ALBERT

Description: All sickle cell anemia patients (HbSS, SCA) have the same genetic mutation, but the clinical phenotype is highly variable and difficult to predict prior to the onset of complications. Transcranial Doppler (TCD) is the only validated predictive indicator of severe SCA currently available, identifying SCA patients age 2 years or older with the greatest stroke risk. Preliminary studies have identified that elevated absolute reticulocyte count (ARC) between 2 and 6 months of age is associated with an increased risk of hospitalization for SCA complications, stroke and death.1,2 To determine if early reticulocyte levels in SCA patients are useful in classifying children at highest risk of developing an abnormal (abTCD) or conditional (cdTCD) TCD, 121 consecutive patients with SCA who had TCD screening were identified after IRB review granted a waiver of consent. Retrospective chart review was then performed to collect the steady state ARC between 2 and 6 months of age; patients were excluded if this value was not available in the medical record. Steady state was defined as a sample collected at least 30 days from an acute illness and at least 60 days since the patient received a blood transfusion. ARC and hematologic data were analyzed using Cox regression analysis to determine the relationship between ARC levels and time to cdTCD/abTCD. TCDs were considered normal (nlTCD) if the Time Average Mean Maximum Velocity (TAMMV) in the middle cerebral artery or distal internal carotid artery was

Description: Highly-conserved LIN28 proteins regulate certain developmentally-timed events in multicellular organisms by decreasing the levels of let-7 miRNAs. It was recently reported that increased expression of LIN28 proteins or decreased expression of the target let-7 miRNAs in human erythroblasts cultured from healthy adult humans causes increased fetal hemoglobin expression. Here LIN28A expression in sickle cell donors’ cells was explored for its potential to regulate fetal and sickle hemoglobin, and to affect the morphological sickling of the mature erythrocytes. After obtaining consent and assent, CD34(+) cells from five pediatric research subjects with HbSS genotype (ages 9-16 years old) were harvested from discarded whole blood following partial manual exchange transfusions. Transgenic expression of LIN28A was accomplished using lentiviral transduction of human CD34(+) sickle cells cultivated ex vivo in serum-free medium for a total of 21 days. On culture day 14, LIN28A over-expression (LIN28A-OE) was confirmed by Q-RT-PCR (control: 8.6E+00 ± 8.1E+00 copies/ng, LIN28A-OE: 2.3E+05 ± 2.1E+05 copies/ng) and Western blot analyses. Erythroblast differentiation and terminal maturation were not affected by LIN28A-OE. Enucleation, as assessed by thiazole orange (TO) staining, was equivalent between the LIN28A-OE cells and control transductions (LIN28A-OE enucleation 40.8 ± 17.0% compared to control 49.9 ± 23.4%, p=0.19). LIN28A-OE strongly suppressed all members of the let-7 family of miRNAs, with average reductions from 66% to 96% for let-7a, let-7b, let-7c, let-7d, let-7e, let-7f-2, let-7g and let-7i. LIN28A-OE caused reduced expression of BCL11A, a known repressor of gamma-globin gene expression. Gamma-, beta (sickle)- and alpha-globin mRNA levels were also investigated by Q-RT-PCR. Gamma-globin mRNA expression levels were significantly increased in LIN28A-OE samples (control: 2.0E+06 ± 7.0E+05 copies/ng, LIN28A-OE: 2.0E+07 ± 6.0E+06 copies/ng, p=0.006), and beta (sickle)-globin mRNA significantly decreased in LIN28A-OE samples (control: 2.0E+07 ± 5.2E+06 copies/ng, LIN28A-OE: 1.6E+07 ± 6.3E+06 copies/ng, p=0.024). Differences in alpha-globin mRNA expression were not statistically significant. Hemoglobin chromatography (HPLC) demonstrated that LIN28A-OE significantly increased the proportion of fetal hemoglobin (HbF control: 10.8 ± 7.1%, LIN28A-OE: 40.1 ± 14.0%; p=0.003) that was balanced by a significant decrease in the proportion of sickle hemoglobin. HbA was not detected. For investigation of the sickling phenotype, enucleated [TO(-)] sickle erythrocytes from LIN28A-OE and control transductions of two subjects’ cells were sorted at the end of the culture period into duplicate tissue culture wells. The sorted erythrocytes were incubated in hypoxia (2% oxygen) for 16 hours, and imaged using inverted microscopy within three minutes after removal from the hypoxia incubator. Four random microscopic field images from each well were acquired. Blinded observers then scored the images from the control and LIN28A-OE transductions according to non-sickled versus sickled morphologies. Cultured erythrocytes from the control transductions demonstrated 86.3 ± 9.5% with sickled morphologies. By comparison, a significant reduction in sickling morphology was observed in the LIN28A-OE cells (56.2 ± 23.1% sickled morphologies; p=0.000009). These results demonstrate that transgenic expression of LIN28A during ex vivo erythropoiesis causes increased gamma-globin gene and protein expression balanced with decreased beta (sickle)-globin at levels that are sufficient to ameliorate hypoxia-related sickling of mature erythrocytes. Disclosures: No relevant conflicts of interest to declare.

Description: Chronic red blood cell transfusion therapy is indicated for primary and secondary stroke prevention in children with sickle cell disease (SCD). The main transfusion goal is achievement of pre-transfusion sickle hemoglobin (HbS) levels of 30%. Unfortunately, there continues to be a population of patients with a history of stroke who have progressive vasculopathy and/or secondary stroke, despite chronic transfusion therapy. Predictive markers for vasculopathy and cerebral events are needed to identify patients at risk for disease progression. Increased reticulocytosis was previously associated with other sickle cell disease complications, and adherent reticulocytes may contribute to the vascular pathology. The objective of this study was to explore the hypothesis that pre-transfusion reticulocytosis may serve as a disease severity marker for cerebral vasculopathy among chronically transfused children with sickle cell disease. After obtaining consent and assent, reticulocytosis was studied in a cohort of pediatric sickle cell patients treated with chronic transfusions (n=33, ages 2-17 years). The group was stratified into three groups: group 1 with an abnormal transcranial doppler (TCD) study in the absence of magnetic resonance angiography (MRA) detected vasculopathy [TCD(+), MRA(-), n=14], group 2 with a history of a stroke in the absence of MRA-detected vasculopathy [Stroke(+), MRA(-), n=5], and group 3 with a history of abnormal TCD or stroke and more severe vasculopathy detected by magnetic resonance angiography [MRA(+), n=14]. Pre-transfusion blood samples were analyzed within 72 hours of collection. Steady-state blood samples were also examined from a control group of pediatric SCD patients (〉6 years of age) with normal TCD studies who were receiving supportive care in the absence of chronic transfusions or hydroxyurea (n=7). Hematologic data, including automated complete blood counts, absolute reticulocyte counts (ARC) with reticulocyte maturity were obtained. In addition, a flow cytometric approach was developed to further examine and quantitate reticulocyte subsets based upon staining with thiazole orange combined with CD36, CD45, CD49d, CD71, and CD235. The pre-transfusion HbS levels were not statistically different among the three transfused groups ([TCD(+), MRA(-)]: 30.2 ± 11.8%; [Stroke(+), MRA(-)]: 28.4 ± 3.3%; [MRA(+)]: 33.3 ± 9%, p〉0.3). The high levels of reticulocytosis in the pre-transfusion samples were similar to those measured in the control group (ARC: 451 ± 126 K/uL in the chronically transfused cohort; ARC: 369 ± 94 K/uL in the control group, p=0.11). Pre-transfusion reticulocytosis was detected in every chronically transfused subject (ARC range 151-701 K/ul). The mean ARC in the [TCD(+), MRA(-)] group was not significantly different from the [Stroke(+), MRA(-)] group (411 ± 135 K/uL and 396 ± 97 K/uL respectively, p=0.82). However, the mean ARC in the [MRA(+)] group (512 ± 107 K/uL) was significantly higher than the control group, the [TCD(+), MRA(-)] group and the [Stroke(+), MRA(-)] group (p

Description: Background Iron overload is a recognized consequence of chronic transfusion therapy in patients with sickle cell disease (SCD), but most of the focus to date has been on the effects of increased liver iron concentration (LIC) with increasing transfusion burden. Even though there is a robust body of literature concerning cardiac iron overload (CIO) in patients with thalassemia major, there remains a paucity of data in how to detect and treat CIO in patients with SCD, particularly in the pediatric and young adult population. While CIO is seen less commonly in sickle cell disease than in thalassemia, patients with SCD remain at risk, with recent studies demonstrating an incidence of 2-5% of CIO in chronically transfused patients with SCD. We performed a retrospective chart review of patients with cardiac MRIs (cMRIs) and LICs by Ferriscan performed at our institution to identify risk factors for CIO, as well as to characterize institutional practice for assessing cardiac iron in the absence of defined practice guidelines. Methods We reviewed clinical characteristics of all patients with SCD who had cMRIs performed at Children's Healthcare of Atlanta between June 2012 and December 2017. We then queried our institutional sickle cell database for patients who were at least 3 years old in 2010, genotype SS or S Beta zero thalassemia, were on chronic transfusions for at least 5 years by 2017, and had not undergone a cMRI. Patients who were status post bone marrow transplant were excluded. For comparison of age, average ferritin, and transfusion duration, significance among means between patients with and without CIO was calculated using a two-tailed unpaired t-test. For comparison of LIC, significance among medians was calculated using the Mann Whitney test. A p value of 43 mg/g in those with CIO vs 34 mg/g in those without CIO, although this was not statistically significant (Figure 1). Interestingly, CIO was seen as young as 7 years of age and after as little as 22 months of chronic transfusions, and with concurrent LIC values as low as 8.1 mg/g. Of the 11 patients with CIO, 6 had follow-up cMRI data available, and all 6 had normalization of cardiac iron (T2* 〉 20ms) on subsequent MRIs (Figure 2 and Table 2). There was 1 patient who did not have full transfusion and chelation history available for analysis. Of the remaining 5, 5/5 had increased or more aggressive chelation added, including 2 who were started on high-dose IV Desferal every 2 weeks; 3/5 also had partial manual exchange (PME) added to their chronic transfusion regimens. There were 80 patients who were on chronic transfusions but did not have a cMRI performed; as a group, they had a median LIC of 17 mg/g (range: 1.7 - 〉43 mg/g), an average 1-year ferritin of 3641 ng/mL (range: 520 - 8478 ng/mL), and had been on chronic transfusions for a mean of 87 months at time of Ferriscan study (range: 14 - 192 months). Overall, these patients had a lower transfusion burden than those who received cMRIs, but there were several in this group who had significant iron overload, including 10 who had LIC values of 〉 43mg/g. Conclusion CIO in SCD may be a more salient issue, and occur earlier, than previously described. We did not find a strong relationship between CIO and ferritin levels or LIC by Ferriscan, but we did find that CIO was reversible with more aggressive chelation or the addition of PME. While guidelines for monitoring for CIO in SCD are largely extrapolated from thalassemia data, the rate and physiology of iron loading may be completely different. Due to a paucity of information in this area, more studies are needed to guide screening and to fully assess risk factors that may put certain individuals more at risk for cardiac iron loading. Disclosures No relevant conflicts of interest to declare.

Description: Chronic transfusion therapy (CTT) for sickle cell anemia (SCA) reduces the risk of sickle complications by diluting hemoglobin S (HbS)-containing red blood cells (RBCs) with HbA RBCs and suppressing sickle erythropoiesis. HbS level is influenced by both endogenous hemolysis and clearance of transfused RBC. Minor RBC antigen (Ag) mismatches may result in alloimmunization and hemolytic reactions, but it is unknown if Ag mismatches independently influence HbA clearance. This study sought to: (1) determine the frequency of RBC minor Ag mismatches that occur in chronically transfused SCA patients receiving phenotype-matched transfusions; (2) determine the rate of clearance of transfused HbA following each transfusion; (3) explore associations of HbA clearance with RBC minor Ag mismatches. Children with SCA ages 3-20 years on CTT by either simple transfusions or partial manual exchange (PME) were enrolled in a prospective observational study. All RBC units were serologically matched for C/c, E/e, K (category 1); additional matching for Fya, Jkb, and Ag-negative for putative antibodies (category 2) was provided to patients with ≥1 clinically significant alloantibody or (in some cases) warm autoantibodies. RBC genotyping of SCA patients was performed using PreciseType™ Human Erythrocyte Antigen (HEA), RHCE, and RHD BeadChip assays (Immucor, Norcross, GA). Genomic DNA was extracted from unit segments of all RBC transfused over 12 months and tested using the prototype HEA Leukoreduced (HEA-LR) BeadChip assay to detect the same profile of RBC Ag. RH variant testing of units was not performed. Pre-transfusion HbA was recorded for all episodes, where HbA (g/dL) = Hb (g/dL) x HbA%. Hb electrophoresis was drawn 15-30 minutes post-transfusion in a subset of episodes. For other episodes, post-transfusion HbA was estimated based on the volume of transfusion (VolT), estimated unit hematocrit (Hct), phlebotomy volume (VolPh, if applicable) and total blood volume (TBV), using the following equation: Post HbA (g/dL) = Pre HbA (g/dL) + [(unit Hct) x VolT / 3 x Wt] - (Pre HbA%/100) x [(Pre Hb (g/dL) x VolPh / 3 x TBV]. HbA loss/day was calculated from the Post HbA to the next pre-transfusion HbA. There were 82 patients (54 category 1, 28 category 2) who received 2123 units in 1014 transfusion episodes; HEA-LR genotypes were obtained on 1827 units and 828 complete transfusion episodes. Altered RHC/c or e genotypes were present in 29 (36%), and partial D+ genotypes in 4 (5%). There were 46 historical alloantibodies, the majority against Rh (C, E, hrB, Goa, f, V, VS) and Kell (K, Kpa, Jsa). During the study, 5 patients developed 6 new alloantibodies: 3 Jsa, 1 Goa, 1 Wra, 1 AUS (category 2 transfusions alloantibody incidence 0.64/100 units; category 1 transfusions 0.074/100 units). The mean genotype-predicted Ag mismatches per transfusion was 3.52 for category 1 and 2.78 for category 2 (despite category 2 being matched for ≥2 Ags beyond category 1 matching). Ag mismatch frequency was different in category 1 vs. 2 transfusions for: C (3.8% vs. 0%, p=0.0009), e (24.5% vs. 36.8%, p=0.0002), and VS (21.2% vs. 13.6%, p=0.008). Overall concordance between HEA-LR and available serologic testing was 98%. Of the 169 transfusion episodes with post-transfusion HbA measurements (103 category 1, 66 category 2), the mean HbA loss was 2.38 g/dL/28 days for category 1 vs. 2.73 g/dL/28 days for category 2 (p=0.0047). Comparison of category 1 vs. 2 transfusions (see table) shows shorter transfusion interval and higher frequency of PME in category 2 transfusions. Estimated post HbA had a bias of -0.31 g/dL (95% C.I.-0.18, -0.46) and precision of 0.85 g/dL (95% C.I. 0.61, 1.18). Comparison of estimated HbA loss to Ag mismatches was limited to 533 transfusion episodes in which unit Hct could be estimated within

Description: Introduction Unlike sickle cell anemia (HbSS disease), little research has been devoted specifically to Hemoglobin SC (HbSC) disease. Though distinct, the disease is often treated simply as a milder form of HbSS disease. As a result, its associated manifestations are less well defined which leads to mismanagement of these patients in the medical community. The aim of this study was to further define HbSC disease and its manifestations, as well as evaluate the burden of chronic opiate use and healthcare utilization in this population. Design and Methods A retrospective chart review was performed utilizing patients from one of the largest sickle cell clinics in North America. Patients with hemoglobin electrophoresis confirmed HbSC disease that had been to the center between 2010 and 2016 were identified. Clinical data relating to baseline laboratory values, HbSC disease-related complications and acute care visits was then abstracted from patient's medical records and analyzed to further define this population. IRB approval was granted for this retrospective review. Results A total of 210 patients were included in the study. The cohort had a median age of 35 years (mean of 37.2 years, range of 18 to 81 years) and 57.1% were female while 42.9% were male. Steady state laboratory values included an average hemoglobin (HGB) concentration of 11.67 mg/dL. The incidence of chronic complications including chronic opiate use, avascular necrosis (AVN), retinopathy and pulmonary hypertension were 34.76%, 23.33 %, 65.22% (in those with a documented fundoscopic exam) and 2.38% respectively. The incidence of acute complications during the study period including venous thromboembolism, stroke, and acute chest syndrome were 7.62%, 0% and 1.9% respectively. A total of 183 patients (87.1%) had an acute care visit for vaso-occlusive pain crisis during the study period of 6 years (average of 3.51 visits/patient/year) while 113 patients (53.8%) required hospital admission for a vaso-occlusive pain crisis (average of 0.43 admissions/patient/year with an average length of stay of 3.91 nights). Subgroup analysis comparing those with and without chronic opiate use (more than 3 months chronic pain) revealed similar baseline demographics with median ages of 38 and 33 (average age of 38.5 vs. 36.5, p-value 0.31), respectively. The two groups were also similar in steady state laboratory values (Hemoglobin concentration of 11.69 mg/dL versus 11.65 mg/dL, p-value 0.86) and incidence of AVN (28.77% versus 20.44%, p-value 0.17). However, chronic opiate users did have a significantly higher number of acute care unit visits per year (average of 6.15 versus 1.35, p-value 0.0001) and hospital admissions per year (0.84 versus 0.22, p-value of 0.0001), with chronic opiate users (34.8% of the study population) accounting for 74% of all acute care visits and 71.2% of all hospital admissions. Conclusions This cohort helps to further define HbSC disease as a distinct entity from HbSS disease and demonstrates that, though often thought of as mild, this disease places a significant burden on the healthcare system. As expected, the degree of anemia was quite mild compared to that seen in HbSS disease and there were significantly less episodes of stroke and acute chest. However, the incidence of AVN was similar to that seen in HbSS disease and in those with a documented fundoscopic exam the degree of retinopathy was higher, demonstrating one of the characteristics that makes this disease unique. Subgroup analysis of those with chronic opiate use reveals a population with a propensity toward high health care utilization. These distinctions help to identify risks of morbid complications in this population as well as areas to focus efforts in screening and prevention. Disclosures No relevant conflicts of interest to declare.

Description: Abstract 1449 Granulocyte colony-stimulating factor (G-CSF) in combination with plerixafor (AMD3100) produces significant mobilization of peripheral blood stem cells in the rhesus macaque model. The CD34+ cell population mobilized possesses a unique gene expression profile, suggesting a different proportion of progenitor/stem cells. To evaluate whether these CD34+ cells can stably reconstitute blood cells, we performed hematopoietic stem cell transplantation using G-CSF and plerixafor-mobilized rhesus CD34+ cells that were transduced with chimeric HIV1-based lentiviral vector including the SIV-capsid (χHIV vector). In our experiments, G-CSF and plerixafor mobilization (N=3) yielded a 2-fold higher CD34+ cell number, compared to that observed for G-CSF and stem cell factor (SCF) combination (N=5) (8.6 ± 1.8 × 107 vs. 3.6 ± 0.5 × 107, p

Description: Background Red blood cell (RBC) alloimmunization occurs at a much higher rate in patients with sickle cell disease (SCD) compared to other multiply transfused populations. Reasons for this include: altered immunologic responses, frequent transfusions during times of elevated inflammatory states, and disparate donor/recipient RBC antigens. Providing Rh and Kell matched RBCs has been shown to decrease but not eliminate RBC alloimmunization in patients with SCD. Although it has been shown in murine models that recipient inflammatory state at time of transfusion has the ability to regulate alloimmunization, direct clinical evidence for this effect is lacking in SCD patients. Methods With IRB approval, medical records of alloimmunized pediatric SCD patients were retrospectively reviewed to determine the influence of SCD-related complications often considered to be pro-inflammatory, at time of transfusion on RBC alloimmunization. The degree of antigen matching, age, and additive solution of each RBC unit were also assessed. Potential pro-inflammatory states were classified as: acute chest syndrome (ACS), acute stroke, acute febrile illness in the absence of another sickle-co morbidity, splenic sequestration, aplastic crisis, priapism, VOC with and without leukocytosis (WBC count ≥ 20k/μL) and elective surgery. Initial alloantibody detection dates were ascertained from blood bank records. Clinical events at time of transfusion were compared between transfusions resulting in a new alloantibody (AlloAb) and transfusions that did not. Univariable analysis was performed using Wilcoxon rank sum and Pearson’s Chi square test. Logistic regression modeling was used to estimate adjusted effects of stated variables on alloimmunization. Results A total of 3166 prestorage leukoreduced RBC transfusions (mean age 19.3 days) were provided to 52 SCD patients; 128 transfusions resulted in RBC alloantibodies; 3038 transfusions did not. On univariable analysis, 14.1% of transfusions during any inflammatory event resulted in a new AlloAb compared to 1.4% of transfusions in the absence of inflammation (p

Description: Background A higher incidence of Red Blood Cell (RBC) alloimmunization exists in Sickle Cell Disease (SCD) than in any other multiply transfused population. The majority of the RBC alloantibodies are to Rh (D, C, c, E, e) and K antigens. Transfusing Rh and Kell matched RBCs substantially decreases alloimmunization rates in SCD patients; however clinically significant Rh antibodies with apparent common specificities persist as a result of altered RH alleles in SCD patients. Methods SCD patients with a history of ≥15 transfusions or having RBC alloantibody(ies) were consented and asked to complete an ethnicity survey defining patients as “African” (patient or both parents African-born), African American (parents and patient US-born), or other. RH genotyping was performed on all patients using RH Variant Beadchips (BioArray, Warren NJ). Medical records of patients were retrospectively reviewed and compared to RH genotype and ethnicity to determine the clinical impact of RH variants on alloimmunization. Fisher’s Exact test was used to determine statistical significance of correlations. Results Among 117 SCD patients genotyped, 67 (57.3%) had alloantibodies, with a median of 50 transfusion exposures. RHCE variant haplotype frequencies for (C)ces, ces, ceAR, ceMO and ces(340) were 6.8%, 20.1%, 0.9%, 1.3% and 0.4%, respectively. Twenty-two patients were either homozygous (7), compound-heterozygous (5), or heterozygous for these RHCE variant haplotypes with a conventional RH E allele in-trans (10). Of these, approximately 32% (7/22) formed an anti-e alloantibodies after a median of 6 Rhe+ RBC transfusion exposures compared to 7.3% (7/95) of all other patients (p=0.0048). No anti-e alloantibodies were detected in 15/22 patients within the RHCE variant subgroup after 1436 Rhe+ RBC transfusions (median 66 transfusions/patient), yielding an anti-e alloantibody frequency of 0.45/100 units. Fifty percent (11/22) of patient in the RHCE variant subgroup formed an autoantibody, compared with 24% (19/78) of all other patients (p=0.0345). Approximately 32% (8/25) of the “African” patients were homozygous or compound heterozygous for a variant, as opposed to 10.7% of “Non-African” patients (p=0.0312); only 12.5% (1/8) of African patients with RHCE variant subgroup formed anti-e alloantibodies versus 46% of “Non-African” patients this subgroup (p= NS). Conclusion SCD patients with RHCE variant haplotypes are at increased risk for the formation of clinically significant anti-e alloantibodies, which may be inaccurately identified as autoantibodies in the absence of RH genotyping. This implies that RH genotyping should be either incorporated into the standard RBC phenotype evaluation in all SCD patients, or at least into the evaluation of any SCD patient with an autoantibody that demonstrates “e” specificity. We report similar RHCE variant allele frequencies compared to previously published SCD population studies; however we found a higher prevalence of homozygous and compound heterozygous RHCE variant genotypes in SCD patients less then two generations removed from African immigration. Confounding factors for anti-e alloimmunization risk in SCD patients with RHCE variant genotypes other than antigen disparity exist which may explain why “African” patients within the RHCE variant subgroup demonstrated lower anti-e alloimmunization compared to other patients in this group. Further study is warranted to further characterize the immunogenic potential of high incidence Rh antigens in individuals with RH variants, and the immunogenetic variables that affect alloimmunization overall. Disclosures: Fasano: ApoPharma: Honoraria.

Description: Abstract 1109 Background: Platelet refractoriness as a consequence of HLA alloimmunization complicates allogeneic hematopoietic stem cell transplantation (HSCT). Little is known regarding the affects of reduced intensity HSCT on the incidence and duration of HLA alloimmune-mediated platelet refractoriness following reduced intensity HSCT. Methods: We retrospectively studied HLA alloimmunization in 16 patients with malignant and non-malignant hematological disorders, who underwent an allogeneic HSCT from an HLA matched relative using reduced intensity, non-myeloablative conditioning. All patients received a G-CSF mobilized, lymphocyte replete, peripheral blood stem cell (PBSC) allograft following cyclophosphamide (120 mg/kg) and fludarabine (125 mg/m2) based conditioning (+/− equine ATG 40 mg/kg × 4 days) with cyclosporine (CSA; beginning on day −4) and IV methotrexate (5 mg/m2 × 3 days) given as GVHD prophylaxis. Eight patients known to be HLA alloimmunized pre-transplant were compared to 8 control patients who were HLA alloantibody negative pre-transplant. Stored patients' serum samples from pre-transplant (day -30) and post-transplant intervals (days +30, +60, +100, +180, and ≥ 365) were analyzed for the presence of IgG antibodies to HLA class I antigens using a membrane-independent solid phase assay involving color-coded microbeads coated with HLA antigens and analysis with a flow analyzer (LABScan 100; One Lambda, Canoga Park, CA). Panel reactive antibody (PRA) and mean fluorescent intensity (MFI) were analyzed for all samples to measure HLA antibody strength and specificity, respectively. HLA alloantibodies were analyzed and compared with the degree of donor myeloid and T-cell engraftment measured on post-transplant blood samples by PCR of short tandem repeats (STRs). Results: Among the 8 alloimmunized patients who required HLA matched platelets pre-transplant, the median time until HLA antibodies disappeared was 100 days post HSCT. Remarkably, among these patients, 3/8 (37%) had HLA class I antibodies detectable for more than 100 days post-transplant including one patient who continued to show high level alloreactivity greater than 1 year after HSCT. Among the 8 control patients who tested negative for HLA alloantibodies pre-transplant, 3 acquired HLA alloantibodies after HSCT which were first detected at day +30 in all 3 cases. In 2 of the 3 cases, the donors for these patients were found to have pre-existing HLA antibodies of equivalent specificity to those found in the patient post-transplant suggesting patient acquisition of HLA antibodies was mediated by passenger donor lymphocytes that were transplanted in the allograft. Overall, HLA antibodies were detectable for more than 100 days after transplantation in 6/16 (37%) patients analyzed with 3/16 (18%) having alloantibodies detectable for more than 1 year post-transplant despite chimerism being 100% donor in myeloid and T cell lineages. Conclusions: Prolonged production of HLA alloantibodies leading to platelet refractoriness can occur following reduced intensity allogeneic HSCT and may persist even after full donor myeloid and T-cell chimerism have been achieved. Remarkably, we observed that transplantation of passenger donor lymphocytes can result in de novo HLA alloimmunization, complicating post-transplant transfusion management. Screening patients and their donors for HLA antibodies before HSCT would identify the majority of subjects at risk for alloimmune-mediated platelet refractoriness after transplantation. This new screening strategy would not only assist in transfusion support after allogeneic HSCT but could also play a role in the decision analysis for selecting optimal stem cell donors. Disclosures: No relevant conflicts of interest to declare.