VIEWS: 71 PAGES: 13 POSTED ON: 4/10/2011
TRANSFUSION PRACTICE Transfusion efﬁcacy of ABO major-mismatched platelets (PLTs) in children is inferior to that of ABO-identical PLTs Friedgard Julmy, Roland A. Ammann, Behrouz Mansouri Taleghani, Stefano Fontana, Andreas Hirt, and Kurt Leibundgut I n everyday transfusion practice, inventory manage- BACKGROUND: ABO major compatibility is essential ment pressures associated with the limited 5-day in transfusions of red blood cells but is not requisite in storage period for platelets (PLTs) result in the pro- PLT transfusions. In adults there is some evidence that vision of many ABO-mismatched PLT transfusions. transfusion efficacy of ABO blood group–identical plate- To minimize PLT outdating, it is standard practice in lets (PLTs) is superior to major-mismatched PLTs. many transfusion services to administer the oldest PLTs However, in children this question has not been investi- ﬁrst. These include transfusion of PLTs that are ABO- gated for more than 30 years. unmatched. While for red blood cell (RBC) transfusions, STUDY DESIGN AND METHODS: In a prospective ABO major compatibility is mandatory, both AABB and study, the efficacy (based on the 1-hour percentage of the British Committee for Standards in Haematology PLT recovery [PPR1hr]) of 400 eligible ABO blood suggest, but do not mandate, transfusion of ABO-matched group–identical or out-of-group apheresis PLT concen- PLTs.1,2 trates (APCs), transfused mainly prophylactically to 50 PLTs express ABH blood group antigens on their children with hematologic malignancies, solid tumors, or surface, some of them at very high levels.3-6 After ABO aplastic anemia was investigated. The primary objective major-mismatched transfusions, transfused PLTs may be was to compare PPR1hr between ABO-identical and directly affected by the recipient’s isohemagglutinins, the major-mismatched transfusions. naturally occurring ABO antibodies, leading to premature RESULTS: After ABO major-mismatched transfusions, destruction of the transfused incompatible PLTs. Further- PPR1hr was signiﬁcantly lower than after ABO blood more, transfusion of ABO minor-incompatible PLTs stored group–identical transfusions (median 21% vs. 32%; in plasma containing anti-A and/or anti-B antibodies can p = 0.034). Multivariate analysis showed major- lead to a hemolytic transfusion reaction in the recipient.7-9 mismatched transfusions to be signiﬁcantly more often unsuccessful than identical transfusions (odds ratio [OR], 3.97; 95% conﬁdence interval [CI], 1.52-10.39; p = 0.005). Using ﬂow cytometry and ﬂuorescent ABBREVIATIONS: APC(s) = apheresis platelet concentrate(s); microscopy, it could be demonstrated that PLTs of BSA = body surface area; BW = body weight; CCI1hr = 1-hour subgroup A1, signiﬁcantly expressing A antigen on their correct count increment; LME = linear mixed-effect (modeling); surface, were rapidly cleared from the circulation of PPR1hr = 1-hour percentage of PLT recovery. group O or B recipients. In contrast, major-mismatched From the Department of Pediatrics and the Department of transfusions of A2 PLTs, expressing no detectable A Hematology, University of Berne, and the Blood Transfusion antigen, were as successful as identical transfusions Service of the Swiss Red Cross, Berne, Switzerland. (OR, 1.13; 95% CI, 0.16-7.88; p = 0.90). Address reprint requests to: Kurt Leibundgut, MD, Division CONCLUSION: These data clearly indicate that in chil- of Pediatric Hematology/Oncology, University Children’s Hospi- dren ABO major-mismatched PLT transfusions result in tal, CH-3010 Berne, Switzerland; e-mail: kurt.leibundgut@ inferior transfusion efficacy, with the only exception of insel.ch. group A2 PLTs. ABO minor-mismatched PLTs showed This work was supported by the Swiss Cancer League comparable efficacy to identical transfusions. (Oncosuisse, Berne, Switzerland; Grants OCS 01470-02-2004 and KLS-01828-02-2006). Received for publication May 26, 2008; revision received July 14, 2008; and accepted July 17, 2008. doi: 10.1111/j.1537-2995.2008.01914.x TRANSFUSION 2009;49:21-33. Volume 49, January 2009 TRANSFUSION 21 JULMY ET AL. Already in 1965, Aster10 demonstrated that ABO ences in transfusion efﬁcacy, measured by the 1-hour per- incompatibility had a negative effect on the recovery of centage of PLT recovery (PPR1hr), between ABO-identical transfused PLTs in adults. Since then, this ﬁnding has been and major-mismatched PLT transfusions; that is, from conﬁrmed by several studies. Lee and Schiffer11 found group A donors to group O or B recipients, from group B that ABO compatibility can affect the results of pooled donors to group O or A recipients, and from group AB random-donor PLT transfusions and that patients who donors to group A, B, or O recipients, respectively. Major- experience poor increments from ABO-mismatched PLTs mismatched and bidirectionally mismatched transfusions may beneﬁt from ABO-compatible PLTs. Heal and col- (i.e., both major- and minor-mismatched) were catego- leagues12 reported that compared with ABO-unmatched rized together. Secondary objectives were to compare the PLTs, transfusion of ABO-identical PLTs is associated with efﬁcacy of ABO-identical PLT transfusions with major- a signiﬁcantly higher corrected count increment (CCI) mismatched transfusions of group A subtypes A1 and A2 as and with a twofold reduction of the risk to develop PLT well as with minor-mismatched transfusions. Further- refractoriness. While in these studies, pooled PLT products more, side effects of PLT transfusions and data on the were used and patients were randomly assigned to receive production process of APCs were recorded prospectively. either ABO-matched or ABO-unmatched transfusions, The latter included information on the apheresis proce- Jimenez and coworkers13 compared transfusion efﬁciency dure, storage medium, storage time, volume, and gamma of split apheresis products from the same donor in paired irradiation. recipients. If one recipient was ABO-identical and the other was ABO-unmatched, mismatched transfusions yielded one-third of the recovery of ABO-identical trans- Patients fusions. For the TRAP trial (Trial to Reduce Alloimmuni- Between January 2004 and June 2006, all consecutive zation to Platelets), Slichter and coworkers14 reported patients with hematologic malignancies, solid tumors, or signiﬁcantly higher PLT increments after transfusion of aplastic anemia requiring at least one PLT transfusion at ABO-compatible PLTs. In lymphoma patients undergoing the Division of Pediatric Hematology/Oncology of the autologous marrow transplantation, leukoreduced ABO- University Hospital Berne (Berne, Switzerland) were eli- identical PLT transfusions were associated with reduced gible for this trial. Exclusion criteria were refusal of study morbidity.15 In cardiac surgery, ABO-mismatched PLT participation, fever of 38.5°C or greater before transfu- transfusions may be associated with a poor outcome.16 sion, clinically enlarged spleen, thrombosis, adverse All of these studies were performed in adult patients. events requiring interruption of transfusion, and hemor- In pediatric patients it has been more than 30 years since rhage of WHO Grade 3 or greater. Before enrollment, transfusion efﬁcacy of ABO-mismatched PLTs has been written informed consent was obtained from the parents investigated. van Eys and coworkers17 reviewed the effects and (as far as possible according to age) from the patient. of 393 transfusions of pooled PLT products in pediatric The study protocol was approved by the Government cancer patients between 1973 and 1974. The authors con- Ethics Committee of Berne, Switzerland. The study was cluded that ABO matching had no effect on the success of conducted in accordance with the principles of good clini- a transfusion. Since this early study, production of PLT cal practice. concentrates has considerably improved. Within the past decade, in many centers pooled PLT concentrates from random units of whole blood have been replaced by Preparation and delivery of PLTs single-donor apheresis PLT concentrates (APCs).18,19 PLTs were obtained from healthy volunteer donors, either Modern apheresis equipment typically achieves high PLT by single- or by double-needle thrombocytapheresis using yields with very low RBC and white blood cell contamina- a cell separator (Amicus Crescendo, n = 105, Software tion.20,21 The present prospective study was designed to Version 2.51, Baxter, Deerﬁeld, IL). Further APCs were investigate the impact of providing either ABO-identical produced by single-needle procedures using one of two or out-of-group APCs to children with thrombocytopenia different cell separators (Trima Accel, n = 277, Software suffering from hematologic malignancies, solid tumors, or Version 5.0, Gambro BCT, Lakewood, CO; or COBE aplastic anemia. Spectra, n = 18, Software Version 7.0, Gambro BCT). Aph- eresis time was limited to 100 minutes. Depending on the PATIENTS AND METHODS yield, the apheresis product was split into units containing at least 2.0 ¥ 1011 PLTs, corresponding to the minimum Study design and objectives content required by the standards of the Swiss Red Cross In a prospective trial in children with thrombocytopenia, at the time. APCs produced with the Amicus device were we evaluated the inﬂuence of ABO blood group matching suspended in 65 percent additive solution (AS; T-Sol, PAS between PLT donor and recipient on posttransfusion PLT II, Baxter) and 35 percent donor’s plasma according to the increment. The primary objective was to detect differ- manufacturer’s instructions, whereas PLTs derived from 22 TRANSFUSION Volume 49, January 2009 TRANSFUSION EFFICACY OF MISMATCHED PLTs the Trima and Spectra devices were suspended in 100 dose [1011/L]. Blood volume was determined according to percent donor’s plasma. PLTs were stored at 22 2°C for a Linderkamp and coworkers.24 Although determination of maximum of 5 days on a ﬂat-bed shaker with constant the sample size by power analysis was based on the CCI1hr, agitation until delivery for transfusion. Older products the ﬁnal analysis was based on PPR1hr rather than on CCI1hr were delivered ﬁrst, in line with the policy of our division because of the inadequacy of the CCI1hr in infants and of transfusion medicine to avoid expiration of APC prod- small children; ergo only PPR1hr is shown throughout. ucts. ABO-identical products were provided ﬁrst; other- wise minor- or major-mismatched PLTs were delivered at the discretion of the staff. Wherever applicable no D+ PLTs Determination of PLT count were given to D– female recipients. Group O PLT donors PLT count in blood samples of patients was measured with were screened for the presence of anti-A and anti-B a PLT analyzer (Cell Dyn 3500R, Abbott Laboratories, hemolysins. If a titer of more than 4 was present, the Abbott Park, IL). In case of a PLT count of less than 30 ¥ 109 product was labeled accordingly and released for group O per L or an abnormal histogram pattern, manual counting recipients only. When required, gamma irradiation of with a modiﬁed Neubauer counting chamber was APCs was performed with a cesium irradiator (Gammacell performed. 3000 Elan, MDS Nordion, Ottawa, Canada) at doses of In each APC, after adhering to the preanalytic recom- 25 Gy. mendations of the apheresis device manufacturers, PLT count was determined with a hematology system (Advia PLT transfusions 120, Bayer, Leverkusen, Germany). Most patients received prophylactic PLT transfusions, with only a few receiving PLT transfusions for therapeutic purposes. In children without purpura, excessive pete- Determination of A antigen expression and A1/A2 chiae, widespread hematomas, epistaxis, mucosal bleed- subgroup typing ing, gastrointestinal bleeding, or genitourinary bleeding, For major-mismatched transfusions from group A donors, no PLTs were transfused, even when the PLT count was A antigen expression on PLT surfaces was measured by fewer than 10 ¥ 109 per L. As shown in Table 2, in deﬁned ﬂow cytometry both in the APC and in the recipient before clinical settings a PLT count of fewer than 10 ¥ 109 per L and 1 hour after transfusion. For the ﬁrst, we used a triggered prophylactic transfusion. Bleeding episodes due sample of the concentrate remaining in the tube of the to thrombocytopenia were managed with PLTs until con- transfusion set, and for the latter, PLT-rich plasma was trolled. In all other situations, infants received 10 to 15 mL obtained by centrifugation of blood samples. PLTs were per kg body weight (BW) of the APC product. Children diluted at a working concentration of 2 ¥ 107 PLTs per mL less than 2 years old received half an APC product, and with 20 mmol per L phosphate-buffered saline (PBS), patients aged 2 years or more received a full APC product, ﬁxed in 1 percent paraformaldehyde (ratio 1:1), and incu- independent of their BW. In line with our routine practice, bated at room temperature for 30 minutes. The following we recorded pulse rate and blood pressure before starting monoclonal antibodies (MoAbs) were used: PLT gate was the transfusion as well as every 3 minutes during transfu- identiﬁed by staining PLTs with MoAb anti-CD61, directed sion. Body temperature was measured before and after against glycoprotein IIIa (Becton Dickinson, Franklin transfusion. Patients were monitored for adverse events Lakes, NJ). A antigen expression on PLT surface was until 1 hour after the end of transfusion. assessed using anti-A reagent BRIC-145 (International Blood Group Reference Laboratory, Bristol, UK). Approxi- mately 106 PLTs were stained with the detecting antibodies Efficacy measures: PPR1hr and 1-hour CCI (CCI1hr) anti-CD61 and BRIC-145. At a working concentration of Transfusion efﬁcacy was determined by two surrogate 10 mg per mL, incubation was performed for 30 minutes outcome measures. The CCI1hr allows to assess the efﬁcacy at 22°C. FITC-conjugated F(ab′)2 fragments of rabbit of PLT transfusion accounting for the body surface area anti-mouse immunoglobulin (Dako Diagnostics, Zug, (BSA) and the number of transfused PLTs. It is calculated Switzerland; ﬁnal concentration 1.5 mg/mL) was added as CCI1hr = BSA [m2] ¥ (posttransfusion count [109/ and incubated in the dark for an additional 30 minutes. L] - pretransfusion count [109/L])/PLT dose [1011/L].22,23 The samples were then diluted at 1:20 with 20 mmol per L Because in infants and children, blood volume is propor- PBS and analyzed within 2 hours on a ﬂow cytometer tional to BW and is only suboptimally related to BSA,24 we (FACScan, Becton Dickinson) equipped with a 15-mW air- also used the BW-corrected measure, PPR1hr, as proposed cooled argon laser. A total of 10,000 events per sample by Hanson and Slichter25 and Davis and coworkers.26 It is were gated and analyzed with computer software (Cell calculated as PPR1hr [%] = (posttransfusion count [109/ Quest, Becton Dickinson). Results are reported as the per- L] - pretransfusion count [109/L]) ¥ blood volume [L]/PLT centage of PLTs positive for A antigen expression. Because Volume 49, January 2009 TRANSFUSION 23 JULMY ET AL. donors are not screened for blood subgroups at our The Hodges-Lehmann method and the Wilcoxon U center, we based the deﬁnition of subgroups A1 and A2 on test cannot be extended to multivariate analyses. The a publication of Cooling and colleagues.6 According to British Committee for Standards in Haematology deﬁnes a their data, no A antigen was detected by Western blot in A2 successful PLT transfusion as one with a PPR1hr of more donors, while by ﬂow cytometry the range of A antigen- than 30 percent; thus, the outcome variable was dichoto- positive PLTs varied between 4.3 and less than 1.0 percent. mized into successful versus unsuccessful transfusions. Being aware that some A2 donors will erroneously be This binary outcome variable was then analyzed by logis- allocated as group A1, we deﬁned individuals with blood tic regression stratiﬁed for patients. First, ABO-matching group A presenting more than 1 percent A antigen- status (identical vs. ABO major-mismatched transfusion positive PLTs by ﬂow cytometry to be of blood group A1 for the main study objective; other comparisons for sec- and donors with 1 percent or less to be of subgroup A2. ondary study objectives) was tested for potential associa- tion with transfusion efﬁcacy, that is, PPR1hr, by exact univariate logistic regression. Second, asymptotic multi- Fluorescent microscopy variate logistic regression was performed in a similar Cytocentrifuge PLT preparations were ﬁxed with 4 percent way to test if ABO-matching status was associated with paraformaldehyde and stained with anti-A BRIC-145 and transfusion efﬁcacy independently from other patient- with secondary antibody Cy-3 (Jackson ImmunoResearch and APC-related variables. Exact multivariate logistic Laboratories, Inc., West Grove, PA). Then, MoAb anti- regression proved not feasible. The multivariate logistic CD61 (Becton Dickinson) was used with a biotinylated regression model started with all variables signiﬁcantly bridge antibody (Jackson ImmunoResearch), followed by associated with transfusion efﬁcacy in univariate analysis anti-biotin Alexa 488 antibody (protein labeling kit (p < 0.05). The backward variable selection procedure was A-10235, Molecular Probes, Eugene, OR) to identify glyco- applied until all variables remaining in the model were protein IIIa on PLTs. Pictures were obtained by an imaging signiﬁcantly (p < 0.05) and independently associated with system (Zeiss Axioplan 2, Zeiss, Zürich, Switzerland). transfusion efﬁcacy. ABO-matching status was then included into this model. Baseline APC-related variables considered for multi- Isohemagglutinin titer testing variate analysis were ABO and D blood group, storage In many recipients, pretransfusion blood counts were time, volume, PLT concentration, absolute PLT count, determined by capillary blood sampling. Thus, isohemag- apheresis procedure including apheresis device, duration glutinin titers could only be analyzed for the patients of thrombocytapheresis, processed blood volume, blood undergoing venous blood sampling. With minor- ﬂow, number of PLTs harvested per minute, yield per aph- mismatched transfusions, isohemagglutinin titers were eresis, weight before splitting, number of splits, and determined in the donor plasma. For this purpose, residu- gamma irradiation. Baseline patient variables considered als of APCs remaining in the tube of the transfusion set were ABO and D blood group, age, weight, and sex. were transferred to test tubes to obtain plasma by cen- Power analysis determined a sample size of 400 trans- trifugation. Isohemagglutinin titers were determined fusions for the primary objective (a = 5%, power 90%, using a microtyping system (DiaMed-ID, DiaMed, Cressier difference in the CCI1hr 5, estimated proportion of major- sur Morat, Switzerland) according to the manufacturer’s mismatched transfusions 25%, and sequential interim instructions. analyses after 100, 200, and 300 transfusions). Two-sided tests were used throughout, and p values below 0.05 were considered signiﬁcant. No correction for the multiple com- Statistical analysis parisons performed for secondary study objectives was Because of nonnormally distributed data, medians, applied. All exact analyses were performed with computer ranges, and interquartile ranges (IQRs) were calculated for software (StatXact 6 and LogXact 6, Cytel Software Corp., descriptive statistics, and nonparametric exact methods Cambridge, MA); another software program (EaSt-2000, were used wherever feasible for analytical statistics. For Cytel Software Corp.) was used for sample size determina- univariate main and secondary analyses, between-group tion, performing of interim analyses, and correction of the differences and their 95 percent conﬁdence intervals main-analysis p values for the interim analyses. (CIs) were estimated using the exact Hodges-Lehmann method. The corresponding p value was calculated using RESULTS the exact Wilcoxon U test, stratiﬁed per patient. These nonparametric methods correspond to parametric linear Patient characteristics, PLT transfusions, and mixed-effect (LME) modeling, which accounts for the cor- indications relation among multiple transfusions given to the same During the study period, 52 children and adolescents patient. received 528 PLT transfusions, of which 128 (24%) were 24 TRANSFUSION Volume 49, January 2009 TRANSFUSION EFFICACY OF MISMATCHED PLTs excluded from analysis for the following reasons: fever given prophylactically and 14 (3.5%) therapeutically of 38.5°C or greater before transfusion (n = 73), missing (Table 2). pre- or posttransfusion PLT count (n = 16), clinically enlarged spleen (n = 16), thrombosis (n = 11), APC obtained from a foreign blood bank (n = 7), refusal of Constellations of ABO blood group matching of study participation (n = 3), adverse events requiring donors and recipients interruption of transfusion (n = 1), and hemorrhage of The 400 PLT transfusions included 282 (71%) ABO- WHO Grade 3 or greater (n = 1). The eligible study identical, 76 (19%) major-mismatched, and 42 (11%) population consisted of 50 children (29 girls, 21 boys) minor-mismatched transfusions, respectively (Table 3). with a median age of 6.7 years at ﬁrst transfusion (range, Among the 76 major-mismatched transfusions, there 0.2 to 16.1 years). They presented with thrombocytope- were 49 (12%) major-mismatched A1, 17 (4.3%) major- nia due to the malignancy itself or the applied chemo- mismatched A2, and 8 (2.0%) major-mismatched transfu- therapy or because of aplastic anemia (Table 1). The sions with other constellations (A1B, B). For 2 donors, 1 median number of APCs transfused per patient was 5 with blood group A and 1 with blood group AB, A1/A2 sub- (IQR, 3-10). Of 400 PLT transfusions, 386 (96.5%) were group typing by ﬂow cytometry was not available. TABLE 1. Patient characteristics, number of PLT transfusions, and Interim analyses of the primary ABO blood group of patients and transfused APCs study variable Patients (%), Transfusions (%), The three planned interim analyses did Variables n = 50 n = 400 not result in a proposal to stop the study Sex Male 21 (42%) 189 (47%) early because of proven difference or Female 29 (58%) 211 (53%) proven futility. The standardized statis- Age (years) tics resulting from LME analysis of <1* 3 (6%) 11 (3%) 1 and <2* 5 (10%) 38 (10%) CCI1hr in ABO matched versus major- 2* 42 (84%) 351 (88%) mismatched transfusions were 0.81 for Diagnosis of patients the ﬁrst 100 transfusions (boundary to Acute lymphoblastic leukemia 23 (46%) 141 (35%) Acute myelogenous leukemia 6 (12%) 140 (35%) reject null hypothesis, 3.83; no bound- Lymphoma 1 (2%) 1 (0%) ary deﬁned to reject alternative hypoth- Tumor of the central nervous system 11 (22%) 58 (15%) esis), 0.65 for 200 transfusions ( 2.80; Extracranial solid tumor 8 (16%) 38 (10%) Severe aplastic anemia 1 (2%) 22 (6%) 0.40), and 1.64 for 300 transfusions ABO blood group ( 2.25; 1.37). At the time of study A 22 (44%) 200 (50%) design and interim analyses, CCI1hr to be B 8 (16%) 18 (5%) AB 2 (4%) 8 (2%) analyzed by LME was deﬁned as the O 18 (36%) 174 (44%) principal outcome measure. At the stage * Patients less than 1 year received 10 to 15 mL per kg BW of the APC product, patients of the ﬁnal analysis, CCI1hr was replaced older than 1 and less than 2 years half an APC, and patients 2 years and older 1 APC. by PPR1hr because of its improved TABLE 2. Indications for transfusion and corresponding pretransfusion PLT counts Pretransfusion PLT count (¥109/L) Indications Median IQR Transfusions (%), n = 400 Therapeutic use Minor bleeding (WHO Grade 2) 17 6 to 37 14 (4%) Prophylactic use Before RBC transfusion* 9 6 to 15 138 (35%) Before invasive intervention† 20 13 to 35 91 (23%) Hospitalization because of severe neutropenia and fever*‡ 9 7 to 15 65 (16%) Prevention of further bleeding after major hemorrhage 23 19 to 40 40 (10%) Before discharge from hospital* 9 4 to 14 38 (10%) Autologous stem cell transplantation‡§ 10 8 to 14 14 (4%) * A PLT count of fewer than 10 ¥ 109 per L triggered prophylactic transfusion. † Intramuscular injection, surgery, insertion of a central venous catheter, lumbar puncture. In case of circulating leukemic cells in the periph- eral blood before lumbar puncture, PLTs were transfused when the PLT count was fewer than 100 ¥ 109 per L.33 ‡ Only transfusions given to patients with a body temperature of less than 38.5°C were considered eligible for this study. § Without fever a PLT count of fewer than 10 ¥ 109 per L and with fever a count of fewer than 20 ¥ 109 per L triggered prophylactic transfu- sion, respectively. Volume 49, January 2009 TRANSFUSION 25 JULMY ET AL. from analysis, the failure rate TABLE 3. ABO blood group matching constellations (n = 400) of major-mismatched transfusions Blood group was even more striking (p = 0.002). Recipient However, minor-mismatched, plasma- Donor A B AB O incompatible transfusions and identical A 129 (32%) 31 (7.8%)*‡ 4 (1.0%)† 36 (9.0%)*§ A1 No typing done 22 (5.5%)*‡ No typing done 27 (6.8%)* transfusions were similarly successful A2 No typing done 9 (2.2%)*‡ No typing done 8 (2.0%)* (p = 0.48). B 2 (0.5%)*‡ 14 (3.5%) 0 2 (0.5%)* In the multivariate analysis, pro- AB 1 (0.3%)* 3 (0.8%)* 3 (0.8%) 1 (0.3%)*§ A1B 1 (0.3%)* 3 (0.8%)* 3 (0.8%) 0 duction of APCs according to the A2B 0 0 0 0 Amicus procedure, which included O 18 (4.5%)† 19 (4.8%)† 1 (0.3%)† 136 (34%) storage in PLT AS (p < 0.001), a longer Total 150 (38%) 67 (17%) 8 (2%) 175 (44%) storage time before transfusion * Major-mismatched transfusions. (p = 0.003), and a higher PLT yield per † Minor-mismatched transfusions. ‡ Bidirectionally mismatched transfusions (i.e., both major- and minor-mismatched). apheresis (p < 0.001) were signiﬁcantly § For one donor with blood group A and one donor with blood group AB, A1/A2 subgroup associated with a higher number of typing was not available. unsuccessful transfusions, whereas a higher PLT count per APC (p < 0.001) and an increasing BW of the recipient accuracy in infants and small children25,26 and exact logis- (p = 0.003) had a positive impact (Table 5). In this context, tic regression replaced parametric LME. Interim analyses we have recently reported on the effects of these variables using exact logistic regression on PPR1hr, however, would on the efﬁcacy of transfused PLTs.27 have resulted in the same decisions, with standardized statistics of 0.35, 1.15, and 1.75 after 100, 200, and 300 transfusions, respectively, in post hoc calculations. Fate of transfused A antigen-positive PLTs In 32 major-mismatched transfusions from group A1 donors to group O (n = 18) or group B (n = 14) recipients, PLT transfusion efficacy the percentage of A antigen-positive PLTs was determined Table 4 shows the transfusion efﬁcacy depending on ABO by ﬂow cytometry in the APC and in the recipient before matching. The median PPR1hr of all transfusions was 31 and 1 hour after transfusion. In the recipient, there were percent. Univariate analysis showed PPR1hr for 76 major- no A antigen-positive PLTs before transfusion (data not mismatched transfusions to be signiﬁcantly inferior to shown). Figures 1A and 1B show that in the majority of that for 282 ABO-identical transfusions (median, 21% vs. group O and B recipients, no A antigen-positive PLTs were 32%; p = 0.034); notably for major-mismatched A1 PLTs detectable in the circulation 1 hour after transfusion. In the median PPR1hr was 44 percent lower compared to contrast, after ABO-identical PLT transfusions from group ABO-identical PLTs (median, 18% vs. 32%; p = 0.007). Con- A1 donors to group A2 recipients, whose RBCs but not PLTs cerning transfusion success or failure, major-mismatched express A antigen, the majority of transfused A antigen- transfusions were signiﬁcantly more often unsuccessful positive PLTs remained detectable in the circulation than ABO-identical transfusions (p = 0.033), especially (Fig. 1C). Figure 2 depicts ﬂow-cytometric histograms of many major-mismatched transfusions from subgroup A1 four typical PLT transfusions with different ABO constel- donors failed (p = 0.002). Major-mismatched A2 PLTs lations. The percentages of A antigen-positive PLTs in the were as successful as ABO-identical ones (p = 1.00). No APC and blood of the recipient before and after transfu- signiﬁcant difference between ABO-identical and ABO sion are indicated. It is noteworthy that A antigen-positive nonidentical transfusions (i.e., the sum of major- and PLTs, which remained detectable 1 hour after transfusion, minor-mismatched transfusions) was seen (p = 0.22; expressed low amounts of A antigen only (Figs. 2A and 2B, Table 4). Row 3). In Fig. 3, using the MoAb anti-CD61 for identiﬁca- Multivariate analysis fully conﬁrmed these results tion of PLTs and anti-A BRIC-145 for expression of A (Table 5). Major-mismatched transfusions were signiﬁ- antigen, depletion of transfused A antigen-positive PLTs cantly more often unsuccessful than ABO-identical trans- from the circulation of a blood group O recipient is visu- fusions (p = 0.005). As shown in Table 6, subgroup analysis alized at a single-cell level by dual-color ﬂuorescent revealed that PLTs from donors with ABO subgroup A2 microscopy. Regression analysis showed that for group A given to group O or B recipients, although major- major-mismatched transfusions, there was a signiﬁcant mismatched transfusions, by deﬁnition, were just as suc- negative correlation (p = 0.002) between the percentage of cessful as ABO-identical PLTs (p = 0.90). Therefore, by A antigen-positive PLTs in the APC and the PPR1hr; that is, reanalyzing ABO-identical versus major-mismatched the more PLTs expressed A antigen, the lower the PPR1hr transfusions after excluding the 17 A2 PLT products (Fig. 4). 26 TRANSFUSION Volume 49, January 2009 TRANSFUSION EFFICACY OF MISMATCHED PLTs To substantiate the hypothesis that in group O or B p Value 0.033‡ recipients, the transfused A antigen-positive PLTs are 0.002 0.007 1.00 0.59 0.22 rapidly cleared from the circulation in vivo, we performed in vitro experiments (n = 3; Fig. 5). Blood treated with eth- ylenediaminetetraacetic acid (EDTA) was obtained from 1.65 to 18.49 1.18 to 6.44 0.14 to 4.21 1.34 to 9.72 0.26 to 1.89 0.81 to 2.81 recipients with blood group O before PLT transfusion. The Reference 95% CI patients were transfused with an A1 major-mismatched 30% APCs. Thereafter, transfused PLTs (100 mL) remaining in the transfusion set were added to the recipient’s blood sample. TABLE 4. Univariate analysis of PLT transfusion efficacy and transfusion failure* depending on ABO matching PPR1hr One hour later, blood samples from both the patients and According to the British Committee for Standards in Haematology an unsuccessful PLT transfusion was deﬁned as one with a PPR1hr of 30 percent or less.2 the in vitro prepared samples were analyzed in parallel by 2.67 4.98 0.78 3.41 0.71 1.49 OR ﬂow cytometry. In the recipient’s blood, no A antigen- positive PLTs were detected, whereas in vitro similar per- Number (%) centages of A antigen-positive PLTs that were previously (47%) (44%) (62%) (71%) (35%) (68%) (40%) (54%) measured in the transfused APCs were recovered. 187 123 47 35 6 39 17 64 Safety review and isohemagglutinin titers p Value During the whole study period in which the transfusion of 0.034† 0.005 0.007 0.78 0.49 0.22 528 APCs was available for safety review, no clinically detectable hemolytic transfusion reaction was observed; however, we did not test for subclinical manifestations of -12.5 to -1.6 -15.9 to -4.5 -15.8 to -4.2 -5.4 to 15.5 -4.1 to 10.1 p Value adjusted from a calculated p value of 0.017 for the interim analyses performed for this primary outcome. p Value adjusted from a calculated p value of 0.016 for the interim analyses performed for this primary outcome. hemolysis. Notably, only 1 (0.2%) of 528 transfusions had References -8.2 to 1.2 95% CI to be interrupted due to a major transfusion reaction (angioedema). Minor transfusion reactions were also rare (6/528; 1.1%). In 32 plasma-incompatible transfusions, isohemagglutinin titers were determined in the APCs revealing no titer greater than 32. In the recipients, titers Estimated difference PPR1hr (%) were determined before PLT transfusion in 228 cases revealing low medians of anti-A and anti-B isohemagglu- -7.2 -10.6 5.1 -10.3 2.9 -3.9 tinins (Table 7). Statistically, no association between transfusion efﬁcacy and the recipient’s isohemagglutinin titers were detected (data not shown). It is noteworthy that during the study period no patient became PLT refractory, shown by an adequate PPR1hr after subsequent transfu- 45 44 40 34 52 37 52 45 sions, and therefore HLA/HPA-matched PLTs were never IQR to to to to to to to to required. 17 19 12 13 17 10 16 14 The sum of major- and minor-mismatched transfusions. DISCUSSION Median 31 32 21 18 36 18 34 26 PLT transfusions are administered to patients with throm- bocytopenia to prevent or stop hemorrhage. Because intensive treatment regimens, including allogeneic and Number 400 282 76 49 17 57 42 118 autologous stem cell transplantation, are increasingly applied, hematology/oncology patients are currently among the largest groups receiving PLT transfusions.19,28 In this context, the clinical signiﬁcance of ABO compat- Major-mismatched non-A2 ibility in PLT transfusion has been a matter of debate since ABO major-mismatched ABO minor-mismatched Major-mismatched A1 Major-mismatched A2 the early studies of Aster,10 more than 40 years ago. He and ABO non identical§ others have recognized that ABO-incompatible PLTs PLT transfusions may be associated with decreased PLT increments after ABO-identical transfusion.11-14 However, not only PLT-incompatible transfusions, that is, major-mismatched transfusions, Type Total but also plasma-incompatible transfusions, that is, minor- † ‡ § * mismatched transfusions, have attracted attention, Volume 49, January 2009 TRANSFUSION 27 JULMY ET AL. because they have been linked to acute TABLE 5. Multivariate logistic regression of factors affecting hemolytic transfusion reactions, transfusion failure* (primary objective) venoocclusive disease, and increased PPR1hr 30% morbidity in allogeneic transplant Variable OR 95% CI p Value recipients.9,29,30 ABO matching–related ABO major-mismatched vs. -identical 3.97 1.52 to 10.39 0.005 The majority of studies have been Product-related performed in adult patients, and PLT Apheresis and storage procedure: 7.82 3.35 to 18.25 <0.001 transfusion in children is poorly investi- Amicus vs. others Storage time of APC (per day) 1.46 1.13 to 1.87 0.003 gated. In addition, in the majority of Absolute PLT count of APC (per 1011) 0.090 0.023 to 0.354 <0.001 older reports, transfusion of pooled PLT Yield of apheresis (per 1011 PLTs) 1.37 1.16 to 1.63 <0.001 concentrates from random units of Recipient-related Weight of recipient (per kg) 0.83 0.74 to 0.94 0.002 whole blood has been studied, but these * According to the British Committee for Standards in Haematology, unsuccessful PLT PLT products have been widely replaced transfusion is deﬁned as one with a PPR1hr of 30 percent or less.2 by APCs within the past decade.18,19 With modern apheresis equipment, a high PLT yield can be obtained allowing the production of two, three, or even more APCs from a single donation.20,21 TABLE 6. Multivariate logistic regression of ABO-matching The aim of this prospective study constellations affecting transfusion failure* (secondary objectives)† was to investigate transfusion efﬁcacy PPR1hr 30% and transfusion success of APCs in Variable OR 95% CI p Value children with thrombocytopenia. The ABO matching–related Major-mismatched A1 vs. identical 9.30 2.53 to 34.24 <0.001 primary objective was to test the Major-mismatched A2 vs. identical 1.13 0.16 to 7.88 0.90 hypothesis that transfusion efﬁcacy of Major-mismatched non A2 vs. identical 5.81 1.94 to 17.38 0.002 ABO major-mismatched PLTs is signiﬁ- ABO minor-mismatched vs. identical 0.67 0.22 to 2.02 0.48 ABO nonidentical vs. identical 1.69 0.83 to 3.45 0.15 cantly inferior to that of ABO blood * According to the British Committee for Standards in Haematology, unsuccessful PLT group–identical PLTs. A randomized transfusion is deﬁned as one with a PPR1hr of 30 percent or less.2 study design would have increased the † Each subgroup analysis was adjusted for the corresponding patient- and product- power of our study. However, in view of related variables (data not shown). the preceding studies in adults,10-13,15,16 we considered such a design incompat- ible with good clinical practice in this vulnerable pediatric population. There- fore, power calculation was based on a retrospective review of the ABO- matching status of 50 APCs that had been delivered between May 2003 and October 2003, in line with the policy of our division of transfusion medicine, which remained unchanged during the study period. Since PLT transfusion efﬁ- cacy not only depends on ABO match- ing, patient- and APC-related factors were also assessed prospectively. While planning the study, we con- sidered basing the primary study objec- tive on a robust clinical endpoint, such as cessation of bleeding. However, a retrospective review revealed that in Fig. 1. Percentage of A antigen-expressing PLTs in APCs and blood of recipients 1 our pediatric hematology/oncology pa- hour after transfusion. Before transfusion, no A antigen-positive PLTs were detected tients, the vast majority of PLT transfu- in the recipients (data not shown). (A) Major-mismatched transfusions of group A1 sions were given prophylactically rather PLTs to group O recipients (n = 18), (B) to group B recipients (n = 14), and (C) identi- than therapeutically,28,31 and the fre- cal transfusions of A1 PLTs to group A2 recipients (n = 3; serving as in vivo control). In quency of major and minor bleeding each panel, the medians are shown by bars. episodes was very low. Therefore, we 28 TRANSFUSION Volume 49, January 2009 TRANSFUSION EFFICACY OF MISMATCHED PLTs Volume 49, January 2009 TRANSFUSION 29 JULMY ET AL. Fig. 2. Percentage of A antigen-positive PLTs in the APCs and blood of recipients before and 1 hour after major-mismatched trans- fusion (A-C) or identical transfusion (D). (A-D) PLTs were stained with BRIC-145 and labeled with FITC to identify the blood group A antigen. (A) Group O recipient transfused with group A1 PLTs, (B) group B recipient transfused with group A1 PLTs, (C) group O recipient transfused with group A2 PLTs, and (D) group A2 recipient transfused with group A1 PLTs. Columns 1, 2, and 3 show respec- tive histograms of the recipient’s PLTs before transfusion, of the APCs, and of the recipient’s PLTs 1 hour after transfusion. Horizon- tal brackets indicate A antigen-positive PLTs. PLTs left of brackets are considered to be A antigen-negative. In each histogram the percentage of A antigen-positive PLTs is inserted as well as the particular ABO blood group of the recipient and/or donor. MFI = mean ﬂuorescence intensity. Fig. 3. Fluorescent microscopy of A antigen-positive PLTs in an APC and in the blood of a blood group O recipient 1 hour after major-mismatched transfusion. Fixed PLTs were stained with anti-A (red) and anti-CD61 (green) as described under Materials and Methods. Column A = PLTs of an APC obtained from a donor with blood group A1. In this APC, 41 percent of the PLTs expressed A antigen as measured by ﬂow cytometry. Column B = PLTs from the recipient 1 hour after transfusion of the APC shown in Column A. Row 1 = PLTs expressing A antigen (red); Row 2 = PLTs identiﬁed with anti-CD61 (green); Row 3 = fusion dual-color picture depicting A antigen-positive PLTs in yellow whereas A antigen-negative PLTs remain green. After transfusion, conﬁrmed by ﬂow cytometry, no (0%) A antigen-positive PLTs remained in the circulation of the recipient. chose to base our statistical endpoints on CCI1hr and PPR1hr Our results demonstrate that transfusion efﬁcacy and as surrogate parameters for PLT transfusion efﬁcacy. CCI1hr transfusion success, the latter deﬁned as PPR1hr of more accounts for BSA and is well established in adults,22,23 than 30 percent, of major-mismatched PLTs are signiﬁ- whereas PPR1hr accounts for BW and is reported to be more cantly inferior to those of ABO-identical PLTs, whereas appropriate in children, particularly in infants and tod- efﬁcacy and success of minor-mismatched transfusions dlers.25,26 Because of the latter the ﬁnal analysis was based are not different from those of identical transfusions. It on PPR1hr. is noteworthy that comparison of transfusion success 30 TRANSFUSION Volume 49, January 2009 TRANSFUSION EFFICACY OF MISMATCHED PLTs of nonidentical (i.e., the sum of minor- and major- efﬁcacy between ABO-identical and major-mismatched mismatched) PLTs with that of ABO-identical PLTs did not PLTs could have been missed.17 reveal any signiﬁcant difference (Table 6). A limitation of In accordance with earlier reports,5,6 we observed a some older studies is that nonidentical transfusions were dramatic diversity in the percentages of PLTs expressing A not categorized into minor- (plasma-incompatible) and antigen among A1 donors (median, 40%; range, 3%-80%). major- (PLT-incompatible) mismatched transfusions, High A antigen expression may be the main pathophysi- bearing the risk that signiﬁcant differences in transfusion ologic mechanism explaining the rapid clearance of these PLTs from circulation after transfusion into group O or B recipients. This notion appears to be supported by our observation that the percentage of A antigen-expressing PLTs and PPR1hr were inversely related; that is, the higher the percentage of A antigen-expressing PLTs, the lower the PPR1hr (Fig. 4). Moreover, our experiments showed that after in vitro incubation of A antigen-positive PLTs in group O blood samples, the percentages of positive PLTs did little change, while in vivo, 1 hour after transfusion, A antigen-positive PLTs were no longer detected in the blood of the recipients (Fig. 5). Therefore, we hypothesize that any PLT increment after transfusing PLTs from group A1 donors into group O or B recipients is due to PLTs expressing no or low levels of A antigen. Because median anti-A titers in the plasma of blood group O recipients was Fig. 4. Linear regression of PPR1hr and percentage of A higher (8) than in plasma of group B recipients (2; Table 7), antigen-positive PLTs in group A major-mismatched transfu- we speculate that this may explain the slightly better sur- sions. The solid line represents the linear regression line, and vival of transfused A1 PLTs in group B recipients (Fig. 1). the dashed line indicates the border between a successful Furthermore, our data conﬁrmed the conclusion of (>30%) and an unsuccessful PPR1hr ( 30%) as deﬁned by the others6 that A2 PLTs can be considered group O compat- British Committee for Standards in Haematology.2 ible. With 8 percent of the Swiss population having blood Fig. 5. Comparison of A antigen expression in vivo and in vitro in a major-mismatched transfusion constellation. A antigen-positive donor’s PLTs (100 mL) were added in vitro to 2 mL EDTA blood obtained from a recipient with blood group O before transfusion and incubated for 1 hour. Histograms are explained in the legend to Fig. 2. (A) A antigen expression in the APC, (B) in the recipient before and (C) 1 hour after transfusion, and (D) in the sample prepared in vitro. In each histogram the percentage of A antigen- positive PLTs is inserted. TABLE 7. Isohemagglutinin titers in APCs and plasma of recipients before transfusion Plasma-incompatible PLT concentrates (n = 32) Recipients (n = 228) Anti-A Anti-B Anti-A Anti-B Blood group Number Median Range Median Range Number Median Range Median Range O 16 8 1-32 4 1-16 108 8 0-128 4 0-128 A 15 4 1-32 86 4 0-64 B 1 4 34 2 0-32 Volume 49, January 2009 TRANSFUSION 31 JULMY ET AL. group A2,32 routine A1/A2 subtyping of group A donors manuscript. We also wish to thank all patients and their families would expand the available inventory of compatible PLTs for having consented to participate in this study. for group O recipients. It remains to be shown if PLTs from blood group A1 donors with a low expresser phenotype REFERENCES would produce a sufﬁcient PLT increase in group O or B recipients. 1. American Association of Blood Banks. Technical manual. Whereas ABO matching clearly inﬂuenced transfu- 15th ed. Bethesda (MD): American Association of Blood sion success or failure, further factors revealed to be sig- Banks; 2005. niﬁcantly associated with transfusion efﬁcacy. Whereas 2. British Committee for Standards in Haematology, Blood the apheresis procedure including storage conditions, the Transfusion Task Force (Chairman P Kelsey). Guidelines for storage time, and the PLT yield per apheresis procedure the use of platelet transfusions. Br J Haematol 2003;122:10- had an unfavorable impact, a higher PLT count per APC 23. and increasing BW of the recipient favored successful 3. Julmy F, Achermann F, Schulzki T, Carrel T, Nydegger U. transfusions (Table 5). Therefore, separate from this study, PLTs of blood group A1 donors express increased surface A we have recently reported on the effects of these variables antigen owing to apheresis and prolonged storage. Trans- on the transfusion efﬁcacy of PLT products.27 fusion 2003;43:1378-85. Safety review of transfusions conﬁrmed the good 4. Ogasawara K, Ueki J, Takenaka M, Furihata K. Study on the tolerability of PLT transfusions. No clinically detectable expression of ABH antigens on platelets. Blood 1993;82: hemolytic reactions were observed in 528 recorded PLT 993-9. transfusions disregarding the matching status, and there 5. Curtis BR, Edwards JT, Hessner MJ, Klein JP, Aster RH. was only 1 (0.2%) transfusion that had to be interrupted Blood group A and B antigens are strongly expressed due to a major transfusion reaction. Each group O PLT on platelets of some individuals. Blood 2000;96: donor was routinely screened for the presence of anti-A 1574-81. and anti-B hemolysins. If a titer of greater than 4 was 6. Cooling LL, Kelly K, Barton J, Hwang D, Koerner TA, Olson present, according to the standards of the Swiss Red JD. Determinants of ABH expression on human blood Cross, the product was labeled accordingly and released platelets. Blood 2005;105:3356-64. for group O recipients only. The latter fact may account 7. Duguid JK, Minards J, Bolton-Maggs PH. Lesson of the for absence of hemolysis after plasma-incompatible week: incompatible plasma transfusions and haemolysis in transfusions. Furthermore, no patient became PLT children. BMJ 1999;318:176-7. refractory, and patients’ anti-A and anti-B isohemagglu- 8. Larsson LG, Welsh VJ, Ladd DJ. Acute intravascular tinins remained low despite having received major- hemolysis secondary to out-of-group platelet transfusion. mismatched PLT transfusions, in many cases more than Transfusion 2000;40:902-6. one. 9. Herman JH, King KE. Apheresis platelet transfusions: does In conclusion, major-mismatched PLT transfusions ABO matter? Transfusion 2004;44:802-4. were signiﬁcantly less successful than ABO-identical 10. Aster RH. Effect of anticoagulant and ABO incompatibility transfusions. In children requiring regular and continued on recovery of transfused human platelets. Blood 1965;26: PLT support, an ABO-compatible transfusion strategy 732-43. should be the ﬁrst choice. In this context, A2 PLTs are 11. Lee EJ, Schiffer CA. ABO compatibility can inﬂuence the equivalent to group O PLTs. For this ﬁnding to be of results of platelet transfusion. Results of a randomized practical value, PLT concentrate suppliers would have to trial. Transfusion 1989;29:384-9. provide A1/A2 subgroup typing of group A PLT donors. In 12. Heal JM, Rowe JM, McMican A, Masel D, Finke C, this study, transfusion of plasma-incompatible PLTs was Blumberg N. The role of ABO matching in platelet as successful as that of identical PLTs, although an inher- transfusion. Eur J Haematol 1993;50:110-7. ent risk for hemolytic transfusion reactions will remain. 13. Jimenez TM, Patel SB, Pineda AA, Tefferi A, Owen WG. Factors that inﬂuence platelet recovery after transfusion: resolving donor quality from ABO compatibility. Transfu- ACKNOWLEDGMENTS sion 2003;43:328-34. The authors thank all members of the medical, nursing, and labo- 14. Slichter SJ, Davis K, Enright H, Braine H, Gernsheimer T, ratory staff of the Division of Pediatric Hematology/Oncology of Kao KJ, Kickler T, Lee E, McFarland J, McCullough J, Rodey the University Children’s Hospital Berne and the Division of G, Schiffer CA, Woodson R. Factors affecting posttransfu- Transfusion Medicine of the Department of Hematology, Univer- sion platelet increments, platelet refractoriness, and plate- sity Hospital Berne. We are especially indebted to the nurses of let transfusion intervals in thrombocytopenic patients. the apheresis unit. We are grateful to Jakob Zbaeren for his excel- Blood 2005;105:4106-14. lent technical assistance in ﬂuorescent microscopy and to Dr 15. Heal JM, Blumberg N, Kirkley SA, DiPersio JF, Rapoport AP, Silvia M. Rogers (MediWrite, Basel, Switzerland) for reviewing the Rowe JM. Leukocyte-reduced transfusions of ABO-identical 32 TRANSFUSION Volume 49, January 2009 TRANSFUSION EFFICACY OF MISMATCHED PLTs platelets and clinical outcome in autologous bone marrow bone marrow hypoplasia: evidence for a ﬁxed platelet transplantation for lymphoma. Bone Marrow Transplant requirement. Blood 1985;66:1105-9. 1994;14:943-8. 26. Davis KB, Slichter SJ, Corash L. Corrected count increment 16. Blumberg N, Heal JM. ABO-mismatched platelet transfu- and percent platelet recovery as measures of posttransfu- sions and clinical outcomes after cardiac surgery. Transfu- sion platelet response: problems and a solution. Transfu- sion 2002;42:1527-8; author reply 8-9. sion 1999;39:586-92. 17. van Eys J, Thomas D, Olivos B. Platelet use in pediatric 27. Julmy F, Ammann RA, Mansouri Taleghani B, Fontana S, oncology: a review of 393 transfusions. Transfusion 1978; Hirt A, Leibundgut K. Effects of high-yield thrombocyta- 18:169-73. pheresis on the quality of platelet products. Transfusion 18. Ness PM, Campbell-Lee SA. Single donor versus pooled 2008;48:442-50. random donor platelet concentrates. Curr Opin Hematol 28. McCullough J. Current issues with platelet transfusion in 2001;8:392-6. patients with cancer. Semin Hematol 2000;37 Suppl 4: 19. Sullivan MT, Cotten R, Read EJ, Wallace EL. Blood collec- 3-10. tion and transfusion in the United States in 2001. Transfu- 29. Benjamin RJ, Antin JH. ABO-incompatible bone marrow sion 2007;47:385-94. transplantation: the transfusion of incompatible plasma 20. Vassallo RR Jr, Wahab F, Giordano K, Murphy S. Improving may exacerbate regimen-related toxicity. Transfusion 1999; technology for collecting platelets by apheresis: ﬁve-year 39:1273-4. experience in one blood center. Transfus Med Rev 2004;18: 30. Lapierre V, Mahe C, Auperin A, Stambouli F, Oubouzar N, 257-66. Tramalloni D, Benhamou E, Tiberghien P, Hartmann O. 21. Fontana S, Mordasini L, Keller P, Taleghani BM. Prospec- Platelet transfusion containing ABO-incompatible plasma tive, paired crossover comparison of multiple, single- and hepatic veno-occlusive disease after hematopoietic needle plateletpheresis procedures with the Amicus and transplantation in young children. Transplantation 2005; Trima Accel cell separators. Transfusion 2006;46:2004-10. 80:314-9. 22. Daly PA, Schiffer CA, Aisner J, Wiernik PH. Platelet transfu- 31. Stroncek DF, Rebulla P. Platelet transfusions. Lancet 2007; sion therapy. One-hour posttransfusion increments are 370:427-38. valuable in predicting the need for HLA-matched prepara- 32. Achermann FJ, Julmy F, Gilliver LG, Carrel TP, Nydegger tions. JAMA 1980;243:435-8. UE. Soluble type A substance in fresh-frozen plasma as a 23. Yankee RA, Grumet FC, Rogentine GN. Platelet transfusion function of ABO and Secretor genotypes and Lewis pheno- the selection of compatible platelet donors for refractory type. Transfus Apher Sci 2005;32:255-62. patients by lymphocyte HL-A typing. N Engl J Med 1969; 33. Gajjar A, Harrison PL, Sandlund JT, Rivera GK, Ribeiro RC, 281:1208-12. Rubnitz JE, Razzouk B, Relling MV, Evans WE, Boyett JM, 24. Linderkamp O, Versmold HT, Riegel KP, Betke K. Estima- Pui C-H. Traumatic lumbar puncture at diagnosis tion and prediction of blood volume in infants and chil- adversely affects outcome in childhood acute lymphoblas- dren. Eur J Pediatr 1977;125:227-34. tic leukemia. Blood 2000;96:3381-4. 25. Hanson SR, Slichter SJ. Platelet kinetics in patients with Volume 49, January 2009 TRANSFUSION 33
"Transfusion efficacy of ABO major mismatched platelets PLTs in"