti‐E, 1 anti‐C, 3 anti‐e, 1 Anti‐K, 1 anti‐s, 1 anti‐Le^a, 1 anti‐Go^a), three antibodies reactive only with Fy(a‐b‐) cells and two antibodies of yet unidentified specificity were detected. In six D‐ patients (3 had been pregnant) anti‐D (together with anti‐C in two patients) was found. In three out of four D+ patients with anti‐D, anRHDvariant gene was demonstrated (2 DAU‐alleles and 1 DIII type 4 or DIVa‐2).

Logistic regression revealed that none of the risk factors analysed was associated with the presence of antibodies in the 34 patients.

Fifty‐eight patients, had experienced an adverse reaction during or shortly after transfusion (12 patients had dark urine). Adverse reactions were associated with the number of units received (OR 1.71 (95% CI, 1.25‐2.33; P=0.001), but not with the presence of antibodies (P>0.7)

Summary/Conclusions: In at least 15% of multi‐transfused patients with SCD alloimmunization could be demonstrated, mainly (80%) directed against Rh antigens. The enzyme only reactivity, coupled with absence of antibodies in seven of 12 patients with probable haemolytic reaction and known evanescence of especially non Rh antibodies suggest possible low titre and disappearance of some clinically relevant antibodies.

Given the high immunization rate together with the high frequency of adverse transfusion reactions, pre‐transfusion screening for RBC antibodies should be considered for patients with SCD.

STRONG PREGNANCY INDUCED ANTI‐D IMMUNIZATION IN DEL PHENOTYPE WITH RHD*01EL.04 ALLELE

M Pisacka¹, J Kralova¹, V Hanzikova², P Calda³, A Horinek⁴, E Pazourkova⁴, S Schneider⁵, S Scholz⁵

¹Reference Laboratory for Immunohaematology, Institute of Haematology and Blood Transfusion ²Transfusion Dpt., General University Hospital ³Fetal Medicine Center, Charles University and General University Hospital ⁴Institute of Biology and Medical Genetics, General University Hospital, Prague 2, Czech Republic ⁵Inno‐train Diagnostik GmbH, Kronberg, Germany

Background: Rh blood group system and mainly antigen D is one of the most immunogenic, diverse and clinically important protein‐based blood group. Antibody anti‐D may induce hemolytic transfusion reactions and hemolytic disease of the fetus and newborn. Anti‐D prophylaxes become ineffective if an anti‐D immunization has occurred. Approximately 1% of the D+ population carries RHD alleles associated with reduced D antigen expression. Qualitative variants, in which some epitopes are lacking and can produce anti‐D antibody, are usually termed partial D. By contrast, D weak is commonly defined as a quantitative variant that have all D epitopes and should not make anti‐D. Del is a very weak form of D antigen and cannot be detected by routine serological tests. Because some of Del individuals have already developed an anti‐D antibody whereas others did not this group contains both qualitative and quantitative changes.

Aims: Investigation was prompted by finding discrepant results in typing of D antigen in a pregnant woman /3^rd pregnancy, 1^st delivery, 2 abortions in 1^st trimester/. Routine serological techniques detected D negativity and the presence of antibody allo‐anti‐D in clinically significant titre. The non‐invasive testing of D status of the foetus from maternal peripheral blood was indicated, but this was not applicable due to presence of the RHD gene in the woman's DNA sample isolated from buccal swab. Our aim was to investigate the discrepancy and determine the underlying RHD genotype.

Methods: Blood samples, DNA from peripheral blood and buccal swab of the pregnant woman were investigated. Routine blood grouping and antibody testing were performed by column agglutination. Two anti‐D sera (ID‐DiaClon anti‐D IgG (cell line ESD1) by BioRad and Anti‐D duo IgM+IgG, clone: TH28+MS26 by Immucor) were used for adsorption/elution test for identification of Del phenotype. Initial RHD genotyping was performed by RT‐PCR (exons 5,7,10) with the DNA from buccal swab; further resolution was performed using PCR‐SSP (FluoGene; Inno‐train Diagnostik GmbH); Sequencing was performed by Sanger analysis (Inno‐train Diagnostik GmbH).

Results: Genotype was identified as RHD positive by CE‐certified PCR‐SSP kits (FluoGene). Sanger sequencing of RHD from exon 1 to 9 revealed presence of a nucleotide deletion in position c.147delA, which is specific for allele RHD*01EL.04. This nucleotide change results in the amino acid change p.Val50Leufs*5 causing the Del phenotype. Presence of antigen D was proved by adsorption/elution technique. Titre of the anti‐D was rising during the pregnancy to the 2000 level two weeks before the delivery. The newborn was delivered by S.C. without a sign of hemolytic disease. Blood grouping of the newborn revealed blood group A, D negative, DAT negative, testing for Del was not performed.

Summary/Conclusions: The case reported here shows that females with RHD*01EL.04 allele are able develop strong anti‐D immunization, so this type of Del phenotype belongs to the “Partial Del subgroup”. Presence of variant RHD gene in mother disabling antenatal fetal genotyping from maternal blood by current methods requires a more attentive approach to care for such pregnancies. Supported by MH CZ‐DRO UHKT 00023736 and RVO‐VFN 64165.

ANTI‐RBA IS A VERY FREQUENT RED CELL ANTIBODY IN GERMANY

EA Scharberg¹, S Rothenberger¹, A Stürtzel¹, N Gillhuber¹, S Seyboth¹, E Richter¹, G Rink², P Bugert²

¹Institute for Transfusion Medicine and Immunohematology, DRK‐BSD Ba‐Wü‐He, Baden‐Baden ²Institute for Transfusion Medicine and Immunology, Heidelberg University, Medical Faculty Mannheim, Mannheim, Germany

Background: Rb^a (DI6) is a low prevalence antigen of the Diego blood group system. It has been found in few families only. The clinical significance of anti‐Rb^a is unknown so far. The SLC4A1*c.1643C>T (p.Pro548Leu; ISBT allele name: DI*02.06) allele is the molecular basis of the Rb^a antigen. In the gnomAD database this gene variant was found in only one of 125,742 sequenced genomes (allele frequency: 0.000004).

Aims: To prove the frequency of the allele in our population and gain an Rb^a positive donor we performed a molecular screening for DI6 in 1,700 blood donors. After our antibody screening test accidentally contained an Rb^a positive test cell we found out that anti‐Rb^a is a very common antibody specificity. The frequency of the antibody in patients and blood donors was proved.

Methods: For the molecular screening of the blood donors we developed a PCR‐SSP method. The antibody screening test in 3,652 patients and in 964 blood donors was performed in the gel technique (BioRad AHG ID‐Cards) using a 3 cell screening panel (DRK‐BSD SRC) including an Rb^a positive test cell. Positive reactions with the Rb^a positive cell were confirmed by an additional Rb^a positive test cell of different source. Additional antibodies were excluded or identified in the same method using an antibody identification panel (DRK‐BSD IRC).

Results: The molecular screening for the DI*02.06 allele in 1,700 blood donors revealed no single positive individual. Within the first 2weeks of usage of our antibody screening test which accidentally contained the Rb^a positive test cell 84 patients with anti‐Rb^a were found. It was 2.3% of 3,652 patients tested in 21 laboratories in different parts of Germany. Some laboratories stopped using the Rb^a positive lot to avoid expensive and time consuming identification and conformation tests. In 18 of 964 randomly tested blood donors (1.9%) anti‐ anti‐Rb^a was also present.

Summary/Conclusions: Despite the very low frequency of the DI*02.06 allele, anti‐Rb^a is a very frequent unexpected antibody in patients and blood donors in Germany. It is obviously naturally occurring and is even more frequent than anti‐Wr^a and anti‐Vw we found in previous studies in around 1% of patients and donors.

THE IMPORTANCE AND THE APPROACH TO SET UP A SYSTEM OF HEMOVIGILANCE

T Aung

National Blood Center, Ministry of Health and Sports, Yangon, Myanmar

Hemovigilance which detects every event not only for patient’ reactions and donor's complications but also incidents and near misses definitely improve quality of blood transfusion services especially for those situations where implementation of all the standards in one time is not possible.

Healthcare system in Myanmar is still in the stage of requiring priority for clinical professions and has limited resources for supportive roles. Supportive services including transfusion service are still not a center of interest from prioritization of health care system. Blood Transfusion service has been practiced in Myanmar since 1935. Real essence of transfusion service is hidden behind laboratory practice and transfusion is regarded as part of laboratory investigation. Hospital laboratories take care of testing of blood donated by replacement donors. This kind of transfusion services under laboratory umbrella is still being practiced in Myanmar except National Blood center (NBC) which was established in 2003 in accordance with Blood and Blood Product law. This Law was formulated cohesively with WHO strategies of blood safety. In 2002, WHO global data‐based study sent questionnaires for assessment of safety status of transfusion service. NBC noticed that there was no data which can support corrective actions for safety. From that time onward, active retrospective review of existing data and introduction of records, prospective finding of process errors and any events from hospital blood banks were recorded and taking into actions at local level. In 2003, review of screening test pattern between voluntary and replacement donations gave an alarm for importance of education and systematic recruitment of voluntary donors in such a situation of high prevalence of viral carrier rate in general population. Introduction of donor deferral system, checking of previous donation and testing history were the outcome of this active analysis. Results from international EQAS in 2003 further motivated us for changing of testing methods of HBs Ag. By reviewing the history of Japanese transfusion service, systematic recruitment of voluntary donors, changing testing strategies and planning of blood transfusion services to be nationally well organized were gradually introduced. In 2018, NBC supported 140560 units of blood components to meet the demand of 14 HBBs by 99% Voluntary donations using systematic planning based on data from system. TTI testing is now used Minipool of 6 NAT for new donors and Eclia for repeated donors. Cost of every unit of blood is supported by Government. In 2017, National blood and Blood Product committee was established. The steering committee is working hard to get cooperation from every service by aiming to prevent those undesirable events before establishment of national level policy, standards and guidelines for sustainable service quality.

In conclusion, by using essence of hemovigilance as a tool, quality of transfusion service can be improved step by step to fulfil the gap in spite of limited resources. The system started in local, extends to regional level by getting agreement of importance from hemovigilance results and is finally approaching to National level endorsement.

LEARNING FROM TRANSFUSION ‘NEVER EVENTS’ – REVIEW OF UNINTENTIONAL ABO INCOMPATIBLE TRANSFUSIONS AS REPORTED TO SERIOUS HAZARDS OF TRANSFUSION 2010‐2017

S Narayan¹, J Addison¹, D Poles², H Mistry³, S Carter‐Graham¹, A Watt⁴

¹Clinical‐ Haemovigilance ²Research analyst ³Laboratory Incidents Specialist, Serious Hazards of Transfusion ⁴Human Factors Expert in SHOT Working Expert Group, Independent, Manchester, United Kingdom

Background: Erroneous transfusion of ABO‐incompatible(ABOi) blood almost always reflects a preventable breakdown in transfusion protocols and standard operating procedures and can have disastrous consequences, with significant morbidity and mortality. These incidents need to be investigated in a systematic manner to identify system vulnerabilities to mitigate risks and improve patient safety. Since 2016, reporters to SHOT have been asked to score(0‐10) the extent to which the cause of incidents can be attributed to key factors: staff, environmental, organisational and government/regulatory which helps recognise the key factors identified whilst investigating these incidents.

Aims: To understand why unintentional transfusion of ABOi blood components continue to happen despite standard procedures and national guidance available.

Methods: Retrospective analysis of unintentional transfusion of ABOi blood components reported to SHOT between 2010‐2017(inclusive) was done to identify common themes and recognise areas of improvement. Information provided using the SHOT human factors investigation tool (HFIT) between 2016‐2017 was reviewed to understand more about why the errors occurred.

Results: Sixty‐seven unintentional ABOi transfusions were reported between 2010‐2017; majority (56/67, 83.6%) were red cell transfusions but ABOi plasma (9/67) and platelet transfusions (2/67) were also seen. Most errors occurred in the clinical area (45/67, 67%), and could have been detected at point of administration. In 21(31%) cases, the error could not have be detected at the point of administration with a primary laboratory error in 10/21(48%) incidents.

Reviewing data from HFIT for cases in 2016‐2017 (13 ABOi cases), the total score for staff culpability was 100, compared to a total score of 99 for all the other three organisational and system factors. This disparity is most obvious for the 4 ABOi red cell cases, all of which scored the maximum 10 for staff culpability, i.e. 40/40 compared to 8/120 as the combined total score given to the other factors. In the preceding years (2010 to 2015), there were no HF scores available; however, the emphasis on staff‐related culpability is demonstrated by 37 cases that included an outcome of the local case review and 14 (37.8%) mentioned staff‐related retraining or disciplinary procedures. The risk of haemolysis and serious harm is more likely with ABOi red cells than with other components with 2/56(4%) that resulted in death, 14/56(25%) major morbidity and 40/56(71%) no or minor adverse reaction. Of these cases, one resulted in conviction for manslaughter and at least two staff dismissals.

Summary/Conclusions: Transfusion never events continue to occur, and it is evident that investigations into such incidents focus mainly on staff failings and do not consistently identify system wide changes that need to be incorporated to address prevalent issues. National recommendations and a safety alert to ‘use a bedside checklist’ immediately prior to administration were issued between 2015‐2018 to support prevention of such errors but never events continue to persist. Current approach is ineffective because it often leads to apportioning blame, rather than understanding the often‐complicated and multidimensional factors contributing to the error. This must be replaced by a holistic approach which addresses local work pressures and embraces advances in automated technology like electronic prescribing and barcode scanning.

HEMOVIGILANCE ON MIRASOL PATHOGEN‐REDUCED WHOLE BLOOD IN GHANA

A Owusu‐Ofori¹, L Asamoah‐Akuoko², M Acquah², S Wilkinson³, J Ansa², B Brown³, S Owusu‐Ofori¹

¹Komfo Anokye Teaching Hospital, Kumasi ²Korle Bu Teaching Hospital, Accra, Ghana ³Terumo BCT, Lakewood, United States

Background: Data collection and monitoring are critical elements of a hemovigilance (HV) system. Terumo BCT, Inc. (TBCT) and AABB's consulting services collaborated with the National Blood Service Ghana to streamline HV data collection. The Japan International Cooperation Agency funded this project.

Aims: The project objectives were to improve transfusion practice by establishing a system to monitor and evaluate safety of blood transfusions within two hospitals, Korle Bu Teaching Hospital (KBTH) in Accra and Komfo Anokye Teaching Hospital (KATH) in Kumasi, Ghana. An assessment of acute transfusion‐related adverse reactions (TRAR) among the Mirasol Pathogen Reduction Technology System (PRT)‐treated Whole Blood (WB) versus conventional WB transfusions was done.

Methods: Trainings on HV was conducted in May 2017. HV training based on ISBT standards included: 1) quality systems, 2) TRAR, 3) HV forms and reports, and 4) database development. Routine use data was collected prospectively. Data collection focused on evaluating acute TRAR from WB transfusions. In December 2018, data was analyzed to evaluate the safety of Mirasol‐treated WB in routine use in Radiotherapy, Obstetrics and Gynecology, Oncology, Paediatrics, Haematology, Gynaecology emergency, and Maternity wards.

Results: As of December 2018, 2181 transfusion records were collected; 1019 Mirasol‐treated WB transfusions and 1162 conventional WB transfusions. Overall, there were 4.02% TRARs in recipients of Mirasol‐treated WB and 5.59% in recipients of conventional WB. In recipients of Mirasol‐treated WB, there were n=7 (0.69%) allergic reactions (ARs), n=17 (1.67%) febrile non‐hemolytic transfusion reactions (FNHTRs), n=3 (0.29%) cases of transfusion‐acquired circulatory overload (TACO), no transfusion‐related acute lung injury (TRALI), and n=14 (1.37%) unclassified transfusion reactions. Of the confirmed TRARs, n=27 were possibly related to treatment, n=2 TRARs were probable, and n=2 were definitely related to treatment; n=37 TRARs were Grade 1, n=4 were Grade 2, and none were Grade 4. In recipients of conventional WB, there were n=13 (1.12%) ARs, n=32 (2.75%) FNHTRs, n=1 (0.09%) TACO, n=0 TRALI, and n=16 (1.38%) unclassified transfusion reactions. Of the confirmed TRARs, n=55 were possibly related to treatment, n=1 TRAR was probable, and n=8 were definitely related to treatment; n=54 TRARs were Grade 1, n=1 was Grade 2 and n=1 was Grade 4. There were 21 Mirasol‐treated WB transfusions in pregnant women and 2 TRARs (9.5%), both Grade 1 and probably related. There were 80 transfusions of Mirasol‐treated WB and 84 transfusions of conventional WB in patients<18years old resulting in n=7 (8.33%) TRARs in recipients of Mirasol‐treated WB and n=10 (11.90%) in recipients of conventional WB.

Summary/Conclusions: Timely data reporting of TRARs and expanding the HV infrastructure has helped to improve the HV system in Ghana. Of 2181 WB transfusions in routine use in Ghana, there were 4.02% TRARs in recipients of Mirasol‐treated WB and 5.59% in recipients of conventional WB. Additionally, Mirasol‐treated WB was safely transfused in pregnant women and pediatric patients.

IRRADIATED BLOOD PRODUCT USE IN PATIENTS AT RISK FOR TRANSFUSION‐ASSOCIATED GRAFT‐VERSUS‐HOST DISEASE

H Yuen, K Rushford, T Dunstan, C Michael, A Tey, K Yeoh, E Wood

Haematology, Monash Health, Melbourne, Australia

Background: Transfusion‐associated graft‐versus‐host disease (TA‐GvHD) is rare and usually fatal. It can be prevented by provision of irradiated blood products to at‐risk individuals, such as those receiving nucleoside analogues, alemtuzumab, bendamustine or with Hodgkin lymphoma (HL). Duration of risk is uncertain, so ensuring these individuals correctly receive lifelong irradiated blood components, as currently recommended by ANZSBT and BSH guidelines, is challenging. In Australia, platelets are routinely irradiated, but red blood cells (RBC) are not.

Aims: To determine whether patients receiving fludarabine, cladribine, bendamustine, alemtuzumab, or dacarbazine (for HL), appropriately received irradiated RBCs. Secondary outcomes included rates of TA‐GvHD after unintended exposure to non‐irradiated components, factors influencing correct issue of irradiated RBCs such as transfusion management plans, and provision of adequate clinical information on blood requests.

Methods: We performed a retrospective audit to identify patients receiving therapies indicating risk for TA‐GvHD using pharmacy dispensing records from January 2008 to October 2018 at Monash Health, a multi‐campus university hospital in Melbourne, Australia. Diagnosis, treatment dates, group and hold (G&H) requests, RBC transfusions, and follow‐up information were sourced from laboratory and medical records.

Results: We identified310 patientswho received fludarabine (n=52, 17%), bendamustine (n=29, 9%), cladribine (n=17, 5%), dacarbazine for HL (n=164, 53%) and alemtuzumab (n=48, 15%). The median age of patients was 46years (range 8‐92) and 171 (55%) were male. Median follow‐up was 30months (range 0‐132). Post‐exposure, 42 patients (14%) received transfusionswith 33% correctly receiving irradiated RBCs. The remaining 28, all from haematology/oncology, received a total of 192 unirradiated RBCs. In 8 patients, this was rectified on subsequent transfusions. There were no cases of TA‐GvHD at median follow‐up of 14.5months (range 0‐75) from first RBC transfusion.

After medication administration, 99 patients had G&H requests after a median of 3months(range 0‐129). Only 23% of requests had sufficient clinical information to prompt irradiation, such as HL or medication details, and only 20% asked for irradiated components.

Preventive strategies have now been employed. Transfusion management plans for haematology patients were implemented in March 2017. For audited patients, these were written from 38days prior to 104days after medication exposure. Two were written following inadvertent unirradiated RBC transfusion.

Patients identified in this audit will have a laboratory flag generated and prospectively, pharmacy dispensing records will be sent to blood bank to identifyat‐risk patients. Our hospital is transitioning to electronic medical records (EMR). An alert will be generated in EMR when ordering transfusions if there has been exposure to these medications. However, clinical awareness and documentation remain vital. Additional measures include patient education, alert cards, and ongoing collaboration with medical staff to encourage transfusion planning.

Summary/Conclusions: Recognition of patients at risk for TA‐GvHD remains low, even among haematology units. We are making progress on ensuring provision of lifelong irradiated blood components in patients exposed to nucleoside analogues or alemtuzumab, as well as HL patients. Implementation of an EMR and additional strategies in this domain is important to prevent TA‐GvHD.

TRANSFUSION RELATED ADVERSE EVENT REPORTING AND DOCUMENTATION COMPLIANCE AMONG HEALTHCARE PROFESSIONALS IN DEVELOPING COUNTRY‐ AIMING TOWARDS IMPROVING PRACTICES

N Anwar¹, N Fatima², H Mujtaba², T Shamsi¹

¹Hematology ²Research and development, National Institute of blood diseases and bone marrow transplant Karachi Pakistan, Karachi, Pakistan

Background: Blood transfusion is considered an essential element in the management of patients globally. It might be risky and transfusion related adverse reactions may occur with the less adherence of transfusion policies. Standard guidelines regarding the screening of blood for infectious disease, genuine need of transfusion and ABO compatibility are followed and monitored drastically. However, patient assessment during transfusion especially at patient bedside and post transfusion is also equally important.

Aims: We are a newly established hospital and are working towards the best possible management of patients. In this regard to minimize the transfusion errors and to highlight if any lacking being practiced during transfusion, we conducted this study to observe the compliance rate of documentation of transfusion form by the healthcare staff and also to observe the compliance of line of action taken in case of occurrence of transfusion reactions.

Methods: This was a observational study conducted at NIBD and BMT, PECHS campus from February 2018 to February 2019. Ethical approval was obtained prior to the study. Transfusion form for each transfusion was filled. The form provided information on documentation of blood product receiver name, employee identity number, date and time of receiving blood product, patient name, medical record number on units, on patient's wrist band and on transfusion form. ABO compatibility on the unit and on form, medical record number from wrist band, name and employee identity number of two healthcare staff started transfusion, transfusion start and completion time. Time, temperature, blood pressure, pulse and initials of staff at the time of order, onset, after 15minutes and at the completion of transfusion were also included. Transfusion reaction form was also filled by the healthcare staff. Data was analyzed by using SPSS version 23.0.

Results: A total of 500 transfusions forms were analyzed. Over all compliance rate was 18%. Out of 500, 115(23%) forms were available in source notes and of 115, 88 (76%) were partially and completely filled. Higher compliance was seen in the initial months of hospital establishment than later months (P‐value=0.000). Highest non compliance was seen in documentation of initials of duty doctors on transfusion form at the completion of transfusion(3%) and highest compliance was seen in documentation of name by healthcare nursing staff at the start of transfusion(89%). A total of 02(0.4%) adverse events were reported from red blood cells and platelets. Mean time of start of symptoms was 2hours and 30minutes for red blood cells and for platelets it was 1hour and 13minutes. Transfusion was instantly stopped as the symptoms appeared with no delay of time and actions were taken to resolve the reactions. Time of appearance of symptoms and time of start of medication were documented and error free. All blood bags were returned to the blood bank and discarded after 6hours as per the policy of hospital.

Summary/Conclusions: The study was conducted to highlight the scarce practices that are being implemented by healthcare staff in context of documentation and reporting of transfusion reactions at our hospital. Stringent actions should be taken for the adherence of compliance by healthcare staff to avoid morbidities and mortalities. We believe that it will also be helpful to provide baseline information in the process of preparation of a national guidelines and protocol on blood transfusion procedures.

IMPACT OF PATHOGEN INACTIVATION ON PLATELET SURVIVAL AND ALLOIMMUNIZATION

AS Buser, A Holbro, L Infanti

Regional Blood Transfusion Service, Swiss Red Cross, Basel, Basel, Switzerland

To make blood supply safer, pathogen inactivation (PI) technologies have been developed. They are based on photochemical (amotosalen/UVA or riboflavin/ UV) or UV‐C light treatment to reduce potential pathogens in blood components. This gain of safety might however be offset by “off target” effects of these technologies.

In virtually all clinical platelet transfusion trials, it has been demonstrated that post transfusion increments with PI platelet (PLT) components are lower as compared to conventional components, indicating different biological behaviour such as survival/ clearance of nontreated and treated PLT. Published studies have also suggested shorter survival of platelets in vivo in animal studies.

Additionally, data of the rates of alloimmunization and refractoriness after transfusion of PI platelets are show discrepant results. Animal studies suggest a reduction of the rates of alloimmunization when transfusing (leukoreduced) PI PLT as compare to conventional PLT.

In the clinical setting, published data, including very recent reports, showed different rates of HLA class I and II alloimmunization with the two currently available photo‐chemical based PI technologies.

While PI of PLT components surely benefit patients regarding pathogen safety, the impact of potential off target effects possibly impairing efficacy of PI PLT transfusions need more investigation.

EFFICIENT INACTIVATION OF BRUCELLA CLINICAL ISOLATES IN HUMAN PLATELET CONCENTRATES IN 100% PLASMA WITH AMOTOSALEN AND ULTRAVIOLET A LIGHT TREATMENT

F Alseraye¹, O Alsaweed¹, F Albloui¹, H Alghethber², A Hazazi³, F Aldakheel⁴

¹Pathology and Laboratory Medicine ²Internal Medicine, Hematology, Security Forces Hospital, Riyadh ³Department of Pathology and Laboratory Medicine, Security Forces Hospital, Riyadh ⁴Clinical Laboratory Sciences, King Saud University, Riyadh, Saudi Arabia

Background: Brucellosis is an endemic disease and still a major health problem in Saudi Arabia. Ministry of Health in Saudi Arabia listed Brucellosis as a notifiable disease due to its endemicity. In the last ten years, the incidence has decreased significantly to approximately 15 cases per 100,000 but is still higher than that in developed countries. Human‐to human transmission is extremely rare including breast feeding, transplacental, sexually and blood transfusion. Five cases of Brucellosis through blood transfusion have been reported in the literature. Brucella transmission through blood transfusion is likely underreporting due to the long incubation time of 2–4weeks (range, 5days to 5months),vagueness of clinical presentation and lack of hemovigilance systems in endemic areas.

Aims: Assessment of the efficacy of Amotosalen/UVA technology to inactivate clinical Brucella isolates in human platelet concentrates in 100% plasma. Considering a minimum infectious dose of 10‐100CFU, sterilization to prevent infection is needed.

Methods: Whole‐blood derived interim platelet units (IPUs) in 100% plasma were produced with an automated Reveos system (Terumo BCT). Six IPUs were pooled and inoculated with approx. 12000CFU of a clinical Brucella isolate per pool. Baseline samples from the pool pre‐inoculation were taken and tested for bacterial growth, RBC count, WBC count, pH, platelet count, swirling and HIV/HBV/HCV serology and NAT. The pool was incubated at 20‐24°C under continuous agitation for 24‐30h followed by split into a therapeutic test unit and a small volume control unit. The test unit was treated with the INTERCEPT Blood System Large Volume Processing Set for Platelets (Cerus Corporation), while the control unit was untreated. Both units were stored until day 7 at 20‐24°C under continuous agitation taking samples for blood culture using an automated BacT/ALERT system (Biomerieux) before inactivation and at day 7 of storage. Negative blood cultures were incubated for 21days.

Results: All pools were negative for bacterial growth, showed no presence of aggregates and had an RBC count≤01.x10⁶/mL. Six different Brucella clinical isolates were characterized by 16s rRNA sequencing and used for independent inactivation experiments. The average incubation time post inoculation was 26:47h (25:30‐27:00h). Taking that into consideration, the bacterial load pre‐inactivation was calculated based on a 2.5‐3.5h doubling time and an initial titer of approx. 30CFU/mL to be approx. 5760‐46080CFU/mL. The test units had an average volume of 289mL (281‐300mL) and total platelet dose of 3.3x10¹¹ (2.9‐4.0x10¹¹). Post inactivation, all test units (day 1 and day 7) were negative for 21days of blood culture, while the control units were positive after an average culturing time of 20:17h (19:25‐21:24h) at day 1 and 7:21h (4:26‐8:38h) at day 7.

Summary/Conclusions: All six Brucella isolates were efficiently inactivated in human platelet concentrates in 100% plasma, pointing towards enhanced prevention of transfusion‐transmitted Brucella infections.

THERAPEUTIC RESPONSE TO AMOTOSALEN/UVA‐TREATED PLATELETS WITH UP TO 7 DAYS STORAGE DURING 5 YEARS OF ROUTINE PRACTICE

L Infanti¹, A Holbro¹, J Passweg², D Tappe³, J Irsch⁴, J Lin³, L Corash³, R Benjamin³, A Buser¹

¹Regional Blood Transfusion Service Basel, Swiss Red Cross, Division of Hematology, University Hospital ²Division of Hematology, University Hospital, Basel, Switzerland ³Cerus Corporation, Concord, CA, United States, ⁴Cerus BV, Amersfoort, Netherlands

Background: The Swiss Red Cross blood transfusion service, Basel, introduced the universal use of amotosalen/UVA (INTERCEPT™ Blood System for Platelets) pathogen‐inactivated apheresis (80‐90%) and pooled whole blood‐derived, buffy‐coat (10‐20%) platelet components (PC) in 65% Platelet Additive Solution (PAS) (Intersol) and 35% plasma in 2011. PCs were routinely stored for up to 7days.

Aims: The study evaluated the impact of PC storage age on therapeutic efficacy of amotosalen/UVA‐treated PC under routine use conditions for hematopoietic stem cell transplant (HSCT) and non‐HSCT hematology/oncology (Heme/Onc) patients treated at the University Hospital Basel.

Methods: Amotosalen/UVA‐treated PC stored for up to 7days were transfused for all indications with routine collection of pre‐transfusion and 1‐4hour post‐transfusion platelet counts. Platelet component and recipient characteristics of all platelet transfusions were prospectively captured and retrospectively analyzed by storage day. Platelets were transfused on a first in, first out basis without regard to PC age. The risk of hemostasis failure was assessed by the need for additional PC or RBC transfusion on the day of or the day following an index PC transfusion.

Results: Between January 10, 2011 and May 17, 2016, 22,579 INTERCEPT PC were transfused to 2,809 general hospital patients. Platelet dose was 3.0±0.4×10¹¹ with a storage age of 4.2±1.4days. Around 18.3% were>5days old at the time of transfusion. Heme/Onc and HSCT patients used 77.5% of PC and comprised 982 (35.0%) of the transfused population, 441 (15.7%) Heme/Onc, 411(14.6%) allogeneic (alloHSCT) and 130 (4.6%) autologous (autoHSCT) HSCT patients, with mean corrected count increments (CCI) of 8.7×10³, 8.0×10³ and 7.2×10³, respectively. Mean CCI decreased in a linear fashion between Day≤2 and Day 7 PCs (9.0×10³, 9.1×10³ and 8.5×10³ at≤2days; 6.7×10³, 5.8×10³ and 4.9×10³ at 7days, respectively), although the number of PC transfused on Day 7 to autoHSCT patients was small (n=71). The proportion of transfusions followed by additional PC on the same or next day was 47.8%, 73.4% and 50.4% for Day≤2 PC vs. 46.4%, 73.6% and 66.2% for Day 7 PC in Heme/Onc, alloHSCT and autoHSCT patients, respectively. Use of RBC on the same or next day as an index PC transfused was 49.7%, 53.2% and 44.8% for Day≤2 PC vs. 49.6%, 48.0% and 49.3% for Day 7 PC, respectively. The median time to the next PC transfusion was≤1.0day for Day≤2 PC and 1.0days for older PC in all three groups. The incidence of all transfusion reactions, febrile non hemolytic transfusion reactions and allergic reactions was comparable for PC≤5 or>5days‐old in all groups.

Summary/Conclusions: Transfusion of pathogen inactivated PCs older than 5days to hematology/oncology, allogeneic and autologous HSCT patients is safe and effective and did not affect hemostasis as assessed by the need for additional PC or RBC transfusions on the day of, or the day following an index PC transfusion.

NIPAH VIRUS IS EFFICIENTLY INACTIVATED IN PLATELET CONCENTRATES BY UVC LIGHT USING THE THERAFLEX UV PLATELETS TECHNOLOGY

U Gravemann¹, M Eickmann², W Handke¹, F Tolksdorf³, A Seltsam¹

¹Research and Development, Red Cross Blood Service NSTOB, Springe ²Institute of Virology, University of Marburg, Marburg, ³Macopharma, Langen, Germany

Background: Nipah virus (NiV) is a paramyxovirus (genus Henipavirus) that emerged in the late 1990s in Malaysia and has since been identified as the cause of sporadic outbreaks of severe febrile disease in Bangladesh and India. NiV infection is frequently associated with severe respiratory or neurological disease in infected humans with transmission to humans through inhalation, contact or consumption of NiV contaminated foods. Nipah virus (NiV) belongs to the list of pathogens identified by the WHO to have the potential for a global pandemic.

Aims: This study aimed to investigate the efficacy of the THERAFLEX UV‐Platelets system to inactivate NiV in platelet concentrates (PCs). The THERAFLEX UV‐Platelets system (Macopharma) uses UVC light without the need of any additional photoactive compound.

Methods: Plasma reduced PCs from 4 BCs (35% plasma in additive solution SSP+) were spiked with virus suspension (10% v/v). PCs (n=2, 375mL) were then UVC‐irradiated on the Macotronic UV machine (Macopharma) and samples were taken after spiking (load and hold sample) and after illumination with different light doses (0.05, 0.1, 0.15 and 0.2 (standard) J/cm²)). The titre of the NiV (Malaysia) was determined as tissue culture infective dose (TCID50) by endpoint titration in microtitre plate assays on Vero 76 cells (ATCC^® CRL‐1587™).

Results: The results of the infectivity assay demonstrated that UVC irradiation dose‐dependently inactivated NiV. After spiking a NiV titer of 6.2 (bag no. 1) and 6.5 (bag no. 2) log10 TCID50/mL was received in the PCs. At a UVC dose of 0.10J/cm² and higher NiV was inactivated down to the detection limit of the system (1.9 log10 TCID50/mL), resulting in log10 reduction factors of ≥4.3 (bag no. 1) and ≥4.6 (bag no. 2).

Summary/Conclusions: Our results demonstrate that the THERAFLEX UV‐Platelets procedure is an effective technology to inactivate NiV in contaminated PCs.

USE OF PLATELET PATHOGEN REDUCTION TO IMPROVE STOCK MANAGEMENT IN A NATIONAL BLOOD SERVICE

N Arnason^1,2, R Landro¹, G Svansdottir¹, B Hardarson¹, I Hjalmarsdottir¹, T Jonsson¹, S Gudmundsson¹, OE Sigurjonsson^1,2,3, A Halldorsdottir^1,3

¹The Blood bank, Landspitali‐The National University Hospital of Iceland ²School of Science and Engineering, Reykjavik University ³Faculty of Medicine, University of Iceland, Reykjavik, Iceland

Background: The Reykjavik Blood Bank (BB) is a nationwide blood transfusion service in Iceland serving a population of 350.000 inhabitants. Due to Iceland′s unique geographical situation our blood stock management is both a challenge and a matter of public safety. To stabilize the supply of platelet concentrates (PC) and to reduce bacterial contamination risk, pathogen reduction (PR) of all PC was implemented in 2012. The INTERCEPT Blood System™ (IBS) allowed for extension of platelet storage from 5 to 7days. Prior to 2012 PC stocks fluctuated considerably, with resultant spells of PC shortages and disposal of outdated units. Implementation of IBS with 7day storage was expected to decrease fluctuations, have a positive impact on wastage and to reduce incidences of PC shortage.

Aims: To compare PC use in Iceland in two periods, before and after implementation of IBS in 2012, including age of transfused units and number of units transfused per year and per patient.

Methods: Data on platelet ordering were extracted from the ProSang blood bank information system. The number of PC ordered per year, number of PC transfused per patient, PC age (in days) at transfusion, and PC platelet content were compared in two five year periods, before (2007‐2011) and after (2013‐2017) implementation of PR.

Results: There was a shift in PC age at time of ordering when periods before and after implementation of IBS were compared, as the mean PC age at time of ordering before IBS was 2.0±0.24days (range 0‐5days) compared to 4.16±0.23days (range 0‐7days) after IBS implementation (P=1.93×10^‐8). The number of PC per year at our center did not change after implementing IBS, as the number of platelet concentrates ordered annually before PR was 1682±383 (range 1035‐2087 units) compared to 1923±59 (range 1763‐2090) after PR (P=0.19). Furthermore, the mean number of PC per patient did not change, as the mean use of PC per patient per year before PR was 5.42±0.35 units/patient (range 4.93‐5.83 units/patient) compared to 5.86±0.38 (range 5.39‐6.17 units/patient) after PR (P=0.07). The platelet content of buffy coat PC was significantly lower after PR implementation (274±37.7×10e11 vs. 364±20.6×10e11 platelets/unit, P<0.001), whereas the platelet content of apheresis PC did not change (305±9.9 vs. 300, ±16.1, P=0.57).

Summary/Conclusions: Pathogen reduction resulted in the transfusion of older PC on average, but without altering the number of PC ordered or the use of PC per patient. Pathogen reduction has improved PC stock management without an increase in platelet demand, despite lower platelet content of buffy coat PC after PR implementation.

KEEPING THE BLOOD FLOWING

L Eberhart

Donor Management, Austrian Red Cross, Wien, Austria

The transfusion procedure is the last step in a multi‐process supply chain. The task of matching supply with demand requires donor managers to consider average consumption rates on a weekly or monthly basis, but to also have insight into variability in order distribution and possible attribute (blood groups) requirements. Since hospitals and blood banks are usually not deeply interwoven and often only ex‐post data is available, forecasting methods should be implemented. A thorough analysis of order pattern to set weekly target inventories and safety levels is required to close the information gap. A collection plan needs to identify possible bottlenecks which can be prevented through the planning of inter‐shipping, changes in message urgency and building of reserve donor pools.

Constant analysis of collection and mobilization KPIs allows donor managers to implement the rolling‐wave planning approach and continually adapt to changing requirements, unexpected events and overall systematic variability. The variability happens on the demand side, as order quantities and their attributes, such as blood group distribution, are subject to change. However, also the supply is subject to significant variation, as donor response rates, attrition, deferrals and overall availability of donors are not constant. Unfavorable combinations of variable factors often lead to bottlenecks if only long‐term average values find their way into the planning process.

THE BEHAVIOURAL CHARACTERISTICS OF FIRST‐TIME BLOOD DONORS IN TURKEY: AN EXTENSIVE ANALYSIS OF THEORY OF PLANNED BEHAVIOUR MODEL

NN Sozmen, M Ulukanligil, S Caglak, S Coplen

General Directorate of Blood Services, Turkish Red Crescent, Ankara, Turkey

Background: Totally 7,338,696 donors had donated blood in Turkey between 2009 and 2017, with a donation frequency decreases proportionally (Ulger, 2019, Vox Sanguinis ISBT Series). The percentage of donors who donated blood only once is 51.8%, however who donated sixteen times and twenty‐four times are 0.14% and 0.005% in 8years. These results indicate that the first‐time donors constitute a large part of the donor pool of Turkish Red Crescent. It is necessary to understand donors’ behaviour and the factors motivate them to donate blood again after first donation for improving the retention of donors.

Aims: In this study, we focused on donor retention by examining a number of factors that may contribute to first‐time donors’ intentions to donate blood again using structural equation modelling to present the relationships among the extended Theory of Planned Behaviour (TPB) in the literature and additional factors.

Methods: The data was collected with the face‐to‐face interview method right after the donation. 1478 first‐time donors has attended to the study in 18 Regional Blood Centres in 29 cities in Turkey.

The survey included items in accordance with the standard TPB predictors of attitude, self‐efficacy, and intention. Self‐identity, anticipated regret, donation anxiety, paraphernalia anxiety, personal moral norm, descriptive norm, satisfaction, motivation also assessed for the first‐time donors.

The relation between the predictors and intention confirmed with correlation analyses. The predictors’ distribution analysed by multiple linear regression. A number of goodness‐of‐fit indices were calculated and examined for each tested models (IBM, AMOS SPSS).

Results: The predictors significantly correlated with intention. The basic TPB model was compared with the proposed model derived from the results of France's model (France, Transfusion, 2007) afterwards with Masser's model (Masser, Transfusion, 2009) (Cmin/Df: 354.94, 40.505 and 29.785, respectively). The results of goodness‐of‐fit tests for proposed model provided a better fit to the data than these models (Cmin/Df=14). Moreover, this result indicated that the fit between the proposed model and the data could be improved with further modifications with the inclusion of paths between motivation and attitude, self‐identity and intention. Moreover, inclusion the paths between donation anxiety and intention and between self‐efficacy and attitude, on contrary to recent analyses suggesting opposite paths.

Evaluation of goodness‐of‐fit tests showed good result for revised model with a value of Cmin/Df=3.55, close to perfect fit. The revised model revealed that attitude was the strongest positive direct predictor of intention followed by personal moral norm, self‐identity, motivation and anticipated regret (path coefficients: 0.47, 0.28. 0.19, 0.17, and 0.5, respectively). Donation anxiety was the negative direct predictor of intention (‐0.05). Satisfaction was the strongest positive indirect predictor of intention via attitude and followed by self‐efficacy (0.90 and 0.08). Paraphernalia anxiety was the negative indirect predictor of intention (‐0.03). Descriptive norm did not show any significance. Our model accounted for 75.3% of the variance in intention.

Summary/Conclusions: These findings suggest several potential avenues for enhancing donor retention. The results obtained with this study provide important data from the standpoint of donor retention, which should be, implemented in the future strategies of Turkish Red Crescent.

TOWARDS MORE UNIFORM DONOR ELIGIBILITY CRITERIA: OVERVIEW OF FIRST RESULTS FROM TRANSPOSE

G Mori, E Merz, K van den Hurk, S van Walraven, M van Kraaij

Sanquin, Amsterdam, Netherlands

Background: TRANSPOSE‐TRANSfusion and transplantation: PrOtection and SElection of donors, is a European consortium project, including partners from 16 countries, reviewing donor selection and protection policies for substances of human origin (SoHO).One of the main issues in the current donor selection system, which TRANSPOSE aims to tackle, is that for many, if not most criteria, is not evidence based. The TRANSPOSE consortium therefore tries to re‐assess selection criteria, revised them where needed and provide recommendations as evidence‐based as possible. TRANSPOSE additionally adds to the current European Directorate for the Quality of Medicines & HealthCare (EDQM) Guidelines by emphasizing donor safety.

Aims: The aim is to compare existing donor eligibility criteria throughout Europe, and to compile a list of risks to consider, with evidence‐ or consensus‐based deferral criteria to provide more uniform donor screening criteria.

Methods: There are three horizontal work‐packages (WPs); WP1 Coordination, WP2 Dissemination, and WP3 Evaluation of the project, and four technical ones with specific deliverables and milestones to be regularly produced:

• ‐

  WP4 Inventory of Donor Selection & Protection Practices;

• ‐

  WP5 Development of risk-based Guidelines for donor selection and protection;

• ‐

  WP6 Development of a Standard Donor Health Questionnaire (DHQ);

• ‐

  WP7 Training Course/Workshop on the Use of the Guiding Principles, Guidelines and the DHQ.

The TRANSPOSE project launched in September 2017 and will complete in Spring 2020. WP4 has completed its work in October, WP5 will complete its work in June 2019, and WP6 and WP7 have recently commenced.

Results: With the use of the deliverables created by WP4, we have created an in‐depth inventory of current practices in donor selection and protection, including overview of similarities and differences across European countries and across SoHO types. There is an agreement amongst experts that existing guidelines are often based on the precautionary principle rather than on risk assessment. Consequently, in the development of WP5's Guidelines for donor selection and protection, we now make an effort to also emphasize donor safety, in a more evidence‐based way via the use of risk‐based assessments. This will result in a standardized DHQ with a common trunk and more in ‐depth questions per SoHO.

Summary/Conclusions: The impact of the outcomes of TRANSPOSE will be threefold. First, outcomes are expected to be of help in revising donor selection and protection related EU Directives. Second, the set of guiding principles and Donor Selection & Protection Guidelines will facilitate EU member states to take a next step in implementing donor selection and protection policies in a consistent and clear‐cut way to the benefit of both donors and recipients of SoHO. Third, a standard Donor Health Questionnaire with carefully guided local/regional/national adjustments will become available per SoHO which can be used widely and will consequently enable comparisons of the prevalence of certain risks and risky behaviours throughout Europe.

CURRENT PRACTICES IN DONOR SELECTION AND PROTECTION ACROSS EUROPEAN COUNTRIES AND SUGGESTIONS FOR IMPROVEMENT – THE CASE OF BLOOD AND PLASMA

E Merz¹, G Mori¹, K van den Hurk¹, S Van Walraven², M Van Kraaij²

¹Donor Studies ²Donor Affairs, Sanquin Blood Supply, Amsterdam, Netherlands

Background: TRANSPOSE–TRANSfusion and transplantation PrOtection and SElection of donors is a European consortium project, including partners from 16 countries, that reviews donor selection and protection policies for blood, plasma, tissues, assisted reproductive technology (ART) and stem cells (together SoHO). Donor selection criteria (DSC) in Europe are based on EU‐directives, guidelines and countries’ own additional criteria. Literature shows that particular criteria are outdated or not risk‐based, often leading to unnecessary donor deferral or an underestimation of risks for donors.

Aims: To 1) provide a comprehensive inventory of current systems for selection and protection of donors and donations, 2) critically review them and 3) recommend an over‐arching Donor Health Questionnaire (DHQ) including all necessary criteria currently used by different EU‐Member states (EU‐MS).

Methods: In‐depth semi‐structured interviews with key stakeholders in blood collection were conducted to identify main topics for improvement in the current DSC. These formed the basis for a survey sent to professionals from collection institutions of all SoHO to get feedback on current systems from as many EU‐MS organisations as possible. Questionnaires were sent to a total of 163 experts (40 blood; 40 plasma; 27 tissues; 47 stem cells; 9 ART) and 39 (24%) completed questionnaires were received. Where information was lacking, additional experts were asked to recommend upon DSC.

Results: For blood and plasma donation four main areas of concern in DSC were identified: risk‐based selection, adaptability, flexibility and consistency. The stakeholders agreed that DSC are often outdated and lack evidence, hence leading to unnecessary deferral of donors and underestimated risks for donors. They suggested to base DSC on group risk‐assessment (risk‐based selection) and on conducting more research to achieve standardized risk perceptions and evidence‐based deferrals, either for safety of recipient or donors. Criteria could be made more detailed to fit specific groups to defer less donors (adaptability). Furthermore, implementing criteria was considered easy, but abolish criteria when not regarded as a risk anymore seems almost impossible (flexibility). Additionally, deferral periods are perceived too long, seen as both negative, i.e. jeopardizing donor return intention and positive, i.e. no risk for safety (consistency). Changing legislation into guidance was an often‐mentioned suggestion to improve DSC.Specific feedback on plasma donations revealed that many whole blood topics are not applicable to plasma‐only donors, e.g. parasite infections such as malaria (no deferral needed); travel history (no deferral needed), and recent bacterial and viral infections (deferral periods currently too long). A clear need for more research on plasma collection‐related issues was identified.

Summary/Conclusions: DSC are perceived redundant on a substantial number of aspects by most stakeholders. Besides achieving the goal of save and sufficient SoHO for patients, many regulations could be improved to diminish deferrals and decrease donor risks. TRANSPOSE will add to reviewing, improving and harmonising these regulations and criteria. Furthermore, TRANSPOSE will provide suggestions to improve directives and guidelines and a DHQ, focusing on both donor health protection and safety of donations, but also removing deferral criteria that are not relevant (anymore), and offer a future research agenda to make DSC more evidence‐based.

LARGE DIFFERENCES IN THE REPORTING OF ADVERSE REACTIONS IN BLOOD AND PLASMA DONATION WITHIN THE EUROPEAN UNION – RESULTS FROM THE TRANSPOSE PROJECT

C Mikkelsen¹, J Castrén², C Eguizabal³, J Fernández‐Sojo⁴, S Fontana⁵, M Hansen⁶, K van den Hurk⁷, M van Kraaij⁸, R Kullaste⁹, A Navarro Martinez‐Cantullera¹⁰, G Mori¹¹, F Urbano³, M Vesga³, E Veropalumbo¹², S Zahra¹³, H Ullum¹

¹Department of Clinical Immunology, Copenhagen University Hospital, Copenhagen, Denmark ²Finnish Red Cross Blood Service, Helsinki, Finland ³Basque Center for Blood Transfusion and Human Tissues, Galdakao ⁴Blood Bank and Tissue of Catalonia, Barcelona, Spain ⁵Interregional Blood Transfusion SRC, Berne, Switzerland ⁶Blood Center Copenhagen, Copenhagen, Denmark ⁷Sanquin Research ⁸Sanquin Blood Bank, Depts Transfusion Medicine and Donor Affairs, Amsterdam, Netherlands ⁹North Estonia Medical Centre's Blood Centre, Tallinn, Estonia ¹⁰Organització Catalana de Trasplantaments (OCATT), Barcelona, Spain ¹¹Sanquin, Amsterdam, Netherlands ¹²Italian National Blood Centre (CNS), Rome, Italy ¹³Scottish National Blood Transfusion Service, Edinburgh, United Kingdom

Background: TRANSPOSE ‐TRANSfusion and transplantation: PrOtection and SElection of donors, is a European Commission co‐funded project with participation of 24 stakeholders from both not‐for‐profit and private blood collecting organizations as well as researchers and officials. The project aims to create new evidence‐based donor selection criteria as well as guiding principles for risk assessment of threats to the safety of all substances of human origin (SoHO) except solid organs. As part of this, an inventory of current donation‐related risks was performed, including an investigation of both type and number of adverse events reported.

Aims: We here aim to present an overview of reported adverse events in plasma and whole blood donation in Europe and to compare this to the anticipated risks rated by TRANSPOSE stakeholders.

Methods: National or local data on adverse reactions from the years 2014‐2017, both serious and mild, in whole blood and plasma donors was collected from the relevant stakeholders (eighteen and nineteen respectively). Stakeholders were also asked to grade the most important anticipated donor risks according to severity, level of evidence and prevalence. We then compared the relevant risk categories as evaluated by the stakeholders with the categories of the provided data, as well as the heterogeneity of category numbers.

Results: Thirteen stakeholders provided data on adverse events during whole blood donation in a given year, including in total thirty‐three different categories of adverse events, ranging from only one unspecified reaction to seventeen different categories, with an average of nine categories per stakeholder. The most frequently used categories were hematoma (included by 77%), arterial puncture (77%) and nerve damage (54%). Vasovagal reactions were also frequently included (75%); however, this was being done variably as vasovagal reactions unspecified, and acute and/or delayed vasovagal reactions. Only one stakeholder reported iron deficiency.

For plasma donation, seven stakeholders provided data on adverse events. A total of twenty‐seven different categories were reported, ranging from one to seventeen per stakeholder, with an average of nine. The most frequently reported adverse events were hematoma (86%), citrate reactions (86%) and arm pains and nerve damage (both 57%, respectively).

Anticipated risks in blood donation were rated by nine stakeholders rating iron deficiency, vasovagal reactions and hematomas the greatest risks to donors. For plasmapheresis, six stakeholders rated vasovagal reactions, hematomas and citrate reactions as highest risk.

Summary/Conclusions: As shown, categories used to describe adverse events in blood donation vary tremendously across Europe, with some countries only being able to provide total numbers of adverse events without further specification. Furthermore, there is a gap between perceived high donor risks and reported adverse reaction categories in donor vigilance for whole blood, as reports on iron deficiency are virtually absent despite being considered the most significant risk.

Our findings show the need for international collaboration on creating an international standardized donor vigilance system, to gather more insight into donor risks to protect the health of donors.

INVISIBLE ORGANS

R Blasczyk

Institute for Transfusion Medicine, Hannover Medical School, Hannover, Germany

Modern transplantation medicine has made significant progress within the last decades due to a better immunological understanding of rejection and advances in immunosuppression. However, the severe side effects of long‐term, typically lifelong, immunosuppression and the shortage of donor organs remain the major restrictions in transplantation. The idea behind all research to improve transplant outcome has always been the modification of the recipient's immune system to ideally induce a specific tolerance towards the donor's graft. In fact, the immunological blindness of the recipient towards the donor's graft is achieved by a general reduction of the immune system's competence and represents a major burden for transplant patients. The idea of invisible organs is an entirely different approach to solve the problem: instead of inducing an immunological blindness of the recipient's immune system an immunological invisibility of the donor's organ is created. This is achieved by genetically engineering the transplant to eliminate the organ's immunogenicity defined by the gene products of the major histocompatibility complex (MHC) and minor histocompatibility antigens. In addition to manipulating the expression of MHC genes required for immune recognition, immune cloaking strategies are used to evade immune rejection. These approaches take advantage of creating an immunosuppressive environment and expressing immune suppressive molecules by immunomodulatory transgenes. MHC engineering and immune cloaking in an entire organ is achieved during ex vivo perfusion by lentiviral transduction of gene expression modifiers and transgenes to induce a permanent immunological invisibility of the organ. Importantly, MHC engineering also prevents the presentation of minor histocompatibility antigens, which usually are not possible to match between donor and recipient, but which trigger potent immune responses and graft rejection. Eliminating the targets of cellular and humoral rejection as well as creating an allograft‐specific immune environment through immune cloaking camouflages the organ and equips it with a powerful set of defense weaponry. Immune‐engineering of transplants achieved during the inevitable ex vivo period of the allograft after explantation without the need to accept off‐target effects allows keeping the recipient's immune system fully functional and capable to combat infections and cancer. In pre‐clinical in vivo studies from rodents to minipigs a clear survival advantage of ex vivo engineered transplants could be demonstrated. This approach has the potential of eliminating the burden of organ rejection and immunosuppression, thereby sustainably increasing transplant survival, organ availability and quality of life.

GENE EDITING IN SICKLE CELL DISEASE

D Bauer

Pediatric Hematology, Boston Children's Hospital, Boston, United States

Gene editing for sickle cell disease

Re‐expression of the fetal γ‐globin genes (HBG1/2) could be a universal strategy to ameliorate the severe β‐globin disorders sickle cell disease (SCD) and β‐thalassemia by induction of fetal hemoglobin (HbF, α2γ2). We have previously identified BCL11A erythroid enhancer sequences, marked by HbF‐associated common genetic variants, that are required for repression of HbF in adult‐stage erythroid cells but dispensable in non‐erythroid cells. Recently we have optimized conditions for selection‐free on‐target CRISPR‐Cas9 editing in human HSCs as a nearly complete reaction without detectable genotoxicity or deleterious impact on stem cell function. We demonstrate that Cas9:sgRNA ribonucleoprotein (RNP) mediated cleavage at core sequences of the+58 BCL11A erythroid enhancer results in highly penetrant disruption of GATA1 binding motif, reduction of BCL11A expression, and induction of fetal γ‐globin. Erythroid progeny of edited engrafting SCD HSCs express therapeutic levels of HbF and resist sickling, while those from β‐thalassemia patients show restored globin chain balance. Moreover we find that HSCs preferentially undergo nonhomologous as compared to microhomology mediated end‐joining repair. NHEJ‐based BCL11A enhancer editing approaching complete allelic disruption in HSCs appears to be a feasible therapeutic strategy to produce durable HbF induction. In this presentation, I will compare and contrast BCL11A enhancer editing to other autologous curative gene therapy and gene editing approaches at various stages of clinical and pre‐clinical evaluation.

THE WATT WORMS – TAKE A DEEP BREATH

F Zal

Hemarina SA, Morlaix, France

Oxygen is vital for life. Without oxygen death is assured for aerobic organisms. Although everybody knows this fact a lot of medical acts forget to take care of it, leading to a lot of potential troubles. Indeed, during cell respiration the glucose oxidation by oxygen gives carbon dioxide, water and energy. This energy also called ATP is necessary for cellular metabolism and consequently for life. We have identified an extracellular hemoglobin coming from a marine worm, called Arenicola marina, which is able to deliver oxygen to this animal living in the intertidal areas on the Atlantic coast in France between the North Sea and Biarritz. This molecule called M101 was developed in the medical device named HEMO2life^®. We have showed that this product was very efficient to protect organs before transplantation. A multi centers clinical trial performed under the supervision of Pr. Le Meur from the CHU of Brest, on 60 patients waiting kidney grafts showed a delay graft function reduced roughly by three between the two kidneys harvested on the same donor with and without HEMO2life^® and grafted on recipients. In 2018, a world first was realized in France by the Pr. Lantieri to Georges Pompidou hospital in Paris, France. Indeed, it was the first time that a patient received a second graft face. This surgery was realized with HEMO2life^® and showed a very nice result according the Pr Lantieri, the anastomosis were very easy and no edema was observed. Furthermore, we have developed dressing incorporating M101 making a product called HEMHealing^®. Preclinical data on diabetic mice showed an increase of healing process. HEMOXYCarrier^®, a therapeutical oxygen carrier, is also in progress of development in order to address ischemic diseases such as the sickle cells disease, myocardiac infraction and stroke. This universal oxygen carrier without blood typing, which is the ancestor of our Red Blood Cell containing hemoglobin showed that it is able to deliver oxygen at different biological levels, cellular, tissues and organs and could address a multitude of medical applications.

WHO IS A DONOR RECRUITER; AN UP DATED PERSPECTIVE

NN Solaz, M Bayik, R Uluhan, G Emekdas, N Pelit

Turkish Blood Foundation, Istanbul, Turkey

Background: Main goal of transfusion is saving life and/or improve the health status of human by “safe blood” which needs regular, voluntary, unpaid blood donors. Donor recruitment is being more sensitive and challenging part of the blood supply system in actual global socio‐economic conditions. Achievement to enough voluntary non‐remunerated blood donation (VNRBD) can be established by an efficient donor recruitment. Efficiency of the donor recruitment has still close relation with blood donor recruiter although there are so many new tools. Occupational specifications, rights and responsibilities of blood donor recruiter have wide range differences between countries which cannot be explained completely by the specific conditions of each country. Also, a concrete document which has an international consensus was not existing on this subject.Turkish Blood Foundation (TBF) has been organizing an international workshop since 2012; Anatolian Blood Days (ABD). “Who is a blood donor recruiter?” was the topic of ABD‐VII at 9–11 March 2018.

Aims: Main aim of the workshop was to check and evaluate the existing systems of the participant countries. Than create a model for clearly defining occupational specifications, responsibilities and rights of blood donor recruiter.

Methods: Experts from 25 countries participated in the workshop. Those countries are Albania, Algeria, Bosnia‐Herzegovina, Estonia, France, Germany, Hungary, India, Kazakhstan, Lithuania, Macedonia, Montenegro, Oman, Portugal, Qatar, Romania, Russia, Saudi Arabia, Serbia, Slovenia, Sri Lanka, Tajikistan, Turkey, Uganda, Uzbekistan. These countries reflect almost all religious, ethnical, social, cultural and economic situations of the world. A questionnaire which was analyzing existing systems at participant countries sent before the workshop. After country presentations 4 different discussion groups were organized. Below listed topics were announced at final declaration.

Results: Donor recruiter:

1. should have University degree preferably in Marketing and Business Administration field.

2. should have a certificate and/or professional experience in Public Relation

3. should have efficient skill in conversation, sociability, independence, self-confidence, reliability, resilience and conscientiousness as well as to work in a team

4. should get a special training which includes not only social topics such as public relation, marketing, etc. and medical topics related BB&TM before practicing alone as a donor recruiter

5. should be a permanent staff

6. should have basic salary and performance bonus might be given

7. is eligible to monitor and modify mobile team working period at blood drive

8. should participate the mobile blood drive which he/she has organized

9. should participate the group who will create promotional materials for national blood service

10. number at each blood establishment should be defined based on annual blood collection such as 5 staffs for 50,000 whole blood collection annually in Germany.

Summary/Conclusions: In conclusion; both donor recruitment and retention are not easy tasks to undergo while public are aging, and birth rates are decreasing all around the world. Dedicated blood donor recruiter whose occupational specifications, rights and responsibilities are clearly defined will be the corner stone of the success for providing enough safe blood for transfusion.

IS AFRICA ON THE WAY TO A SAFE AND SUFFICIENT BLOOD SUPPLY?

CT Smit Sibinga¹, J Emmanuel²

¹IQM Consulting, ZUIDHORN, Netherlands ²BTS Consultants, Harare, Zimbabwe

Background: Africa is a large continent with 55 independent States and a total population of 1,275,710,034 (February 2018). Healthcare policies and strategies are developed through WHO's advocacy, guidance, and support from HQ in Geneva and the 2 WHO Regional Offices; Eastern Mediterranean Regional Office (EMRO) supporting 8 Arabic speaking countries and the African Regional Office (AFRO) responsible for 47 Sub‐ Saharan Countries. Population distribution is approximately 40.6% urban. There are a large number of different local dialects and languages spoken. The main languages spoken are English, French, Portuguese, Spanish and Arabic. Countries are mainly classified by UNDP as being of low and medium Human Development Index

The Africa Society for Blood Transfusion (AfSBT) has Members in most countries, advocates for the development of sustainable and effective blood services, and has developed a stepwise 3 level accreditation program.

In 2016 EMRO held a consensus meeting developing a “Strategic Framework for Blood Safety and Availability for 2016–2025” with a set of priority interventions focusing on Leadership and Governance, Cooperation and Collaboration, Provision of Safe Blood and Blood Products, Appropriate Clinical Use of Blood, and Quality System Management.

In 2001 all 54 Member States of the African Union (AU) countries, in Abuja, Nigeria, pledged that national budget for Health should be at least 15% of the National Fiscal Budget.

In 2013 Ministers of Health of WHO Member States endorsed that blood and blood products be included in the Essential Medicines List; these endorsements and WHO's Universal Health Coverage (UHC), have yet to be fully implemented.

Aims: To analyze (gap‐analysis) to what extend countries in Africa implement the World Health Assembly Resolution WHA63.12 on Availability, Safety and Quality of Blood Products, which urges governments to ensure safe, accessible, affordable and available supplies of blood and components from voluntary non‐remunerated blood donors, which meet clinical transfusion requirements and achieve national self‐sufficiency, following WHO guidelines and recommendations.

Methods: To provide an overview of the current status of the blood supply in Africa – strengths and weaknesses, data from WHO's 2016 Global Status Report on Blood Safety and Availability were analyzed and used. The study has been descriptive and explorative.

Results: The 2016 Report identified a number of areas requiring attention; principle amongst these were

• ‐

  inadequate funding;

• ‐

  lack of governance and leadership;

• ‐

  ineffective public education on blood donation;

• ‐

  absence of capacity building for clinicians on rational use of blood;

• ‐

  lack of Haemovigilance and implementation of quality management systems;

• ‐

  the need for regulatory or oversight mechanisms.

Summary/Conclusions: National Authorities should address areas requiring attention if progress towards ensuring a sustainable safe and sufficient supply of blood products is to be achieved. Key is the commitment and support of National Governments, which should implement Resolutions and Recommendations agreed by Ministers of Health at WHA and African Union.

REVIEW ON STAFF ROSTERING FOR BLOOD DONATION TESTING (BDT) LAB

B Mohamed Siddique, K Chia, C Goh, Z Huang, W Tan, J Liew, X Su, T Leong, S Chua, S Lam

Blood Services Group, Health Science Authority, Singapore, Singapore

Background: The core function of the Blood Donation Testing (BDT) Laboratory is to screen every unit of blood collected from a donor for blood group type and infectious disease markers to ensure safety of the national blood supply prior to transfusion. The lab operates daily on two work shifts, comprising of 4 staff on the morning (AM) shift (from 8:00 to 17:30) and 5 staff on the afternoon (PM) shift (from 13:00 to 22:00) on weekdays and 3 staff on the AM shift and 5 staff on the PM shift on weekends. BDT Lab has a staff strength of 15 to be rostered for the 2 work shifts. Each staff is on a five‐day work week and has to work 11 PM shift and 9 AM shift per month on average. The higher number of PM shift leads to staff feedback that they do not get sufficient time in the evenings for family and social or leisure activities. A Lean Six Sigma project was initiated to review the work rostering to improve the work‐life balance of the staff.

Aims: The project aims to reduce the number of staff working on the PM shift without affecting the downstream processes and continues to meet the timely release of blood supply to the hospitals.

Methods: Lean Six Sigma tools were used to study the BDT Lab workflow process and to identify factors that contributes to the higher number of PM shifts that the staff has to take on. Data on the turn‐around time and the man‐effort required for each screening tests performed was analyzed. A survey was also conducted to understand the preference of the staff on the acceptable number of PM shifts per month.

Results: The main contributing factor for more staff required to perform the PM shift is due to majority of the daily donation samples being received only in the evening. As this factor is beyond the control of the BDT Lab, redeploying work from the PM shift to the AM shift was eventually selected as the solution to reduce the number of staff needed for the PM shift. The screening test that was shifted was determined based on the test system that has the shortest turn‐around‐time and is able to allow continuous release of results. At the same time, most of the staff must be trained for that test system. A trial on the new roster involving 5 staff on AM shift and 4 staff on PM shift was conducted. The total number of PM shift per month was reduced from 124 to 112 using the re‐defined process. The 10% reduction translates to fewer number of PM shifts that the staff has to undertake and was able to meet the staff's expectation.

Summary/Conclusions: With the adoption of the new process workflow, BDT Lab was able to reduce the number of PM shifts that the staff needs to be rostered using evidence based process improvement method. Most importantly, the lab has a satisfied team of staff with better work‐life balance.

A MODEL FOR CRISIS MANAGEMENT IN EMERGENCY TRANSFUSIONS: TRANSFUSION EMERGENCY KIT FOR DISASTER & CRISIS SITUATIONS

FY Ayhan^1,2, H Sarihan²

¹Transfusion Center ²Haemovigilance Department, Dr Behcet Uz Children's Hospital, Izmir, Turkey

Background: Preparedness of blood transfusion services for emergencies and crisis situations is an important issue concerned with patient and transfusion safety.

Aims: Having an experience of delay in blood component supply in an emergency situation due to partial interruption of Hospital Information System (HIS), it is aimed to create a crisis kit and constitute an alternative work flow for emergency in crisis situations.

Methods: It is stipulated that the failure of HIS which is normally conducts all process for transfusion would be disabled in a disaster or crisis situation. A brain storm was made on possible challenges associated with disability of HIS during transfusion emergencies. According to the scenarios a kit was developed for the process management of transfusion emergencies.

Results: A flow chart was designed in proper with transfusion emergency definitions of WHO and instructions were written to explain the flow chart. All forms categorized with different colour codes are designed to fill with handwriting. The kit consists of flow chart and instructions, Analysis Request Forms (blue coloured), Blood Component Request Forms (pink coloured), Proceeding Forms (green coloured), pens and blood sample tubes with EDTA were put into a plastic folder labelled as Transfusion Emergency Disaster & Crisis (TEDC) Kit. Additionally, the kit is placed in a sealed clear plastic bag and delivered to all inpatient and intensive care units of pediatrics and pediatric surgery.

A training programme concerned with transfusion emergency situations and usage of the TEDC kit was developed for health care workers involved in blood transfusion process. Pre and post‐assessment tests were developed for the evaluation of effectiveness of the training programme.

Summary/Conclusions: It's challenging to improve the responsecapacity ofblood transfusionservices during emergencies and crisis situations.

COMPUTERISATION IN BLOOD BANKS: CHALLENGES AND SOLUTIONS

T Chandra¹, D Agarwal², M Agarwal³, R Agarwal⁴

¹Department of Transfusion Medicine, King George Medical University, Lucknow, Lucknow ²GSVM Medical College ³Era Medical college, Kanpur ⁴St. Francis’ College, Lucknow, India

Background: India is a developing country having 2760 licensed blood banks Majority have manual documentation which causes inaccuracies and errors in blood bank activities. Monitoring is a herculean task.Computerisation is the need of the hour but this goal involves many hurdles and challenges

Aims: The aim of this study is to discuss the challenges faced during computerisation of blood bank activities and the remedial solutions for it

Methods: Department of Transfusion Medicine, King George medical University, Lucknow is one of the biggest blood bank of the country with annual collection of 70,000 blood units. Two years ago, the blood bank worked on totally manual system. Computerisation involved challenges associated with hardware and software installation and personnel training. Hardware was installed in two phases. Initially HP system but later shifted to Apple iMac due to frequent breakdowns. With HP server. Software installation (Easy software) involved erratic internet connectivity hence changed to LAN. Customisation involved radical changes according to our needs. At times we had to change our way of working to suit the software. Biometrics linking, medical registration, cash id generation, donation, serological crossmatch, automated blood grouping, labelling, chemiluminescence & NAT testing, blood component preparation and camps were all included with challenges at every level. Remedial actions were taken from small to big. Training of the staff was the most essential part of the implementation of computerisation who initially showed considerable resistance and at time faking ignorance due to apprehension that their mistakes will be highlighted and they will be penalised for it. It was a herculean task in creating their password protected identity and enforcing them to use it. Gradually the staff realised that computerisation made their task easier as it cut down on paper work and repetition and also prevented serious mistakes from happening. Hard copies at certain essential areas were still maintained to continue work in the event of major breakdown of computer

Results: Computerisation aided us in regulating the movement of the donor which at times was repetitive due to manual entry. Transfer of data ensured a safe supply and the mistakes could be retraced very easily. Implementation which included installation, training and enforcement took a period of 6months. After overcoming all the challenges we minimised hard copies to 6 registers and started taking printouts of the other necessary details.

The turnover time for the employees due to computerisation decreased by 20%. Waiting time for attendant decreased by 10%. Traceability of all the units became 100%.Supervision of the activities being carried out was 90% accurate. Identification of the donors was easy due to biometrics which included thumb impression and iris scanning. The decision making time for donors decreased by 50% thus making the system more efficient.

Summary/Conclusions: Manual to computerisation involves involvement from source to supply and it is essential to anticipate the challenges and be prepared for solutions in order to make its implementation successful

BLOOD SYSTEMS OF COUNTRIES IN THE WHO EASTERN MEDITERRANEAN REGION – EXISTING LEGISLATIVE INSTRUMENTS

CT Smit Sibinga¹, Y Abdella², F Konings²

¹IQM Consulting, ZUIDHORN, Netherlands ²WHO Eastern Mediterranean Office, Cairo, Egypt

Background: WHO defined essential medicines (EMs) as medicinal products that satisfy health‐care needs of the majority of a population. They should be available at all times, in adequate amounts, in appropriate dosage forms, with assured quality and affordability. In 2013 blood and blood products (whole blood, red cells, platelets, plasma, and plasma‐derived products) were added to the WHO Model List of EMs.Appropriate and effective regulatory framework (legislations, regulations, etc.) and a functioning regulatory authority (RA) is crucial for management of blood products as EMs.However, particularly in the less developed world, these prerequisites have barely been implemented.

Aims: To analyze and advise on existing legislation and regulations.

Existing legislative instruments of the 22 Member States of WHO Eastern Mediterranean Region (EMR) were collected and analysed for relevance and appropriateness for preparation and use of blood and blood products as well as use of associated substances and relevant medical devices. A literature search was done on matching combinations of regulatory system, regulatory framework, legislation, regulation, with production and use of blood and blood products, which resulted in almost exclusively references with respect to national and international legislation. Benchmark: WHO recommendations (Aide Mémoires) and EU Directives.

Methods: Existing legislative instruments of the 22 Member States of WHO Eastern Mediterranean Region (EMR) were collected and analysed for relevance and appropriateness for preparation and use of blood and blood products as well as use of associated substances and relevant medical devices. A literature search was done on matching combinations of regulatory system, regulatory framework, legislation, regulation, with production and use of blood and blood products, which resulted in almost exclusively references with respect to national and international legislation. Benchmark: WHO recommendations (Aide Mémoires) and EU Directives.

Results: Various formal legislative documents of only 9/22 countries are put in force by Governments [1960 (Egypt) till 2017 (Pakistan – Sindh)]. Most are detailed descriptions of RA, operational establishments, and specific requirements. However, none of these legislations complies with WHO and EU recommended formats and contents, and will not support effective regulatory oversight to promote and enhance quality, safety and availability of these EMs.

Summary/Conclusions: Government should provide effective leadership and governance in developing a national blood system (NBS, fully integrated into the national health‐care system. Essential functions of a NBS include an appropriate regulatory framework with legislations, regulations and other non‐legislative instruments administered by a RA. These documents should spell out principles and cadres, standard setting, and organization of the blood system to ensure an adequate supply of blood and blood products and safe clinical transfusion for which a Model was designed.

The structure of NBS will depend on organization and level of development of the health‐care system. However, all critical activities within NBS should be coordinated nationally to promote uniform standards, economies‐of‐scale, consistency in staff competency, quality and safety of these EMs, and best transfusion practices. Key is formulation of an appropriate regulatory framework administered by a RA responsible for regulating the vein‐to‐vein chain in the preparation and use of these EMs.

BLOOD TRANSFUSION SERVICE IN POLAND IN THE PERIOD 2008–2017

A Rosiek, J Antoniewicz‐Papis, E Lachert, A Tomaszewska, M Letowska

Department of Transfusion Medicine, Institute of Hematology and Transfusion Medicine, Warsaw, Poland

Background: The capacity of blood transfusion service to provide adequate supplies of blood components is the issue of concern for health providers worldwide; longer term observations of trends in this respect are therefore of crucial value.

Aims: The study aim was analysis of some basic activities of the Polish Blood Transfusion Service in 2008–2017 including organizational changes, numbers of donors, donations and blood components as well as activities directed at increasing their safety.

Methods: Retrospective analysis of data supplied by the regional Blood Transfusion Centers (BTCs).

Results: In the discussed period, blood and blood components were collected in 21 regional BTCs and local collection sites as well as during mobile collections.

Although the number of local collection sites decreased from 170 in 2008 to 133 in 2017 in favor of mobile collections, which increased from 8,672 to 13,189, the former is still the number one location for donating blood. On average 47.36% of all donations were performed in local collection sites.

The total number of blood donors both at the beginning and the end of the discussed period was similar (583,908 in 2008 and 588,184 in 2017); over 99% of all donors were non‐remunerated. However, the number of first‐time donors dropped significantly (from 237,658 in 2008 to 143,038 in 2017).

The total number of donations increased from 1,076,655 in 2008 to 1,249,655 in 2017; most frequent were whole blood collections (from 1,016,411 in 2008 to 1,171,302 in 2017). Some blood components (mostly plasma and platelet concentrates) were also collected by apheresis.

Most frequently prepared blood components were red blood cell concentrates – RBCs (996,678–1,154,239 units per year), fresh frozen plasma – FFP (1,090,369–1,289,021units) and platelet concentrates – PCs (81,692–129,143 units, with significant increasing tendency).

Additional processing methods such as leukocyte depletion and irradiation were more frequently applied to PCs (52–57.6% in respective years irradiated, 75.18–92.07% leukocyte‐depleted), than to RBCs (4.46–8.46% irradiated, 7.65–20.7% leukocyte‐depleted).

In 2010, the pathogen reduction technologies in plasma and the PCs were implemented. Up to date however the use of these technologies is limited in most BTCs. In 2017 approximately 11.5% of PCs and 8% FFP units issued for transfusion were subjected to pathogen reduction technologies.

Summary/Conclusions: Our study data may contribute to the assessment of some long‐term tendencies observed in Polish Blood Transfusion Service and may serve practical‐benchmarking. This in turn may prove beneficial to the transfusion community as a whole.

DEVELOPMENT OF GUIDELINES FOR CHANGE MANAGEMENT IN BLOOD TRANSFUSION SERVICE FOR MORE EFFECTIVE SUPPLY OF BLOOD USING BIG DATA AND DATA MINING METHODS – STAGE 1

A Mikołowska, J Antoniewicz‐Papis

Department of Transfusion Medicine, Institute of Hematology and Transfusion Medicine, Warsaw, Poland

Background: In Poland 80% of hospitals depend on blood for the treatment of patients; over 1.5 mln units of blood components are annually transfused. It is therefore purposeful to expand the knowledge on factors impacting on blood transfusion service (BTS). The Institute of Hematology and Transfusion Medicine (IHTM) as competent authority is responsible for collection of data related to the activity of all Polish Blood Transfusion Centers (BCTs). This data is exploited to a much lesser degree than the recently available statistical methods and data processing tools would allow. Moreover, survey of research in the field of public health indicates a negligible share of issues related to BTS. It seemed therefore necessary to “fill in the gap” with true assessment of performance of the Polish BCTs for improvement of BTS activity. 1^st stage of our investigation refers to collection, merging of data from different sources, their unification and preparation (big data) for further analysis to be performed using multidimensional statistical analysis and data mining methods.

Aims: Assessment of the activity of the Polish BTCs over the 20year‐period (1997–2017) in two stages.

Goals at 1^st stage:

1. Data digitalization; scanning of paper documents.

2. Development of a uniform template for collecting digital data from various sources.

3. Standardization, unification and quality improvement of available data: filling in missing data, elimination of errors, duplicate records etc, that may distort the outcome of analyzes.

4. Selection of data for analysis.

Methods: Digitalization and big data methods for processing various types of data:

1. stored in paper form (1997–2004),

2. digital stored in two file types (.doc and .xls, for the years 2005–2010 and 2011–2017, respectively).

For each data‐type, a separate Excel file model was created. The models were then merged into one analytical table with data processing methods.

Results:

• 1

  1344 pages of paper documents were scanned.

• 2

  Models developed for data from 3 different sources:

  1. Paper-data were rewritten and ascribed to its model; outcome – 3 tables, 272 columns, 10,400 rows.

  2. .doc and .xls. files – data were ascribed to 2 other models; outcome – 6 tables, 1656 columns, 788 rows.

• 3

  The 3 models were merged into 1 analytical table to create a 588MB database (comparable to approx. 784min of music).

• 4

  The data was subjected to standardization, unification and quality improvement: filling in missing data, elimination of errors and duplicate records that may distort the results of analyzes.

• 5

  Selection of data for analysis at 2^nd stage.

Summary/Conclusions: The 1^st stage provided a set of selected data for analysis in the 2^nd stage which will rely on multidimensional statistical analysis and data mining methods.

The outcome of such analysis will contribute to optimal realization of objectives:

1. gaining in-depth knowledge about the fundamental phenomena that shape Polish BTS,

2. identification of potential changes BCTs,

3. development of overall guidelines for change management.

POSITIVE IMPACT OF NATIONAL TRANSFUSION SOCIETY; 20 YEARS’ EXPERIENCE OF BLOOD BANKS & TRANSFUSION SOCIETY OF TURKEY

NN Solaz, M Bayik, R Uluhan, G Emekdas, N Pelit

Blood Banks & Transfusion Society of Turkey, Istanbul, Turkey

Background: Although Turkey has had blood transfusion practice since early 1920s there has been no specific education for Blood Banking & Transfusion Medicine (BB&TM) either in pre or post graduate medical education until 2000. The negative impact of this situation became a major threat for safety of the blood transfusion. A group of dedicated doctors who were involved in the field of BB&TM decided to solve this problem by a civilian initiative and established a national blood society; “Blood Banks and Transfusion Society of Turkey (BBTST)” in 1996. Today; BBTST has 718 members. 75% BBTST's members is medical doctors rest are technicians, nurses, etc.

Aims: The main aims were based on:

1. Closing the knowledge gap of blood bank and clinical staff

2. Establishing official and academic education programs

3. Creating informative publications in Turkish such as text books, hand books, journals, guidelines, etc.

4. Establishing close collaboration with the national health authority and other related parties

5. Integrated with international BB&TM society

Methods:

1. Close collaboration established with Ministry of Health (MoH), Turkish Red Crescent (TRC) and universities.

2. Different curriculums were created with the collaborators for residential national courses, daily courses, symposiums, workshops, etc.

3. Translate EU Guidelines, prepare national guidelines,

4. Participate various committees at MoH, TRC.

5. Organize national congresses

6. Establish close contact with international organizations

Results:

1. 21 national courses, 11 national congress, 20 seminars, 105 symposia and 15 special sessions at the national congresses of different clinical branches have been done so far at different parts of Turkey

2. 2 international congresses; VIII. European ISBT Congress 2003 and XII. AATM – XX. BBTST Congress 2016

3. 3^rd International Thalassemia Summer School 2004, Joint Course with ESTM

4. World Health Organization (WHO) workshop at 2008. Participated WHO Global Forms at 2007 and 2013.

5. Initiated international annual workshops since 2012; Anatolian Blood Days; aimed to touch untouched or less touched topics of BB&TM. So far 8 workshops were organized and each year around 30 countries were participated from almost all religious, ethnical, social, cultural and economic situations of the world.

6. Supported realization of major changes in BB&TM in Turkey;

   1. convincing medical individuals and agencies mainly MoH to give the deserved consideration to BB&TM

   2. encouraging the recognition and establishment of national blood program

   3. issuing a new blood act and numerous necessary bylaws, etc.

   4. creating an appropriate standard donor questionnaire form

   5. changing blood transfusion practice from %96 whole blood (at 1996) to 5%.

   6. changing donated blood screening criteria; while anti-HCV screening became obligatory Malaria, screening cancelled

   7. preparing national guidelines

   8. promoting Haemovigilance nurse post

   9. promoting Patient Blood Management

7. Around 11,000 blood bankers attended national courses, 7,000 attended national congresses, 18,000 attended nationwide symposiums.

Summary/Conclusions: BBTST can be accepted as a sample how a scientific non‐governmental organization may give a very positive impact on developing and progressing BB&TM activities with close collaboration MoH and other related organizations

IMPLEMENTATION OF INFORMATION TECHNOLOGY GOVERNANCE FRAMEWORK FOR BLOOD SERVICES IN RESOURCE CONSTRAINED SETTING

T Mapako¹, C Zaugg², L Marowa³, R Chikwereti⁴

¹Planning, Information and Research Department, National Blood Service Zimbabwe, Harare, Zimbabwe ²Blood Safety Programme, Swiss Red Cross International Cooperation, Berne, Switzerland ³Coordination Department ⁴Finance Department, National Blood Service Zimbabwe, Harare, Zimbabwe

Background: Globally there is growing investment in information technology (IT) in business. This similar trend has been observed in blood establishment computerized systems (BECS). The IT investment can be high hence IT decisions need to be properly informed. The Africa Society for Blood Transfusion (AfSBT) encourages the use of ITs in African blood services as this optimize quality blood services, thereby improving patient's outcomes. AfSBT established in 2016 an IT working group (AfSBT ITWG) with the support of the Swiss Red Cross (SRC) to spearhead the IT standards among member blood services. A number of priority IT thematic areas were identified. These includes IT governance which focuses on creation (strategic alignment) and preservation (risk management) of business value. There is absence of published literature on how a structured IT governance framework can be implemented in a resource constrained setting. A review of the National Blood Service Zimbabwe (NBSZ) IT governance was done based on published IT governance framework.

Aims: To explore how a structured IT governance can be developed, implemented and monitored in a resource constrained setting.

Methods: A published MIT‐CISR framework which has six components was used to assess the strength, gaps and opportunities of the IT governance.

Results: NBSZ has been implementing an evolving structured IT governance system. In terms of Service strategy and organization there is a well‐established IT function which is reflected in the NBSZ strategic plans. This ensures IT annual budgetary support, which averages 4.3% of the total budget. The IT governance arrangements are such that decision rights are assigned to different IT staff (executives, IT specialist, and users). A range of IT solutions have been embedded within the NBSZ operations such as BECS, financial, donor mobile application, social media, temperature monitoring, and human resources. The business performance goals are defined and are congruent across the various business units. IT organization and desirable behaviors are documented in the ICT policies and procedures and were needed remedial actions are available through the code of conduct. The IT metrics are included within the NBSZ monitoring and evaluation system which use a four colored traffic lighting reporting system. It was noted that the IT accountabilities are undesirably tilted to the IT specialist only hence some ICT projects tend to have delayed deliverables. The IT governance mechanisms are supported with tools such as service level agreements and established communication approaches. Simple excel based solutions are used to track critical performance metrics such as on the interactive blood supply management status, which averages 122.8% (2018) based on a 5‐day projected stocking and supply levels. NBSZ need to properly document the return on investments on all these ICT initiatives, which is estimated (2016/2017) to be at 3.2% of annual savings.

Summary/Conclusions: Blood services in resource constrained settings can implement a properly structured IT governance and this will ensure maximum return on IT investments. The NBSZ approach will be shared and further developed in the AfSBT ITWG to support other blood services in improving their IT governance.

IMPACT OF THE INTRODUCTION OF AN INTEGRATED INPATIENT ELECTRONIC MEDICAL RECORD ON QUALITY OF PATHOLOGY SPECIMENS

A Haberfield, S Morgan, G Kelsey, J Box, H Schneider

Haematology/Blood Transfusion, Alfred Health, Melbourne, Australia

Background: In October 2018, an integrated electronic medical record (EMR) was implemented at an Australian metropolitan multi‐campus heath service using Cerner Millennium™, aiming to achieve HIMSS (Healthcare Information and Management Systems Society) level 6. Prior to implementation, large numbers of blood specimens were collected from patients unnecessarily and sent to Pathology without a test request attached (no blood test requested – NTR). These specimens required additional processing in the laboratory. Electronic specimen collection using Cerner Specimen Collect™ allowed streamlining of specimen processing by eliminating paper requests. As part of the new workflow, individual specimen labels are printed with the specified blood test and correct tube type. This helps prevent the practice of collecting additional specimens due to uncertainty of the collection requirements.

Aims:

• To quantify the expected reduction in NTR specimens following introduction of electronic specimen collection, and outline the benefits

• To determine the impact on collection errors and Wrong Blood In Tube (WBIT) events

Methods: Data was obtained directly from Cerner Millennium™ using a CCL (Cerner Command Language) query which is run monthly by Pathology IT staff. This data includes all specimens registered for the month with indication of rejected specimens, WBIT & NTR samples. ‘Rejected specimens’ includes incomplete specimen and/or request certification, unlabelled specimens/requests and mismatched specimens. Further information about WBIT events was collated from Riskman reports and staff interviews.

Results: Data from the 12months prior to EMR implementation was compared with 3months post. NTR numbers reduced from 4220/month to 2019/month (52% reduction), freeing up more storage space in fridges. Rejected specimens due to inadequate patient request labelling reduced from a mean of 27/mth to 6/mth. WBIT numbers have increased slightly: before having median 1 (range 0–2), after with median 2 (range 0–3). Although it was hoped that WBIT incidence may reduce with the new EMR, 4 of the post implementation 6 WBITs involved electronic specimen collection. Departure from planned protocols involving a lack of working printers, causing staff to print patient labels away from the patient's bedside, as well as multiple patient labels printing on individual printers appear to be a main cause of the EMR WBITs.

Summary/Conclusions: EMR implementation has led to a reduction in NTR, and rejected specimens due to inadequate request labelling, as well as increased storage space in laboratory refrigerators. Associated benefits include: • Decreased financial costs of the wasted equipment • Decreased staff time collecting and processing unusable specimens • Decreased environmental impact of manufacture and disposal of unused specimens • Decreased potential of iatrogenic anaemia Work in preventing the occurrence of further WBITs is ongoing, by ensuring that label printers are in working order, are in plentiful supply and easily accessible to staff; and also ensuring positive patient identification and blood collection by the patient's bedside remains a priority.

IMPLEMENTATION OF A PAPERLESS ELECTRONIC BLOOD BANK TEST REQUISITION SYSTEM IN A GENERAL AND ACUTE CARE HOSPITAL

JM Mustaffa¹, K Teo², S Tsai¹, P Heng¹, R Sagun¹, M Wong¹

¹Laboratory Medicine ²Khoo Teck Puat Hospital, Singapore, Singapore

Background: Khoo Teck Puat Hospital (KTPH) is a 761‐bed general and acute care hospital, opened in 2010, serving more than 800,000 people living in the northern sector of Singapore. The Blood Bank of KTPH Department of Laboratory Medicine provides specimen testing and blood transfusion services for KTPH as well as the neighbouring Yishun Community Hospital (YCH), one of the largest community hospitals in Singaporeproviding intermediate care for recuperating patients including rehabilitative services. The process of ordering transfusion‐related test requests in both hospitals is through printed forms.

Aims: In line with the hospital directive to move towardselectronic patient management, the KTPH Blood Bank intended to implement an electronic Type and Screen (e‐T&S) system. The goal of the project is to ensure zero patient identification errors and maintain full traceability and accountability for the blood collection process in all transfusion‐related testing. Another aim of the system is to reduce repeated venepunctures when specimens are rejected due to missing essential patient information on the printed forms by implementing mandatory fields in the e‐T&S form.

Methods: The e‐T&S was implemented in phases. Phase 1: an online version of the printed form was signed electronically by the ordering doctor and a witness within the electronic medical record system, Sunrise Clinical Manager (SCM) system with the doctor counter‐checkingby signing on the specimen label to ensure correct patient identification.Phase 2:the ordering doctor is not required to sign on the specimen label since fingerprint biometrics are required for the electronic sign‐in.Phase 3:elimination of the witness step for blood collection.Specimen collection and rejection data from 2016 to 2018 wasanalysed. Specimen rejection rate was presented as percentage of rejected specimens (mislabelled, unlabelled and clerical errors)over total specimencount for each month.

Results: Between January 2016 and March 2017, before the implementation of the e‐T&S phase 1, the average rejection rate for blood bank specimens was 0.16% and 1.14% for identification and clerical errors respectively. During phases 1 and 2 of implementation, rejection rate increased due to unfamiliarity to the new work processes. By February 2018, with the implementation of the final phase of the e‐T&S system the specimen rejection rate was0.38% and 0.12% for identification and clerical errors respectively. Rejected specimens were mostly from the few locations that had to use paper requisition due to workflow or infrastructure limitations.

Summary/Conclusions: The e‐T&S systemwas implemented successfully inKTPH. Full traceability and accountability of the blood collection process was maintained with the fully electronic system. Theadoption of electronic documentation has also reduced the number of preventable repeated venepunctures that were due to incomplete order information on the printed forms.Future developments in technology and full implementation of e‐T&S system in all hospital locations may make zero patient identification error achievable and ensure transfusion safety in all patients in the near future.

ELECTRONIC TRANSFUSION RECORD: A PILOT STUDY FOR SCWEB^® SYSTEM APPLICATION AT BEDSIDE TRANSFUSION

C Melli, D Camilot, C Dolfini, M Medeot, C Battaglia, P Zamaro, A Bertolutti, S Urban, S Pigani, S Gallo, V De Angelis

Transfusion Medicine, Udine University Hospital, Udine, Italy

Background: Blood component administration represents a critical phase due to the possible occurrence of errors during the different steps from the identification of the patient to the infusion of the product. Error occurrence can be reduced by the implementation of validated information systems. We tested the SCWeb^® System at the bedside in a transfusion outpatient clinic.

Aims: The aim of the study is to validate a system designed to assist and to control blood administration step by step using electronic devices to ensure traceability and documentation of the process

Methods: The SCWeb^® System is based on IT monitored checklists which guide the personnel to follow the procedure, according to best practices; the system must initially be activated by the operator which is recognized by an auto‐signing system based on Bluetooth Low Energy which avoids the operator having to identify himself/herself beforehand. Appropriate privacy protection is provided. Thereafter the system takes up the task to give instructions and to verify the adherence, by asking an active confirmation of the proper fulfillment of the activities; a continuous registration and documentation is made by the system. Standards and specifications for each step of the procedure have been configured on SCWeb^® System to track in detail operator and patient identification, presence of informed consent to transfusion, blood pressure, pulse and temperature recording, vein access, verification of the blood unit. An alarm has been set after 15min, to ensure the control of patient's conditions. For each step, an active confirmation of the action is required and nurse and doctor direct involvement must be actively confirmed on the device by both operators. The system has been tested at the bedside on 30 patients admitted to the outpatient clinic for 45 red cell concentrate transfusion; compliance of the personnel and organizational impact has been recorded.

Results: The System required a very short training: ease of SCWeb^® system allows its implementation without negative impact on organization of transfusion outpatient clinic and without difficulties by operators (nurses and doctors), who appreciated the help given by the IT check system. The registration of the electronic check list offered a reliable tool for the traceability of the transfusion procedure, also granting a paperless and timely available documentation of the entire process through a registration in electronic format of all the operator's action in every single phase of the transfusion process. When prescribed, confirmation of the checklist was only possible in the presence and with the active confirmation of two operators (doctor and nurse).

Summary/Conclusions: The SCWeb^® system is useful as a barrier against the mismatch of transfusion (preventive measure), as a traceability and documentation measure and as a tool for training of personnel in blood transfusion administration; it avoids paper registration during the transfusion process, due to the timely registration of the activities performed by operators recognized by the system thanks to the Bluetooth Low Energy auto‐signing device. The SCWeb^® system will be connected to the transfusion data management system, to monitor all the process from the arrival of the unit from the blood bank.

ANALYSIS AND DESIGN OF ELECTRONIC MEDICAL RECORD SYSTEM IN BLOOD TRANSFUSION SERVICES IN A GOVT. HOSPITAL

T Chandra¹, D Agarwal², M Agarwal³, R Agarwal⁴

¹Department of Transfusion Medicine, King George Medical University, Lucknow, Lucknow ²GSVM Medical College, Kanpur, Kanpur ³Era Medical College ⁴St Francis College, Lucknow, India

Background: he blood banks aims at reducing cost and increasing customer satisfaction by providing quality in service. The quality in service can be attained by streamlining the processes and restructuring the supply chain of the organization by implementing IT tools.

Aims: Aim is to understand the complex flow of information and processes within the supply chain of the blood bank. The requirement of such a study is a part of the integrated ERP modeling for the integrated functioning of a blood bank.

Methods: he approach used to understand and map the sequence of processes, and the work responsibilities of each process and the operational decisions involved at each step is process mapping and data flow diagrams for front end system modeling and analysis. The processes are mapped and represented in a schematic diagram. DFD (Data Flow Diagram) are constructed for representing the system. A context diagram is also constructed for understanding the entities interacting with the system. The EMR systems aim at replacing (or supporting) the paper based medical records. The whole model of the system is divided into two parts‐Front end and Back end. The front end design and analysis is done using EPC (Event‐Driven Process Chains), Resource Views, Data flow Diagram for Data view. Reporting was on Donor Selection, Finance and Collection of Blood Bag, Blood Collection Process, Component Preparation, Blood Testing and Blood Distribution

Results: Process mapping using Event Driven process Chain generated a whole view of the processes involved. The resource view gave an organizational structure and the personnel involved. The data view using context diagram and data flow diagram gives a flow of data and amount of data involved. This framework can be used for business process reengineering for the blood banks by conducting a time study and removing non value added activities. Data view helps analyze redundant data in each process. It also helps in staff training and orientation within the department.

Summary/Conclusions: A systematic overview presented in this paper facilitates in removal of non value added processes, duplication of data, bottlenecks, reduction of cycle time and thereby improving service quality in blood banks.

TRANSFUSION ASSOCIATED COSTS IN A PEDIATRIC INTENSIVE CARE UNIT

G Atakul¹, FY Ayhan²

¹Pediatric Intensive Care Unit ²Transfusion Center, Dr Behcet Uz Children's Hospital, Izmir, Turkey

Background: The transfusion of blood components, one of the most prevalent interventions in clinical practice is a major expenditure item in healthcare services which tend to increase in recent years.

Aims: It is intended to investigate the impact of transfusion associated costs to hospital costs in Pediatric Intensive Care Unit (PICU) patients.

Methods: During a year period (January 2017‐December 2017) 76 patients, 40 females and 36 males receiving transfusion with blood components along the stay in PICU were included in the study. Transfusion associated costs and total costs for healthcare services for children treated in PICU was collected by using Hospital Information System (HIS). Statistical analysis of data was performed by SPSS software (version 22.0, SPSS Inc., Chicago, IL, USA). Mann‐Whitney U test and Kruskal‐Wallis test was performed for comparison of independent categoric variables and numeric data; Chi‐square analysis was performed for comparison of two numeric variables and Spearman correlation analysis was performed for associations.

Results: The median age of patients was 12.0months (interquantile range‐IQR 26). The median length of stay was 16days (IQR 30). In total 400 blood components were transfused in which of 227 red blood cell concentrates, 114 apheresis platelet concentrates, 6 granulocyte concentrates, 51 fresh frozen plasmas, and 1 cryoprecipitate and 1 whole blood.

The ratios of transfusion associated expenditures to hospital costs were categorized in intervals of percentages as<5%, 5–10%, 11–15% and>15%. Most of the patients (63.2%) were ranked in the lowest interval. The medians for hospital cost and transfusion associated cost were 5478.76 euros (IQR=11280.02) and 130.57 euros(IQR=354.86), respectively.

A significant strong positive correlation between numbers of transfusions and hospitalization cost of PICU was detected (r: 0.674, P<0.01). While it was found a significant weak positive correlation between transfusion associated cost and hospital cost (r: 0.247, P=0.032) there was also a significant weak positive correlation between the age and transfusion associated cost (P=0.048, r: 0.227). A significant difference was found between the patients with and without hematological malignancies (P<0.01) for transfusion associated cost.

The reason why pediatric dosages are mostly prefer is that the hospital provides healthcare for only children and splitting of the blood components was common in the hospital. But unexpectedly a significant increase on the transfusion associated costs which is related to splitblood components was detected (P<0.05).

Summary/Conclusions: Studies on the economics of blood transfusion have been conducted mostly in patients who require chronic or multiple transfusions. PICUs, specialized facilities that provide care for patients with severe life‐threatening diseases are major departments often necessitate multiple transfusions.There are many variables to evaluate the impact of transfusion associated cost to hospital cost in PICU patients, but the major factors are underlying conditions, admitting diagnoses and transfusion strategies. Although there are unexpected data in our study demonstrated the increasing impact on transfusion associated cost originated from blood components split for pediatric usage no significant relationship was determined to explain this situation. Further studies on the economics of blood transfusions have to be carried out to clarify the variables of transfusion associated costs.

THE INFLUENCE OF STRATEGIES FOR LIMITING THE NUMBER OF PRETRANSFUSION TESTING TOWARDS COST EFFECTIVENESS OF ORDERING PACKED RED CELL IN DR. HASAN SADIKIN HOSPITAL BLOOD BANK, BANDUNG‐INDONESIA

LL Kosim^1,2, L Rokayah¹, R Suhartini¹

¹Clinical Pathology, Hasan Sadikin Hospital ²Clinical Pathology, Medical Faculty of Padjadjaran University, Bandung, Indonesia

Background: Approximately 55.7% of the transfused blood component is packed red cell (PRC). Over ordering of PRC unit is a common practice and excessive pretransfusion testing was being wasteful of resources and have adverse consequences on cost. High crossmatch to transfusion (C/T) ratio as quality indicator of blood bank implies that crossmatches were performed unnecessarily.

Aims: The aims of this study were to evaluate the cost effectiveness of strategies for limiting the number of pretransfusion testing of ordering PRC.

Methods: All PRC units who ordered from Dr. Hasan Sadikin Hospital from January 2016 to December 2018 were collected in this retrospective study. Number of ordering PRC unit, completed pretransfusion testing of ordering PRC units, and PRC units that were transfused were recorded. Restrictive pretransfusion testing strategies were done based on the hemoglobin level and diagnosis as transfusion indication criteria. Cost effectiveness was measured by multiplying the unit cost of pretransfusion testing and number of PRC unit.

Results: Out of total 177,370 ordered PRC unit, 166,910 (94.1%) were subjected to pretransfusion testing and 63.1% (105,260) of ordering PRC unit which are pretransfusion testing were transfused. This means that 5.9% (10,460) of ordering PRC unit were not subjected to pretransfusion test. This showed savings of 1,098,300,000 Rupiah. C/T ratio was 1.6 which demonstrate a good ordering pattern. However, 16.4% (27,451) of completed pretransfusion testing of ordering PRC unit were not transfused, leading to blood bank loss of 2,882,355,000 Rupiah.

Summary/Conclusions: Strategies for limiting the number of pretransfusion testing on the good C/T ratio was still associated with saving cost effective

IMPACT OF AN EDUCATIONAL INTERVENTION ON THE KNOWLEDGE AND AWARENESS OF NURSES OF A TERTIARY CARE TEACHING HOSPITAL REGARDING BLOOD TRANSFUSION SERVICES AND PRACTICES

S Talati¹, A Gupta¹, A Jain²

¹Hospital Administration ²Transfusion Medicine, Postgraduate Institute of Medical Education and Research, Chandigarh, India

Background: Blood is a precious resource for saving patient lives. The purpose of blood and blood component therapy is to provide suitable and safe blood products to achieve best clinical outcomes. Nurses have an important role in ensuring safe blood transfusion. It is crucial for nurses to have sufficient knowledge about blood donation and collection, storage, component preparation, possible adverse effects of blood transfusion and necessary management and care.

Aims: The aim of this study was to assess the impact of an educational intervention on knowledge and awareness of nurses regarding blood transfusion services and practices.

Methods: The baseline study to assess the knowledge and awareness regarding blood transfusion services and practices of the nurses posted at various areas of the hospital including wards, operation theatres and critical care areas was carried out at our institute hospital which is a tertiary care teaching centre. The nurses were then sensitized and educated regarding blood transfusion services and practices during their day to day activities by referring them to the blood transfusion guidelines of the institute. Subsequently, a self‐developed questionnaire which was used for the baseline assessment of knowledge and awareness of the nurses was again used to re‐assess them. A total of 19 questions were included in the questionnaire pertaining to: general awareness (two questions), blood donation (two questions), testing and blood component preparation related (two questions), storage of blood/blood components (two questions) and pre‐transfusion checks and bed side transfusion practices (eleven questions). Fifty nurses each were included for both the baseline as well as post‐sensitization assessment. For different category of questions, the correct response rates were compared with those obtained in the baseline study using Mann‐Whitney test. The entire study duration was spread over a period of three months (December, 2014 to February, 2015).

Results: The overall mean percentage of ‘correct’ responses for 19 questions in the baseline study was 61.75%, whereas post sensitization it was 77.21%. The mean percentage increase in general awareness related questions was 21.49%, 13.95% for storage of blood/blood components related questions, 17.37% for pre‐transfusion checks and bedside transfusion practices related questions, 21.33% for testing and blood component preparation related questions and 27.27% for blood donation related questions. The percentage increase in correct response was found to be statically significant for each of the five categories of questions. The overall mean percentage increase in correct response rate was also statistically significant (P<0.001).

Summary/Conclusions: This study revealed that after sensitization and educational intervention there was a significant improvement in the knowledge and awareness of nurses regarding the blood transfusion services and practices.

TRAINING ASSESSMENT AND COMPETENCY TOOL (TACT) – A GAP ANALYSIS OF THE TACT PROGRAMME VERSUS OVERSEAS PRE‐TRANSFUSION TESTING GUIDELINES AND PRACTICES – ADAPTING TACT FOR INTERNATIONAL APPLICATION

CL Whitham, J White, R Haggas

BTLP, UK NEQAS, Watford, United Kingdom

Background: TACT, introduced in the UK in 2014 to support managers, provides resource‐saving, continual, ‘real‐time’ monitoring of knowledge‐based competency of staff in transfusion laboratories. TACT is available online 24/7, complementing existing practical competency schemes and external quality assessment. Multiple variations on a standard pre‐transfusion testing scenario are generated using constrained randomisation; logic rules for automatic assessment of sample acceptance, ABO/D, antibody screen and identification (AS/ID), and component issue are based on BSH guidance. During 2018, TACT was offered internationally to transfusion laboratory managers to trial, and saw uptake in five countries. The core TACT programme, based upon UK guidelines, is under review for programming conversion, to be customisable for the international community.

Aims: To assess the feasibility of TACT programming conversion to meet the requirements of country‐specific pre‐transfusion testing guidelines, and to direct future programming in line with feedback from international users.

Methods: Guidelines from 3/5 international users were obtained and translated where necessary. These were compared against the core assessment elements of current TACT programming. International users were approached for their feedback on the current version of TACT, as it compared to their local policies and practices.

Results: The following criteria were cross‐referenced: specification of transfusion request forms, sample label acceptance criteria, reagents used for ABO/D and AS/ID, resolution of grouping anomalies, alloantibody confirmation/exclusion, and selection criteria of blood components for transfusion‐dependent patients and women of child‐bearing potential. Apparent differences included:‐

Australia:‐

‐Selection of red cells for patients with immune anti‐D.

Greece:‐

‐Inclusion of the name of the patient's father on the transfusion request.

Italy:‐

‐Testing of all new patients with an anti‐A,B reagent and two different monoclonal anti‐D reagents.

International users in the same three countries supplied feedback. This included suggestions for:‐ greater complexity of cases presented, provision of patient history, inclusion of follow‐on tests e.g. phenotyping and cells for antibody confirmation/exclusion, broader range of reaction strength grading, and official professional CPD credits. The following differences were noted:‐ nomenclature used, the format and content of the request form, use of English abbreviations of patient clinical details, and the availability, provision and specification of blood components.

Summary/Conclusions: This analysis has shown very few instances where the current TACT iteration differs from the guidelines reviewed, and that it is feasible to expand the use of TACT on a more international basis. The current iteration of TACT has been developed to represent an abbreviated scope of pre‐transfusion testing practices, which can be applied to laboratory practice outside of the UK without difficulty. Further work is required to enable international users to configure TACT such that the system represents all laboratory practice on an international basis.

IMPROVING NEONATAL AND PAEDIATRIC CLINICAL OUTCOMES THROUGH PATIENT BLOOD MANAGEMENT EDUCATION

D Peterson, S Ogley, L English, T Verrall, T Clark

Centre for Education and Training, BloodSafe eLearning Australia, North Adelaide, Australia

Background: BloodSafe eLearning Australia (BEA) is a government‐funded blood transfusion education program that provides courses in clinical transfusion practice and patient blood management.

In mid‐2018 BloodSafe eLearning Australia released a suite of neonatal and paediatric elearning courses based on the Australian Patient Blood Management Guidelines: Module 6 Neonatal and Paediatrics:

• ‐

  PBM for Neonates and Paediatrics (Introduction)

• ‐

  Neonatal: Preterm

• ‐

  Fetal and Neonatal Alloimmune Thrombocytopenia (FNAIT)

• ‐

  Paediatric: Haematology-Oncology

• ‐

  Paediatric: Surgical

• ‐

  Paediatric: Major Haemorrhage

• ‐

  Paediatric: Iron Deficiency Anaemia

Aims: These courses provide education to clinicians on patient blood management and safe transfusion in neonatal and paediatric settings in order to improve patient outcomes and increase awareness of the national patient blood management guidelines. This analysis aimed to investigate the uptake, practical use, and perceived value of the courses by learners.

Methods: A retrospective analysis of course completion statistics and course evaluation data.

Results: There have been 2,376 paediatric and neonatal courses completed from 6 March 2018 to 28 February 2019 with 89.6% of learners being nurses and/or midwives.

Analysis of course evaluation data (n=33) showed that these courses:

• ‐

  Provide knowledge (96.9%)

• ‐

  Improve patient safety and outcomes (84.9%)

• ‐

  Result in change to clinical practice (69.9%)

• ‐

  Are relevant to clinical practice (70.9%)

• ‐

  Are easy to use (67.7%)

• ‐

  Are readily accessible (58.1%).

Examples that learners provided of how they can apply this learning to their clinical practice include:

• ‐

  “[I am now] more aware of special requirements for neonatal blood transfusion”

• ‐

  “[I] feel more confident especially when talking with parents”

• ‐

  “[I will now be] checking the patient's blood results and will speak up for unnecessary blood sampling”

• ‐

  “[It's good that] when there is ambiguity in clinical practice [this is] very well shown by explanation from experts in the field”

• ‐

  “We don't do a lot of transfusions [and this is] a reminder that transfusions are not always the first answer to the baby's clinical picture”.

Summary/Conclusions: Analysis of course and user evaluation data demonstrates that these courses are being used by nurses and doctors working in the neonatal and paediatric settings and that they provide knowledge of PBM that can be applied to clinical practice, thereby contributing to improved patient care.

EVALUATION OF KNOWLEDGE AND TRANSFUSION PRACTICES AMONG NURSES ACCORDING TO THEIR SENIORITY OF EXERCISE

S Mahjoub¹, D Bahri², R Hidoussi², N Ben Romdhane¹

¹Hematology ²Hopital La rabta, Tunis, Tunisia

Background: Blood transfusion is a high‐risk clinical activity that must be mastered both theoretically and practically in order to guarantee the required result without any incident or complications. The mastery of transfusion knowledge among nurses represents a very important link in the transfusion chain.

The objective of this work is to compare the theoretical and practical knowledge of transfusion among two groups of nurses divided according to their seniority.

Aims: This is a cross‐sectional descriptive study conducted over a period of 1month [1^st April‐ 30^th April] 2017. We selected two groups of care staff: the 1^st group consists of 50 students at the end of their training at the higher Institute of Nursing Sciences. The 2^nd is made up of 50 nurses working in 5 university hospitals of Tunis, currently practicing blood transfusion. The evaluation's tool used was a questionnaire of 15 simple or multiple choice questions, 7 were related to theoretical knowledge of labile blood products and 8 to transfusion practice. Ten questions were considered “life‐threatening” if their answers were false. A comparative study was made between the two groups.

Methods: This is a cross‐sectional descriptive study conducted over a period of 1month [1^st April‐ 30^th April] 2017. We selected two groups of care staff: the 1^st group consists of 50 students at the end of their training at the higher Institute of Nursing Sciences. The 2^nd is made up of 50 nurses working in 5 university hospitals of Tunis, currently practicing blood transfusion. The evaluation's tool used was a questionnaire of 15 simple or multiple choice questions, 7 were related to theoretical knowledge of labile blood products and 8 to transfusion practice. Ten questions were considered “life‐threatening” if their answers were false. A comparative study was made between the two groups.

Results: The participation rate in the survey was 100%. The 2^nd group participants had an average seniority of 9years [0–30]. More than half of them (64%) had seniority of less than 10years. Only 12% had more than 20years of experience.

The rate of correct answers for all items combined was 44.4% for students versus 42.5% for practicing nurses. The theoretical knowledge part was more mastered in the 1^st group than that of practicing nurses (44.8% vs 33.4% of correct answers). On the other hand, the control of the transfusion act was better in 2^nd group (44% vs 50.5%).

The overall “dangerous” response rate was 47% for students and 41.7% for practicing nurses. False practical knowledge was more common in group 1 (59.5% vs. 41.5%).

Summary/Conclusions: The theoretical as well as the practical knowledge remains not well mastered by the care staff. Our study highlighted the best theoretical mastery for young students and practical for practicing nurses. This could be explained by the freshness of knowledge in the first group and the daily practice in the second group.

INCONGRUENCE IN WHOLE BLOOD DONOR SELECTION CRITERIA AS EVALUATED BY EXPERTS IN THE TRANSPOSE PROJECT IN COMPARISON WITH DIRECTIVE 2004/33/EC

S Fontana¹, J Castrén², S Van Walraven³, S Zahra⁴, A Chandrasekar⁵, E Veropalumbo⁶, K Seidel⁷, M Kvist⁸, C Mikkelsen⁹, H Ullum⁹, G Mori¹⁰, M Van Kraaij³

¹Interregional Blood Transfusion SRC, Bern, Switzerland ²Finnish Red Cross Blood Service, Helsinki, Finland ³Sanquin Bloodbank, Amsterdam, Netherlands ⁴Scottish National Blood Transfusion Service, Edinburgh ⁵NHS Blood and Transplant, Bristol, United Kingdom ⁶Italian National Blood Centre, Rome, Italy ⁷CSL Plasma GmbH, Marburg, Germany ⁸Clinical Immunology and Transfusion Medicine, Karolinska University Hospital, Stockholm, Sweden ⁹Department of Clinical Immunology, Copenhagen University Hospital, Copenhagen, Denmark ¹⁰Sanquin Bloodbank, Amsterdam, Netherlands

Background: The European Commission (EC) Directive 2004/33/EC on blood donor selection criteria is 15years old. In the meantime, knowledge on risks related to blood donor selection has progressed and challenged several obligatory rules. TRANSPOSE – TRANSfusion and transplantation: PrOtection and SElection of donors, is a European Commission co‐funded project with participation of more than 24 stakeholders from both not‐for‐profit and private organizations providing substances of Human Origin (SoHO). The project aims to provide evidence‐based donor selection criteria and guiding principles for risk assessment of threats to the safety of SoHO. As part of this work, an inventory of current blood donor selection criteria in Europe and an evaluation of the evidence behind current practice was performed by experts working on this project.

Aims: To identify the gap between the EC Directive 2004/33/EC on whole blood donor selection criteria and a current evaluation of the clinical relevance of the criteria based on scientific literature by a panel of European experts within the Transpose project.

Methods: In 2018, we performed an inventory of blood donor and transfusion recipient risks in participating European countries. Project members were asked to provide the existing donor selection criteria related to these risks and to carry out a risk‐based evaluation for each of them. The evaluation was based on the available scientific literature and on a risk assessment template based on the ABO Risk‐Based Decision‐Making Framework, developed by Transpose. All risks with divergent assessments within the panel were resolved through discussion; in all cases an expert consensus was established. Subsequently we compared the results with the content of the EC Directive 2004/33/EC for every risk, thereby identifying discrepancies and missing items in the Directive.

Results: The panel identified 64 risks considered to be significant, distributed between donors and recipients. For 35/64 (55%) of them the expert evaluation deviated from the content of the EC Directive, or the EC Directive provided no information about the decision making. In particular, a discrepancy was observed for 9/20 criteria concerning general health and medication, 12/22 for transfusion transmissible infections, 10/13 for high‐risk behaviour and travel, and 4/9 for other diseases.

Summary/Conclusions: Our results highlight a significant gap between the whole blood donor selection criteria stated in the EC Directive 2004/33/EC and the scientific evaluation performed by a panel of Transpose participating experts. This gap includes both new risks not addressed in the EC Directive and addressed risks that are however evaluated differently. This involves both blood donor and transfusion recipient safety, and various medical and epidemiological topics covering several aspects of the blood donation criteria.

We strongly recommend a change in the European legislation, allowing a procedure to guarantee that blood donor selection criteria are updated regularly within the framework of the European institutions, to keep aligned with scientific progress, epidemiology and changes in medical practice, in order to enable an updated risk‐ and evidence‐based framework for donor selection criteria. The risk‐assessment method elaborated in the TRANSPOSE Project is a valuable instrument for this purpose.

RISK MANAGEMENT OF BLOOD SERVICES: THE RESULTS OF A 10‐YEAR BRAZILIAN EXPERIENCE

CD Costa, J Junior, R Martins, H Sousa, U Junior

GSTCO, Anvisa, Brasilia, Brazil

Background: The Brazilian Health Regulatory Agency – Anvisa has developed the Method for Assessment of Potential Risk in Hemotherapy Services (MARPSH) which is based on the data collected during the inspections of blood services carried out by regulatory authorities. Using MARPSH any blood service can be classified in one of 5 possible potential risk categories: High, Medium‐High, Medium,Medium‐Low and Low Risk. Each category represents a different potential risk level, according to the proportion of compliance with the established regulatory requirements. MARPSH has been used since 2007, showing a trend of risk reduction on blood services evaluated all over the country.

Aims: This work aims to describe the utilization of MARPSH as a tool for an integrated risk management model. Also, it shows the main results obtained after 10years of data monitoring and coordination of regulatory actions and policies by Anvisa, targeting quality and safety of blood products.

Methods: The utilization of MARPSH follows a network risk management model since the inspections are carried out by decentralized organs in all 27 states and some municipalities. The inspectors fulfill a standardized inspection guide containing the regulatory requirements, where each item is associated with a level of risk, varying from I to III as the risk increases. At the end of the inspection, after a statistical calculation, the service is categorized. This classification gives an estimate of its quality profile, guiding the adoption of suitable measures for risk management by local authorities and services. These data are send to the states (if realized by municipalities) and to Anvisa that perform consolidation in a national level. Either states or Anvisa use data to coordinate risk management measures in a broader spectrum. Data are continuously monitored by Anvisa as part of its strategical panel of indicators. Anvisa follows up specially blood services in High and Medium‐High Risk with the aim of helping or complementing local authorities’ actions. Additionally, Anvisa periodically sends this information to the Brazilian Ministry of Health and local governmental organs from Brazilian National Blood System that also support actions to improve quality in their blood services networks.

Results: Since 2007, when the assessment covered 109 blood services, MARPSH reached 1218 blood services in 2017 (57% of the blood services registered) what corresponded to almost 100% of the inspection cover in this year. Over this period (2007–2017) it is possible to notice a dramatic decreasing in trend for proportion of blood services classified as High and Medium‐High Risk, varying from 26% to 9%.

Summary/Conclusions: MARPSH generates data necessary to the categorization of blood services into five levels of potential risk. As a result, the regulatory actions are applied by local organs with the purpose of reducing or eliminating risks involved in the production and use of blood components. Data have shown a significant risk reduction over 10years of MARPSH's utilization. Additionally, monitoring of this data at a national level has been permitting appropriate planning and prioritization of integrated strategies directed to risk management and strengthening of blood services in Brazil.

FAILURE MODE EFFECT ANALYSIS ON PATIENT SAFETY IN THE BLOOD TRANSFUSION CHAIN IN THE DEMOCRATIC REPUBLIC OF THE CONGO

NM Ndalingosu¹, A Heroes^2,3, I Van Cauwenberg⁴, J Jacobs^2,3, J Kabinda^5,6, P Vandekerckhove^7,8, O Lunguya^9,10

¹Quality assurance, National Blood Transfusion Center, Kinshasa, Congo, The Democratic Republic of the ²Department of Microbiology and Immunology, KU Leuven, Leuven ³Department of Clinical Sciences ⁴Department of Public Health, Institute of Tropical Medicine, Antwerp, Belgium ⁵Head Office, national blood transfusion center ⁶Science de la Santé, Université Pedagogique National, Kinshasa, Congo, The Democratic Republic of the ⁷Department of Public Health and Primary Care, KU Leuven, Leuven ⁸Red Cross‐Flanders, Mechelen, Belgium ⁹Institut National de Recherche Biomédicale ¹⁰Université de Kinshasa, Kinshasa, Kinshasa, Congo, The Democratic Republic of the

Background: Sub‐Saharan Africa has the highest need of blood transfusion in the world, mainly for childbearing women and children suffering from malaria. Meeting basic quality and operational requirements to provide patients with safe blood remains a challenge in these settings.

Aims: The aim was to identify and prioritize potential hazards for patients in blood bank practices in the Democratic Republic of the Congo (DRC). We focused on two subsets: (i) sensibilisation and selection of donors, and (ii) qualification and production of blood products, using Failure Mode Effect Analysis (FMEA).

Methods: Two risk analysis workshops were organized at the National Blood Transfusion Centre in Kinshasa, the Democratic Republic of the Congo. In both workshops, a multidisciplinary team was invited to represent different hospitals and profiles in transfusion management in DRC: quality coordinators (n=2), training coordinator (n=1), medical doctor for donor selection (n=2), hemovigilance officer (n=1), laboratory technicians performing donor sampling, blood qualification and production (n=7), biomedical scientists (n=3), microbiologist (n=1), clinical biologist (n=1), nurse (n=3). The principle of FMEA was applied, which implies identification of possible hazards (failures) in a process flow, followed by scoring each hazard according to their impact, probability, detection and feasibility. In the both workshops, participants were guided by an external facilitator who guaranteed common understanding of the methodology. Main focus of the risk analysis was potential harm to the transfused patient in the process of (i) sensibilisation and selection of donors, and (ii) qualification and production of blood products. All ideas were written on coloured cards and mapped on a chart according to their impact, probability, detection and feasibility score. Hazards were ranked according to their final risk score by multiplying these four scores.

Results: In the process of sensibilisation and selection of donors, the three hazards with the highest final score for impact, probability, detection and feasibility were: (i) a paid donor is recruited after sensibilisation by family members of the patient, (ii) donor selection staff approves a non‐eligible donor for blood donation because of personal/financial motivation, (iii) blood donors are not correctly informed about blood donation during sensibilisation by family members of patients.

In the flow of qualification of blood products, highest scores were made for: (i) no double check for validation and sorting of qualified blood products, (ii) stock‐out of reagents, (iii) no check for match between registered test result and tested blood tube.

Regarding production of blood products, the top three consisted of: (i) transport time between blood collection and processing is >8h, (ii) storage of qualified and quarantined blood products in the same fridge (some sites only), (iii) power cuts.

Summary/Conclusions: The risk analysis resulted in three prioritized hazards in the process of donor selection/sensibilisation and blood product qualification/production. Some are very specific to the sub‐Saharan African setting and have been described before (power cut, family and paid donors, stock rupture,…). An action plan needs to be put in place to reduce their final risk score. The risk analysis needs to be continued for the remaining blood transfusion flows.

BLUSTAR.NRW – A PROJECT FOR TYPING REFUGEES AND MIGRANTS AS POTENTIAL BLOOD AND STEM CELL DONORS IN NORTH RHINE‐WESTPHALIA

V Lenz¹, B Wagner¹, C Baumgart¹, L Kordelas², C Jiménez Klingberg², K Gebhardt², T Reimer³, T Zeiler³, J Fischer⁴, J Enczmann⁴, V Balz⁴, P Horn¹

¹Institute for Transfusion Medicine, University Hospital Essen, Essen ²Westdeutsche SpenderZentrale, Ratingen ³German Red Cross Blood Service West, Hagen, Breitscheid, Münster, Bad Salzuflen ⁴Institute for Transplantation Diagnostics and Cell Therapeutics, University Hospital Düsseldorf, Düsseldorf, Germany

Background: 19.3 million of Germany's population, so just under a quarter of residents, have a migration background. The majority of these has roots in regions where the population has a distribution pattern of blood group and HLA‐antigens that differs considerably from the predominant one in the German population. Sufficient supply of these individuals with red blood cell (RBC) and platelet concentrates (TC) will continue to be a major challenge in the future, as blood donors with compatible blood group antigens are dramatically underrepresented in the local donor pools.

Many migrants suffer from severe hematological disorders such as β‐thalassemia or sickle cell disease and will not only need compatible blood transfusions, but an allogeneic stem cell transplantation in the foreseeable future. As healthy family donors often are not available, at present suitable stem cell donors with a similar genetic background can only be found in international donor registries.

Aims: This project was initiated to recruit new donors with a migration background for blood donation and to increase the number of blood stem cell donors among this group.

Methods: Serological extended blood group phenotyping was performed by automated gel card technique (Fa. Grifols, Erytra) and included AB0, Rh (CcDEe), Kk, Fy(ab), Jk(ab), Lu(ab), M, N, S, s.

HLA typing for HLA‐A, ‐B, ‐C, ‐DR, ‐DQ, and ‐DP was performed by Next Generation Sequencing. Allele frequencies were analysed using Genepop Version 4.2; the rare and very rare alleles were defined according to the Allele Frequency Database (www.allelefrequencies.net)

RBC genotyping using Next Generation Sequencing is currently being established and will include additional antigens with the most frequent distribution pattern differences between migrant and resident populations according to literature.

Results: So far, more than 8800 blood donors with a migration background have been recruited for a blood donation in this project. Amongst this group, over 1000 blood donors from more than 20 non‐European countries enrolled as potential stem cell donors.

An initial evaluation of the data revealed a very similar distribution of blood groups compared to the current blood donor population in North Rhine‐Westphalia. Of 600 migrant donors, ten Fy(a‐b‐) donors were identified, which corresponds to a percentage of 1.6%.

Amongst 509 HLA‐typed potential stem cell donors, we found 28 (5.5%) with rare and very rare alleles.

Summary/Conclusions: Blood donors with rare blood group and HLA phenotypes (e.g. null types such as Fy(a‐b‐)), are in demand for adequate medical care of people with a migration background. The technological development of blood group determination by Next Generation Sequencing will significantly improve the supply for all blood transfusion recipients in Germany.

This project is funded by the European Development Fund 2014–2020 (ERDF) and the European Union.

TRANSFUSION PRACTICES IN SEVERELY INJURED TRAUMA PATIENTS IN AN EMERGENCY DEPARTMENT: EXPERIENCE AT LEVEL 1 TRAUMA CENTRE IN INDIA

R Chaurasia¹, A Krishna¹, A Subramanian², T Sinha³

¹Transfusion Medicine ²Laboratory Medicine ³Emergency Medicine, JPNATC, AIIMS, NEW DELHI, India

Background: Mortality due to uncontrolled haemorrhage following trauma is the most important cause of potentially preventable deaths. Trauma care systems in Low and Middle Income countries like India, are still in developing phase. Also, the role of blood component therapy in improving patient outcomes has been mostly derived from combat settings. Application of these protocols in an urban setup has not been well established and marked variation in practice exists. Hence this study aims to identify the key components of transfusion practices to optimize the transfusion protocol in trauma settings.

Aims: To study the current transfusion practices in severely injured trauma patients, admitted to the red area/resuscitation bay after initial triage in ED

Methods: This prospective observational study was conducted over a period of 1year starting from June 2017 to May 2018 at the Department of Transfusion medicine in collaboration with Emergency Department at JPNATC, AIIMS, New Delhi. The study included severely injured patients (ISS≥16) that were admitted within 24h to the red area/resuscitation bay after triage. Data collected included the demographics, injury, laboratory and transfusion details for these patients

Results: During the study period 885 patients (83.5% males) were enrolled. Mean ISS scores was 21.89 (IQR 16–25). Mean time to hospital admission after injury was 9:03 (IQR 3.38–13:48) hours. Mean time to first RBC transfusion following admission was 2:09 (IQR 0:27–2:45) hours. Approximately 49.3% (436) patients were in shock (SBP<90mm Hg &/or pulse rate>110/min). Whereas, 160 (18.1%) patients were coagulopathic (PT≥1.5 times of normal). During initial 24h of admission, these patients were transfused with 2929 (69.7%) RBC, 1986 (51.8%) FFP, 2327 (42.9%) RDP and 384 (81.5%) cryoprecipitate of total blood components utilized for these patients. Massive transfusion (defined as transfusion of≥5 units/4h) was given to 190 (21.4%) patients.

Summary/Conclusions: Significant quantity of blood components were required during initial resuscitation in severely injured patients. Pre‐hospital transfusion can significantly reduce the time to transfusion. Further studies are needed to assess utility of Pre hospital transfusion in severely injured patients.

THE MANAGEMENT OF PLATELET CONCENTRATE SUPPLIES FOR ALLOGENEIC STEM CELL RECIPIENTS IS A REAL CHALLENGE FOR A TRANSFUSION CENTER: THE EXPERIENCE OF THE GENEVA UNIVERSITY HOSPITAL

N Lambeng¹, A Altmeyer², S Huguet², C Chaduc Lemoine², S Waldvogel³

¹Department of Diagnostic ²Department of Medicine ³Department of Diagnostic and Department of Medicine, Geneva University Hospital, Geneva, Switzerland

Background: Allogenic stem cell transplantation recipients are known to be the main consumers of platelet concentrates (PC). The Geneva university hospital is one of the three allogeneic hematopoietic stem cell transplantation (HSCT) centers in Switzerland. Since the blood center is also part of the hospital, data of PC consumption are easily available. As needs rose steadily since several years, with an average increase of 9% per year, PC supply is a serious concern for our center.

Aims: In this study we tried to evaluate if any pre‐transplant indicator could help to forecast the number of PC needed after an allogeneic hematopoietic stem cell transplantation.

Methods: This observational retrospective study was conducted in Geneva hospital on 78 patients suffering from various inherited or acquired disorders of the hematopoietic system who were treated by HSCT in 2016. PC consumption was examined from January 2016 to December 2018. The five indicators were: gender, stem cell source (Bone Marrow (BM) vs Peripheral Blood Stem Cell (PBSC)), donor type (HLA matched (10–8/10) vs haploidentical), conditioning regimen (standard vs reduced intensity), and CMV serology of the recipient.

Results: Data for a total of 78 patients aged from 3 to 74years were analyzed; 48 (62%) were male and 30 (38%) female; 35 (45%) were CMV‐negative and 43 (55%) were CMV‐positive. Out of a total of 78 transplants, 14 (17.9%) were haploidentical and 64 (82.1%) HLA‐matched. According to the stem cell source, BM was transplanted in 22 cases (28.2%), and PBSC in 56 cases (71.8%). Two patients also received a CD34+stem cell boost.

Our analysis showed that, with a mean follow‐up of 652days, the number of PC transfused to our patients treated by HSCT ranged from 0 to 383 units, with an average of 37 and a median of 15, illustrating a high variability. The results indicated that gender, stem cell source (BM vs PBSC), conditioning regimen (standard vs reduced intensity), and CMV serology of the recipient do not have any statistical impact on platelet consumption. However, we observed a tendency of an increased need for platelet transfusion when patients were CMV positive. Our results also showed a statistically significant (P=0.034) higher number of PC transfused for patients treated with a haploidentical (89) versus HLA‐matched (26) transplant.

Summary/Conclusions: This study points out the high variability of platelet consumption after HSCT, which limits the forecast of platelet production needed to support allogeneic HSCT recipients. A larger cohort would be required to confirm a potentially higher platelet consumption in CMV positive patients, and to consolidate our results showing a higher PC consumption for patients treated with haploidentical transplant.

REDUCTION IN WASTE THROUGH THE CARDIAC SURGERY BLOOD COOLER INITIATIVE

K Obrien¹, M Mohammed¹, A Lerner², B Vidal¹, C Luffman¹, L Uhl¹

¹Pathology ²Anesthesiology, Beth Israel Deaconess Medical Center, Boston, MA, Boston, United States

Background: Historically at our institution, a minimum of four red blood cell (RBC) units were crossmatched for all cardiac surgery cases regardless of surgical case‐type or patient characteristics. Two RBC units were pa