ALBERT

Description: INTRODUCTION: The majority of patients with multiple myeloma undergoing autologous stem cell transplant require transfusion support during the engraftment period. A retrospective study was conducted to identify factors predicting need for blood transfusion in this population: METHODS: Charts were reviewed for all patients undergoing their first autologous stem cell transplantation for multiple myeloma (MM) at the Princess Margaret Hospital (PMH) from July 1, 1999 to Oct 31, 2002. In order to ensure all transfusion needs were tracked, patients who did not receive all follow-up care at PMH were excluded from analysis. Use of blood products during the 90-day period following stem cell infusion (Day 0–90) was determined by review of the hospital blood bank laboratory information system. An RBC transfusion was defined as a single unit, and a platelet transfusion as either one apheresis unit or 3–5 pooled whole-blood derived units. The following factors were evaluated as predictors for receipt of an RBC and/or platelet transfusion: patient age and sex, myeloma subtype (IgG vs. other) and stage (Salmon-Durie I/II vs. III); presence of hematologic comorbidity (active blood loss and being on therapeutic anticoagulation) in the post-transplant period; previous treatment with alkylating agent(s) 〉1 month and 〉 4 months; pre-transplant pelvic or lumbar radiation therapy; hemoglobin at transplant admission (Day -2), white blood cell count, platelet count, and creatinine; number of CD34 cells infused. Results were analyzed by both univariate and multivariate logistic regression analysis. RESULTS: Over the 40-month period of review, 280 transplants for MM were performed with 128 cases identified with complete transfusion and follow-up data available. The overall transfusion rate in the post-transplant period was 54.0 % for RBCs, 67.6% for platelets. By univariate analysis, significant predictors for RBC transfusion were female sex (OR 2.03, p =0.046), previous treatment with alkylating agent(s) 〉1 month (OR 3.29, p =0.013) or 〉 4 months (OR 4.83, p =0.018) and decreased admission hemoglobin (OR 0.90, p 4 months alkylator treatment remained as independent predictors. If hemoglobin was dichotomized as ≥110 vs.

Description: Introduction In robust hospital transfusion services, transfusion reaction reporting triggers a structured response to the assessment, diagnosis, and clinical management of the individual. At a population level, the feedback loop of hemovigilance permits the perception of signals applicable to donors, material production, patterns of use, infusion care, and recipient vulnerabilities. Transfusion reaction reporting therefore aims to improve the quality of patient care and safety in transfusion, which remains one of the most commonly performed procedures in medicine today. Large variations between passive/retrospective or active/prospective systems imply underreporting. Reasons for this may lie in unawareness of, nihilism on, or obstacles to this duty. In 2009, our center transitioned from a paper-based to an electronic reporting system (ERS) for suspected patient reaction events (PREs). This study sought to determine the impact of this change on PRE reporting rates. Methods This study was conducted in Toronto, Canada at the University Health Network, a 4-site, 767-bed, ternary care hospital with high transfusion activity (2014: 60,000 component and 30,000 derivative dispensations). In 05/2009, hardcopy mountsheets for transfusion labels were revised to provide space for recording corresponding vital signs, with instructions on PRE reporting. Medical director PRE review followed with event documentation in a transfusion laboratory database (recording imputability, reaction type, severity, and implicated product(s)). An Acute Transfusion Reaction policy was also developed to protocolize and further streamline the approach to various reactions, but was not implemented across all sites until 11/2009. At this time, electronic PRE reporting went live in the existing electronic medical record, with medical director review hereafter culminating in uploaded case conclusions. Technical tutorials on healthcare worker reporting spanned several months before implementation, without emphasizing the theory or evidence-based value of hemovigilance. The quantity and characteristics of reactions pre-/post-ERS implementation were compared. Results Prior to the ERS option (5/2009-11/2009), the reported PRE rate was 0.26/day. Subsequent to launch (11/2009-12/2009), the reported PRE rate was 0.66/day, representing a 2.54 fold increase (p

Description: Introduction: The extent to which febrile transfusion reactions (FTRs) are investigated is the extent to which bacterial contamination (BaCon) may be ascertained; FTR rates in turn vary with policies concerned with their recognition and approach. Microbiology and serology aim to rule out contamination or incompatibility, both potentially fatal. BaCon and acute hemolytic transfusion reactions (AHTR) followed transfusion related lung injury (TRALI) as the leading causes of transfusion-related death in the US in 2013 (FDA: 59 fatalities: 38% TRALI, 15% AHTR, 10% BaCon). Timely BaCon recognition enables interdiction/examination of sister products, while highlighting contamination points in the work sequence. The quality and quantity of hemovigilance data from a large hospital transfusion service were reviewed with respect to overall component utilization, adverse events, and patient/product microbiology, so as to gain a contemporary estimate of BaCon. Methods: The blood transfusion laboratory (BTL) of the tertiary care, 767-bed university hospital is supplied by Canadian Blood Services and managed by a team of technologists, a transfusion safety officer (TSO), and transfusion medicine specialists (TM MDs). Patient Reaction Events (PRE) reported to the BTL were logged over a 5 year period alongside components transfused (red cell units [pRBC], adult dose platelet concentrates [APC], frozen plasma [FP], and cryoprecipitate [crpt]). By policy, PRE are reported and formally investigated, with quarterly analyses. Roughly 3% of product recipients experience a PRE, and 40% are febrile in nature. Patient sampling is discouraged for "lower risk" fevers (asymptomatic Tmax 39C or major symptoms and/or vital sign disturbances) call for cultures of the patient and implicated product(s), as well as AHTR testing. TSO and TM MD review ensue to conclude product imputability, event severity, and final diagnosis. Definite BaCon (Def-BaCon) is defined as product and patient positive (+) for the same microbe, Probable BaCon (Prob-BaCon) as product (+) [but patient negative or untested], and Possible BaCon (Poss-BaCon) as patient (+) [but product negative or untested]. Poss-BaCon was re-classified to high-imputability (Hi-Imp Poss-BaCon) if case review failed to discover a more likely pre-existing source. Results: Between 1/1/2010 to 31/12/2014, 1,624 PRE occurred through 290,044 components dispensed (175,542 pRBC, 43,187 APC, 58,235 FP, 13,080 crpt). Patient cultures occurred in 617 (38%) of PRE, and product cultures occurred in 406 (25%) of PRE. BaCon rates varied significantly according to concluded certainty, with significant re-scaling of poss-BaCon after careful case review (Table).Table 1.rate (95% confidence interval):rate per culturerate per patient reaction event (PRE)rate per component dispensedDef-BaCon (4 cases)0.65% (0.26-1.6)0.25% (0.10-0.63)1.4 x 10^-5 (0.6-3.5) or 1 in 72,511Prob-BaCon (13 cases)3.2% (1.9-5.4) (products)0.80% (0.47-1.4)4.5 x 10^-5 (2.6-7.7) or 1 in 22,311Poss-BaCon (96 cases)15.6% (12.9-18.6)5.9% (4.9-7.2)3.3 x 10^-4 (2.7-4.0) or 1 in 3,021Hi-Imp Poss-BaCon (14 cases)2.3% (1.4-3.8)0.86% (0.52-1.4)4.8 x 10^-5 (2.9-8.1) or 1 in 20,717 Discussion/Conclusions: These data illustrate practical limits to deducing BaCon rates, despite robust hemovigilance. Def-BaCon was rare (1 in 72,511), while Prob-BaCon and Hi-Imp Poss-BaCon were more frequent at ~1 in 20,000. Current as-practiced tools in FTR/BaCon investigation are flawed at various levels. Underestimates stem from under-culturing and test sensitivity, and overestimates occur with incomplete case review for true sources of bacteremia, with Poss-BaCon as high as 1 in 3000. The MD Anderson Cancer Center (Ricci, et al 2014) reported on 999 reactions, with 738 (74%) in 642 central venous catheter (CVC) patients; 606 were cultured within a week of reaction, and 60 (9.9%) were bacteremic. Fevers were concluded to more likely represent the unmasking of CVC colonization rather than BaCon. Systematically incorporating (and adjusting for) CVC data may thus help to reduce inflationary poss-BaCon rates. On the other hand, more rigorous product testing (with biofilm studies) may scale BaCon rates upwards. Clinicolaboratory studies are needed to clarify the true relationship between febrile reactions, bacterial sources, and their significance. Disclosures No relevant conflicts of interest to declare.

Description: Background Passenger Lymphocyte Syndrome (PLS) is a rare complication of solid organ transplantation (SOT) and is marked by production of donor-derived antibodies towards host red blood cell (RBC) antigens. At Canada's largest SOT program, 〉400 transplants are conducted annually. The affiliated Transfusion Laboratory detects possible cases of PLS when previously seronegative hosts develop (delayed, non-transfusion-attributable) post-transplant antibodies (Ab) bearing unexpected, autoreactive (rather than alloreactive) RBC specificities. Donor attributability is established by discovery of the same specificity (on serologic lookback) and/or by the antithetical (seroconversion risk) phenotype. The incidence, duration, and severity of PLS remain unknown due to the infrequency of reported cases and the general absence of active surveillance. Method Between 1/10/2006 – 30/06/2013, PLS cases were discovered by the Blood Bank through pre-transfusion specimens. The specificities, phase (direct or indirect), correlative markers of hemolysis, and temporal profile of the Ab were subsequently followed according to patient and clinician interest. The severity of hemolysis (grade 0 to 5) at the time of nadir hemoglobin was judged according to the tiers defined by the Canadian Blood Services/IVIG Hemolysis Pharmacovigilance Group (CL#2011-34). Implicated organ and donor review (for relevant serology, and outcomes in any parallel organ distributions) were also attempted. Results Over 88 months, 2772 transplants occurred (992 kidney, 187 pancreas [with/after kidney in 113/74 respectively], 909 liver, 529 lung [437 double, 83 single, 9 +heart], 151 heart, 4 bowel). PLS was found in 14 or 0.5% (sex: 5F: 9M; mean age 56 years), with the implicated organ being lung or liver (9 [2%] and 5 [0.6%] respectively). All had a negative pre-transplant DAT with a pre-sensitized donor. PLS Ab specificities included ABO (8: 5A, 3B), RH (5 with ≥1 target [C:3, D:5, E:2, V:1]), and others (KIDD: Jka in 2; MNS: N in 1; unidentified in 1). Multiple Ab occurred in 5 (36%), usually within the same system (3 RH cases: C,D,E in 2; C,D,V in 1), but otherwise towards ABO + ≥1 target (B,Jka,N; B, unidentified). The time to the first detectable expression of an Ab (as signaled by a +DAT and/or screen) was 4-120 days (mean 24, median 15, n=14). Clinically significant hemolysis occurred in 11 (79%), with six grade 3 and five grade 4; the duration of hemolysis was 0-776 days (mean 148, median 38, n=11). The duration of Ab detectability (or the point of “last positivity”) was 12-851 days (mean 196, median 78, n=10), while the point of disappearance (“subsequent sustained negativity”) occurred between day 12-1288 (mean 298, median 45, n=10). Lookbacks were feasible with 5 implicated donors, and in only 1 case (of a contralateral lung's transplantation) was PLS detected in another recipient. PLS did not account directly for any deaths in the cohort, although hemolytic morbidity required treatment in 2 cases. Transfusion support with antigen-negative blood was provided to 11 patients for anemia that may have been due to hemolysis and/or other causes while the Ab was still evident. In one case, severe hemolysis provoked rate-uncontrolled atrial fibrillation with a supply and demand ischemia which culminated in irreversible congestive heart failure and cardiac death 2 months after PLS resolution. Discussion/Conclusions This is the largest series of PLS reported to date. Clinician-driven submissions of blood for transfusion compatibility testing, and the discovery of discrepant serologies, represented an unforced but prospective means by which to identify PLS in roughly 1 in 200 transplants. Several cases transcended the limits of previously described temporal profiles (with the most delayed onset at 120 days, and the most prolonged duration at 2.3 years). The duration of hemolysis was consistently shorter than the duration of seropositivity, and was rarely life-threatening, while the most severe grades associated with RH rather than ABO or other targets. The surveillance achieved in this cohort revealed a longer-than-previously-appreciated duration of Ab productivity, and raised the question of whether or not such prolongation is a function of observation bias or the current-day armamentarium of anti-rejection therapeutics, which may be more permissive to donor-derived chimerism. Disclosures: No relevant conflicts of interest to declare.

Description: Introduction: Serologic testing of post-transfusion reaction specimens aims to ascertain potentially accountable immune hemolytic incompatibility. With the exception of low-risk fevers or uncomplicated allergic reactions (ie- reactions with a likelihood of incompatibility that is deemed too low to justify testing), all transfusion reactions undergo serologic investigation as a matter of local institutional policy. However, compliance with guidelines, and the yields of testing according to reaction types, remain unknown. These measures may indicate the quality of applied practice, and provide evidence for maintaining (or changing) investigation algorithms. Study Design and Methods: Interrogation of two hemovigilance databases identified all possible-to-definite transfusion reactions over a 4-year period (2013-2016) at four academic hospitals (with 1493 adult-care beds). The performance and outcome of reaction-oriented serology were assessed by site, year, reaction type, implicated product, and patient location. Serologic testing of transfusion reaction samples entailed the performance (and pre-transfusion comparison) of grouping (ABO, RHD type) and red cell antibody screening (indirect antiglobulin testing [IAT]), with a post-transfusion Coomb's/direct antiglobulin test [DAT]. Appropriate reflex tests (elutions, panel investigations, or IAT re-crossmatching) proceeded from pertinent positives. Allergic reactions were "complicated" if significant vital sign changes occurred, while fevers were "high-risk" (HRF) if symptomatic or if the Tmax rose to ≥39⁰C. Cardiorespiratory reactions (CRR) involved symptomatic and/or objective disturbances in heart/lung function, while unclassifiable presentations (changes in sensorium or non-precordial pain) were placed in an atypical/"other" category for analysis. Results: Sites received 338-367 reaction referrals per year. By referral proportions (and with overlaps), fevers accounted for 47% of events, allergic disturbances for 40%, and CRR events for 37%, with unclassifiable reactions in 7% (Table). Serologic examination occurred in 773 (55%) of 1412 referrals (of which 1346 were deemed to be transfusion-attributable disturbances, among 1119 unique recipients). The majority of cases (1153 or 82%) were compliant with guidelines. Similar proportions deviated to over-testing (85/550 [15%]) as to under-testing (174/862 [20%]). Overall, 34 (4.4%) of 773 cases yielded a new finding, with 6 (0.8%) reflecting (new or recrudescent) host-derived anti-erythrocyte antibodies, for a number-needed-to-test (NNT) of 129. Serologic yields occurred in all categories where testing was mandated, with most yields (62% or 26/42) owing to HRF and CRR events. Whereas these were often non-ABO (minor antigen-targeting) antibodies (76% or 25/33) and followed reactions to red blood cell transfusions (RBC), the yields from complicated allergic reactions and "other" reactions were entirely due to passively acquired isoagglutinins (11/11). The former non-ABO antibodies were revealed by IAT (with no additional gains from elution studies), while ABO isoagglutinins were revealed by DAT alone (and type-specific eluates), and followed plasma (antibody-) containing products. IVIG-associated reactions exhibited the highest serologic yields (in 48% of cases, or in 70% of non-O type recipients), with 60% experiencing some degree of hemolysis. Conclusion: A fifth of reactions were either over-tested or under-tested. Analysis of the performance and contextual yields of serologic tests revealed that certain products and presentations merit greater attention while others merit less. The IAT is a greater priority than the DAT in HRF and CRR following RBC, while the DAT alone is informative in isoagglutinin-risk cases, irrespective of presentation type, and with predictable eluate specificities. A re-evaluation of traditional serologic testing reflexes may reduce costs and allow re-investment in other more informative reaction-specific assays. Table Table. Disclosures No relevant conflicts of interest to declare.

Description: Introduction: In patients undergoing ABO-incompatible stem cell transplants, recipient isoagglutinins with specificity for donor antigens are expected to slowly become undetectable during the post-transplant period, presumably secondary to the loss of recipient plasma cells via either conditioning chemotherapy or graft-versus-host-disease. Conversely, the persistence or return of these anti-donor isoagglutinins may induce hemolytic anemia or pure red cell aplasia, and may even signify disease relapse. In order to define the expected time to clearance of these donor isoagglutinins at our institution, a retrospective study was performed Methods: The transfusion records of all major-side or bidirectionally ABO-incompatible stem cell transplants performed at an academic cancer hospital between May 2009 and July 2018 were reviewed. Patients without detectable anti-donor isoagglutinins at time of transplant were excluded. Isoagglutinin clearance was defined as non-reactivity of patient plasma against donor-type reverse grouping cells occurring during the post-transplant period; relapse was defined as a 2+ or stronger reaction after one or more non-reactive results Results: A total of 280 patients were included in analysis, of which 44 were native group A, 42 group B and 194 group O. Average age was 52 (±13) years and 130 (46%) were female. Compatibility samples were sent to the UHN blood bank during the post-transplant period for a median of 156 days (IQR 74-370 days). During this time period, donor-targeting/host-derived isoagglutinin clearance was documented in 192 (69%) patients at a median of 48 days (IQR 23-116.5) post-transplant. Of these, 52 (30%) experienced a subsequent return of reactivity at a median 98 days post-transplant (IQR 54-216). Amongst the 128 patients with at least 28 days of post-transplant monitoring and no return of previously cleared isoagglutinins, the median time to clearance was 60 days (IQR 26.5 - 127.5 days). There was no statistically significant difference in time to isoagglutinin clearance between those with and without a subsequent return of reactivity, and time to durable clearance did not correlate with patient age, sex or native ABO group. Conclusions: Amongst patients undergoing major-side or bidirectionally ABO incompatible stem cell transplants, the majority will have cleared their donor-specific isoagglutinins within 120 days of stem cell infusion, although this does not guarantee against later return of reactivity. Thus, the persistence of donor-specific isoagglutinins beyond 120 days post-transplant, or the subsequent return of previously cleared isoagglutinins, may be considered abnormal results which warrant notification of the clinical team by the hospital transfusion service. Disclosures No relevant conflicts of interest to declare.

Description: Rate of Sickle Hemoglobin Recovery in Sickle Cell Disease Patients Undergoing Red Blood Cell (RBC) Exchange Transfusion is Associated with Age of Patients and Number of RBC Units Transfused Introduction: Automated and manual red blood cell exchange (RBCX) transfusions are useful in the primary and secondary prevention of sickle cell disease (SCD) complications (Ware et al., 2012). The ability to consistently maintain sickle hemoglobin (HbS) below target (30% or 50% depending on indication) is quite variable (Kuo et al., 2012). With emerging indications such as silent cerebral infarction, it is imperative that effective means of chronic transfusion to maintain appropriate hematological and clinical targets be identified. We hypothesize the rate of HbS recovery is dependent on the individual's hemolytic and erythropoietic rate. The purpose of this study is to evaluate the effect of the rate of erythropoiesis and hemolysis on HbS recovery in SCD patients undergoing RBCX. Methods: Fifteen (15) patients were prospectively recruited from the adult SCD transfusion program (9 automated, 6 partial manual), from December 2018 to July 2019, and followed through one exchange cycle (4 weeks). Automated and partial manual exchange transfusion protocols have been previously described elsewhere (Canadian Haemoglobinopathy Association Consensus Statement on the Care of Patients with Sickle Cell Disease in Canada, Version 2.0, Ottawa; 2015). Exclusion criteria included active hydroxyurea or erythropoietic stimulating agents use, reported ill health in the preceding 4 weeks, co-morbid hemolytic condition or non-HbSS genotype. Hemoglobin, hematocrit, HbS, lactate dehydrogenase (LDH), reticulocyte count, indirect bilirubin, and serum erythropoietin level were determined for each patient: pre- and post- first exchange, weekly for 3 weeks and pre- second exchange (the 4th week). Descriptive variables were either expressed as means ± SD or median (IQR), based on normality, while linear regression was performed for continuous variables. Co-variates were included in multivariable analysis if P 〈 0.10. Multivariable linear regression was conducted to examine the potential association between the change in HbS over one RBCX cycle and age of patients, pre-RBCX hematocrit, LDH, and number of RBC units transfused. Results: We identified 36 eligible patients from the Program database, after which 15 consented to participate in the study. Mean age was 32.9 ± 12.3 years, consisting of 7 males and 8 females. There was an association between the rate of change in HbS and age of patients (p=0.035), pre-RBCX hematocrit (p=0.030) and number of transfused RBC units (p=0.030). LDH showed a trend towards reduced rate of change in HbS (p=0.069). Rate of change in HbS was not associated with automated vs. partial RBCX (Figure), female vs. male patients, pre-RBCX HbS, erythropoietin, indirect bilirubin, reticulocyte and age of transfused RBCs. Age of patients (p

Description: Background Spontaneous loss of RHD expression is unusual and may present on routine RHD typing with new “mixed field” (separate positive and negative subpopulation) reactions. Its detection has practical and mechanistic implications for underlying pathology, and such events have been associated with the development or progression of malignant disorders. Case A 49-year-old gentleman with Acute Myelogenous Leukemia (AML) underwent standard induction chemotherapy with cytarabine and daunorubicin, achieving a remission. He moved from Saudi Arabia to Canada and presented to our center with an infected central venous catheter, whereupon a repeat marrow was performed, confirming remission (myeloblasts

Description: Abstract 4272 Introduction: Demand for blood components in Canada is a significant and growing contributor to health care spending, but the drivers of this demand remain poorly characterized. While most institutions do not routinely monitor the transfusion requirements of individual diagnoses and procedures, an institution's overall demand for blood components and the types of clinical activities it provides are both well documented. Development of an institution-level inferential model may allow for more accurate demand forecasting for blood components. Methods: Hospital shipments of all blood components from Canadian Blood Services (red blood cells, platelets, plasma and cryoprecipitate) in the province of Ontario during the period 2006–2009 were merged with institution-level administrative data containing hospital characteristics and selected diagnostic and procedural codes as obtained from inpatient and ambulatory clinical databases maintained by the Canadian Institute for Health Information. Simple linear correlation and several nested multivariable linear regression models were fitted and compared. Results: From 2006–2009, our sample included 137 healthcare facilities, representing approximately 1.4 million units of RBCs, 400 000 units of plasma, 166 000 doses of platelets (defined as either one pool of whole blood-derived platelets or a single apheresis platelet), and 20 000 units of unpooled cryoprecipitate. Institutional demand for these blood components correlated with both hospital-level characteristics and diagnostic and procedural activity data. Institutional demand for red blood cells correlated most strongly with visits for cirrhosis and chemotherapy. Demand for plasma correlated most strongly with visits for cirrhosis and cardiac surgery procedures. Demand for platelets correlated most strongly with bone marrow and solid organ transplant procedures. Demand for cryoprecipitate correlated most strongly with visits for valvular heart disease and cardiac procedures. The strength of these correlations generally remained stable over the four years of analysis, during which time demand for red blood cells and platelets increased, demand for plasma decreased, and demand for cryoprecipitate remained stable. Conclusions: In the largest province of Canada, institutional demand for blood components correlate strongly with disease burden measured by diagnostic and procedural codes contained within administrative data. Disclosures: Pendergrast: Canadian Blood Services: Research Funding. Zagorski:Canadian Blood Services: Research Funding.