Risk Factors for Thrombotic Microangiopathy after Hematopoietic Stem Cell Transplantation By Disease Groups in Competing Risk Analysis

Background: Allogeneic hematopoietic stem cell transplantation (allo-HSCT) associated thrombotic microangiopathy (TA-TMA) is a recognized risk factor for mortality after allo-HSCT. In an observational study (Epperla et al, BJH. 2020), researchers identified risk factors (RF) for developing TA-TMA in competing risk analysis. Another study (Schoettler et al, BJH. 2024), reported concordant and discordant RF for TA-TMA, between children and adults. Recognizing significant differences in incidence of TA-TMA and mortality post-HSCT between various underlying diseases, we evaluated RF for TA-TMA by underlying diseases, with death for all causes as the competing risk.

Methods: This secondary analysis of an existing CIBMTR dataset, included all patients with hematological diseases who underwent allo-HSCT between 2008-2016. In the absence of any specific diagnostic criteria during this timeframe, TA-TMA was identified by presence or absence of the condition as judged by the treating physicians. Death was considered a competing risk in the Fine and Gray sub-distribution Cox proportional hazard model. Acute graft-versus-host-disease (aGVHD), a recognized time-dependent risk factor for TA-TMA cannot be incorporated as a time-dependent covariate in the Fine-Gray model. To be evaluated as a RF in competing risk analysis, aGVHD grade 2-4 cases were only included as RF if they occurred in patients before TA-TMA. Sub-distribution Cox multivariable regression models were used to calculate hazard ratios for developing TA-TMA, treating death as a competing risk.

Results: There were significant differences in the mortality rates between the malignant and non-malignant disease states. Significant differences in mortality were also seen within the malignant and non-malignant groupings themselves. While there were no significant differences in the incidence of TA-TMA between the malignant and non-malignant disease states, there were significant differences between diseases within the malignant grouping.

For malignant diseases, we found that aGVHD (grade 2-4), female sex, prior autologous transplant, GVHD prophylaxis (sirolimus + calcineurin inhibitor), and donor type (matched unrelated donor, related donor not including HLA-identical sibling, partially matched unrelated donor, cord blood donor with ≤4/6 HLA matched) had increased risk of developing TA-TMA. Factors associated with decreased risk of developing TA-TMA included reduced-intensity conditioning/non-myeloablative conditioning (RIC/NMA) and treatment with anti-thymocyte globulin when applicable regardless of alemtuzumab treatment.

For acute myelogenous leukemia (AML), adjusting for all significant variables, only acute GVHD (grade 2-4), female sex, hematopoietic cell transplantation comorbidity index (HCT-CI) ≥3 morbidities, African American race, other races (not including African American and Caucasian) had increased risk of developing TA-TMA. All acute GVHD prophylaxis except for sirolimus (without CNI or post-transplant cyclophosphamide (pt-cy)) regimen had significant lower risk of developing TA-TMA.

For acute lymphoblastic leukemia (ALL), adjusting for all significant variables, only acute GVHD (grade 3-4) and African American race had increased risk of developing TA-TMA. All acute GVHD prophylaxis except for sirolimus (without CNI or post-transplant cyclophosphamide (pt-cy)) regimen had significant lower risk of developing TA-TMA.

For non-malignant diseases, adjusting for all significant variables, only acute GVHD (grade 3-4) and donor type with mismatched unrelated donor had increased risk of developing TA-TMA. No variables were associated with a lower risk of developing TA-TMA.

Discussion: Competing risk analysis by malignant, non-malignant disease groups, disease groups - AML and ALL, add to the previous CIBMTR reports that identified significant risk factors for developing TA-TMA, by identifying the concordant and discordant significant RF. This information is important as it allows providers to be more aware of the significant RF for developing this rare transplant complication for patients with different underlying diseases.