Testing for Platelet Refractoriness: Optimizing Testing Algorithms

Abstract Platelet refractoriness occurs when recipients fail to achieve an appropriate platelet count increment following a platelet transfusion. The evaluation of platelet refractoriness assesses nonimmune and immune-mediated etiologies of platelet clearance. Chief among the immune-mediated causes are human leukocyte antigen (HLA) alloantibodies, which form in response to foreign HLA proteins and account for 30% to 40% of these cases. Therefore, testing for HLA antibodies is crucial in the workup of these patients. Multiple platforms have been developed to detect HLA antibodies, including complement-dependent lymphocytoxicity assays, enzyme-linked immunosorbent assays (ELISAs), and microparticle bead-based assays using multiplex flow cytometry. Many centers employ a qualitative screening assay to detect HLA antibodies. However, the sensitivity of these various assays is not uniform. Ideally, a screening assay should have high sensitivity to limit false-negative reactions. In this study, two screening assays were compared, namely, the ELISA platelet antibody screen used in the blood bank and the FlowPRA screen used by the transplant laboratory to detect HLA antibodies. Data from January 2010 to December 2017 were retrospectively reviewed. Only ELISA indirect platelet antibody screens (LIFECODES PakPlus; Immucor, Norcross, GA) that had an accompanying flow cytometry screen (FlowPRA; One Lambda, Canoga Park, CA) were included. Results of the two assays were compared and sensitivity and specificity were calculated. The average optical densities (ODs) of the ELISA screen were also compared to the percent reactivity (PRA) of the flow cytometry screen. Within the study period, 348 ELISAs were performed, of which 63 (18%) had corresponding flow cytometry screens. The flow cytometry screen was concordant (positive ELISA, positive flow) with the ELISA in all cases, but the ELISA was discordant (negative ELISA, positive flow) to the flow screen in 13 cases. These false-negative cases represented 21% of the total study cohort and 37% of the ELISA-negative cases. Though the ELISA had 100% specificity, the sensitivity was only 68% with a negative predictive value of 63%. The degree of discordance between false-negative ELISAs and corresponding PRA values was higher than expected, with PRA values ranging from 17% to 99% (average of 45%). OD values appeared to directly correlate with PRA but did not differ significantly between true-negative and false-negative ELISA scre