Impact of Timing of Acute Graft-Versus-Host-Disease on Incidence of Thrombotic Microangiopathy after Hematopoietic Stem Cell Transplantation

Background: Risk factors associated with allogeneic hematopoietic stem cell transplantation (allo-HSCT) associated thrombotic microangiopathy (TA-TMA) were reported from Center for International Blood and Marrow Transplant Research (CIBMTR) (Epperla et al, BJH. 2020). Incidence of acute graft-versus-host-disease (aGVHD), grades 2 and 3-4 were significant risk factors for developing TA-TMA. Timing of aGVHD was not considered in the evaluation. Fifty-four percent of the TA-TMA cases were preceded by aGVHD grades 2-4. In this study, we evaluated the impact of timing of aGVHD grades 2-4, along with other risk factors on the incidence of TA-TMA, using a dynamic landmark approach.

Methods: This retrospective analysis of an existing CIBMTR dataset, included all patients with hematological diseases who underwent allo-HSCT between 2008-2016. In this landmark study, we selected 8 landmark time points at month 1, 2, 3, 4, 5, 6, 7 and 8. Death before TA-TMA was considered a competing event to TA-TMA. A dynamic landmark analysis was performed by fitting the marginal Fine-Gray model, stratified by malignant and non-malignant diseases. Incidence of acute aGVHD grade 2, grade 3-4, and death-free patients were considered at each landmark time point. Age, pre-HSCT renal health (decreased kidney function defined as GFR <60 mL/min for adults and <90 mL/min for children vs normal kidney function), donor type, prior autologous transplant, Karnofsky performance score, sex, race, donor-recipient cytomegalovirus status, conditioning regimen, anti-thymocyte globulin alone/alemtuzumab alone, HSCT-comorbidity index, graft type and aGVHD prophylaxis were evaluated as fixed covariates in the analysis. Interactions between landmark time points, aGVHD, and age groups were evaluated.

Results: This study included 23,090 patients, with 2.8% incidence of TA-TMA, 22.3% incidence of aGVHD grade 2, 16.6% incidence of aGVHD grade 3-4 and 47.0% deaths for all causes. Eighty percent of the incidence of TA-TMA occurred during first eight months following allo-HSCT, fifty-five percent of such cases were preceded by aGVHD grades 2-4.

Dynamic landmark analysis, adjusting for all significant risk factors, provided a negative estimate of landmark timepoints, -0.116 (p<0.001) indicating that patients who survived without TA-TMA to later time points had a lower subsequent incidence of TA-TMA. The interaction effects of an aGVHD grade 2 or grade 3-4 with landmark time points were significant for incidence of TA-TMA, with p-values 0.007 and 0.035, respectively. The negative signs of interactions indicated that the differences in the hazard ratios between the aGVHD and none or Grade 1 aGVHD is reduced as landmark time points move forward from Month 1 to Month 8. In univariate analysis with aGVHD as main effect, age categorized as <50 and ≥ 50 years was a significant risk factor (HR 1.28, 95% CI 1.41, 1.63) for TA-TMA. While interaction aGVHD grade 2 with age groups was not significant, there was significant interaction between aGVHD grade 3-4 and age groups. Consequently, effects of aGVHD grades 2 and 3-4 were evaluated for patients <50 and ≥50 years of age at each of the landmark time points.

Risk (HR, 95% CI) for TA-TMA in patients <50 years, with aGVHD grade 2 compared to no aGVHD or grade 1 aGVHD, declined from 2.35 (1.89, 2.92) to 1.40 (1.06, 1.84) and in patients ≥50 years, with aGVHD grade 2 compared to no aGVHD or grade 1 aGVHD, declined from 2.12 (1.71, 2.64) to 1.26 (0.95, 1.68) over first 8 months post-HSCT. HR (95% CI) for TA-TMA in patients <50 years, with aGVHD grade 3-4 declined from 4.85 (4.00, 5.88) to 3.29 (2.52, 4.29) and in patients ≥50 years, with aGVHD grade 2, declined from 2.72 (2.16, 3.42) to 1.84 (1.37, 2.48) compared to no aGVHD or grade 1 aGVHD over first 8 months post-HSCT.

Discussion: This landmark study results add to the previous CIBMTR reports that not only acute GVHD, but also its timing and age of patients have significant effect on development of TA-TMA. Age has confounding effect with acute GVHD grade 3-4, with patients <50 years at a significantly higher risk for developing TA-TMA as compared to patients ≥50 years. These results are important in informing clinical practice for transplant patients, suggesting that patients who do not develop early TA-TMA or aGVHD are less likely to develop these complications the further from HSCT.