Prospective Evaluation of Immune Hemolysis in Minor-ABO Incompatible Reduced-Intensity Peripheral Blood (PB) Hematopoietic Cell Transplantation (HCT).

Background: Life-threatening immune hemolysis may complicate minor-ABO incompatible HCT. Strategies for laboratory monitoring and proposed risk factors for hemolysis such as reduced intensity conditioning, non HLA-matched-sibling donors, CsA alone as GVHD prophylaxis, and PB rather than marrow have not previously been evaluated prospectively.

Study Design: Following fatal massive hemolysis in an index patient, we prospectively evaluated 40 consecutive, lymphocyte-replete, ABO-minor incompatible PB HCT following fludarabine/cyclophosamide(CY)-based reduced-intensity conditioning, with daily CBC, LDH, and DAT d5-15, and donor-compatible RBC transfusion to maintain the hgb > 9.5 g/dL. All donors were 6/6 HLA matched relatives who received GCSF 10 ug/d x 5 d as mobilization. GVHD prophylaxis was CsA, either alone (n=15), or combined with mycophenolate mofetil (1 g PO BID, n=18, CsA/MMF) or methotrexate (5 mg/m² IV d 1, 3, 6, n=7, CsA/MTX). Outcomes were compared with a concurrent series of 15 T and B-cell depleted ABO-minor incompatible PB HCT following CY/TBI myeloablation.

Results: The three GVHD prophylaxis groups were similar for age, weight, gender, and CD34, B and T cell dose. Hemolysis (mean peak LDH 2445 vs 365 U/L, 9 vs 3 RBC transfusions, p<0.0001 vs no hemolysis) occurred in 5 cases; 4/15 with CsA alone, 1/18 with CsA/MMF; 3/38 sibling donor, 2/2 non-sibling donor transplants (1 parent→child with CsA/MMF, 1 child→parent with CsA alone); all 4 cases after the index case survived. Host-type RBC disappeared more quickly in cases with hemolysis (median 12 vs 138 d, p<0.001), with non HLA-matched-sibling donors (12 vs 117 d, p=0.017), and with hematologic diseases or pediatric malignancies compared to adult solid tumors (35 vs 198 d, p<0.001). A positive DAT occurred in 4/5 cases prior to hemolysis and in 28/40 cases overall; the DAT had poor positive predictive value (PPV) for hemolysis (15%) and did not impact the rate of disappearance of host-type RBC (p=0.98). Similarly, RBC eluates were positive for the relevant ABO antibody in 3/5 cases with hemolysis, and in 12/23 cases without hemolysis (PPV=20%). Donor isohemagglutinins appeared more rapidly in cases with hemolysis (9 vs 341 d, p<0.001), with non HLA-matched-sibling donors (p=0.003), and with use of CsA alone as GVHD prophylaxis (p=0.03). The allograft B-cell dose (n=22) was inversely related to the time until the DAT became positive (p<0.01) in patients who did not receive MTX, but was not associated with hemolysis. Host-type RBC disappeared more rapidly (19 vs 78 d, p=0.026), and donor-type isohemagglutinins appeared more rapidly (23 vs 434 d, p=0.04) in reduced-intensity, lymphocyte-replete HCT vs myeloablative, lymphocyte-depleted HCT treated with similar GVHD (CsA without MMF or MTX) regimens.

Conclusion: Lymphocyte-replete reduced-intensity PB HCT produce potent anti-host immune effects. Hemolysis in minor-ABO incompatible cases can be severe, and may be increased when using non HLA-matched-sibling donors or CsA alone as GVHD prophylaxis. Serial DAT monitoring has limited utility for predicting hemolysis. Daily CBC testing and aggressive transfusion support are recommended in all cases.