(P-QU-4) A data-driven approach to identify an optimal mean fluorescence intensity cutoff for platelet component transfusion in highly HLA-sensitized patients

Procuring compatible platelet components for highly human leukocyte antigen (HLA)-alloimmunized patients can be difficult. Prior studies report permissive mismatching (PERMM) of ≥1 donor specific antibodies (DSA) to increase the number of possibly compatible donors, but mean fluorescence intensity (MFI) cut-offs for clinically significant DSA vary between institutions. An academic transfusion service evaluated the impact of expanding the MFI threshold from 4,000 up to 8,000 to increase the number of donors that can be recruited for highly sensitized patients. The main outcome was adequacy of corrected count increment (CCI).

Study

Design/Methods:

A retrospective chart review was performed on six HLA-alloimmunized patients with calculated panel reactive antibodies (cPRA) from 45% to 100% who were transfused ≥1 HLA compatible leukoreduced, irradiated, and apheresis platelet components between January 1 and May 4, 2024, as inpatients. CCIs were calculated as described by Cohn ASH 2020 for post-transfusion platelet counts collected within two hours of cessation of transfusion with an acceptable CCI threshold of ≥5000. The MFI of each HLA class 1 A and B antigen and the corresponding CCI values from each transfusion were correlated to generate the probability of a MFI value producing an adequate CCI. The probabilities at each threshold were used to fit a locally estimated scatterplot smoothing (LOESS) curve and predict the probability of an inadequate CCI according to MFI.

Results/Findings:

Six patients met inclusion criteria out of 494 patients transfused 2,660 platelet components. Of those, the six patients received 96 HLA-selected platelet components out of 338 total HLA-selected components. Most of the HLA-selected platelet donors were recruited from a local, dedicated FDA-registered donor center.

Both maximum MFI and time until measurement were significantly associated with CCI (P < 0.001 and P < 0.01, respectively, Wald Chi-Squared Test for fixed effects in Linear Mixed Effects Model). The lower the maximum MFI and the shorter the amount of time until the platelet count was measured resulted in higher CCIs. The empirically predicted probability of observing an inadequate CCI increased with the maximum MFI against antigens expressed by the donor of the platelet unit. This probability crossed 50% at an MFI of 4,821 against HLA class 1 A and B antigens.

Conclusions:

This analysis suggests an MFI threshold of approximately 5,000 may result in an acceptable CCI response going forward at our institution. Implementation of this new cutoff has the potential to increase the number of potential donors within the local donor database and decrease the number of HLA-selected platelets purchased from outside suppliers.