Earlier detection of transplant-associated thrombotic microangiopathy using harmonization criteria: A real-world comparative evaluation with conventional criteria

Abstract

Abstract Aim: Transplant-associated thrombotic microangiopathy (TA-TMA) is a frequently underrecognized complication of allogeneic hematopoietic stem cell transplantation (allo-HSCT), linked to significant mortality. The lack of a specific biomarker has led to the development of various diagnostic criteria based on nonspecific laboratory parameters. Recently, new harmonization criteria for the diagnosis and risk stratification of TA-TMA were proposed, warranting validation, particularly in adult populations. However, independent analyses by Acosta-Medina et al. and Vasu et al. have yielded markedly discordant findings, reporting TA-TMA incidence rates of 6.3% and 56.9%, respectively. We aimed to further validate the criteria with our large, long-term cohort of adults with acute myeloid leukemia (AML). Methods: We screened 1360 patients who underwent first allo-HSCT in Catholic Hematology Hospital between 2012 and 2024 to diagnose TA-TMA according to the harmonization criteria (H-TMA) and Cho's criteria (C-TMA). Patients diagnosed with H-TMA were classified as high-risk (HR-TMA) or standard-risk (SR-TMA) to compare non-relapse mortality (NRM) from the time of diagnosis. We determined which diagnostic criteria were met, in addition to the temporal sequence of their appearance. Lastly, we grouped patients into 4 categories, H-TMA+/C-TMA+, H-TMA+/C-TMA-, H-TMA-/C-TMA+, and H-TMA-/C-TMA- and compared their NRM from allo-HSCT to investigate the diagnostic discrepancy and potential gain in the newly proposed criteria. Results: The cumulative incidence of H-TMA at one year following allo-HSCT was 30.8% (n=457, 95% CI: 28.4 - 33.3%) with median time to onset of 1.8 months (IQR: 1.1 – 4.2 months). On the other hand, the cumulative incidence of C-TMA was 3.7% (95% CI: 2.8-4.8%), with a later onset of 3.8 months (IQR: 2.1 – 7.6 months). Among H-TMA patients, 321 patients (70.2%) were classified as HR-TMA, with cumulative incidence of 22.1% (95% CI 19.9-24.3%). HR-TMA demonstrated significantly higher 1-year NRM rate from diagnosis compared to SR-TMA (48.4 vs. 23.5%, P<0.001). At diagnosis, the majority of patients with H-TMA presented with anemia (n=451, 98.7%), elevated LDH (n=445, 97.3%), schistocytes (n=387, 84.7%), and thrombocytopenia (n=369, 80.7%). The diagnosis was further supported by the presence of hypertension (≥140/90 mmHg) in 131 (28.7%) patients and proteinuria in 69 (15.1%) patients. The appearance of schistocytes occurred earliest, with a median of 17.2 days (IQR 3.8 - 29.8 days) before diagnosis. This was followed by elevated LDH (16.6 days, IQR 6.7 - 34.6 days) and anemia (14.5 days, IQR 6.6 - 31.0 days). Thrombocytopenia (4.7 days, IQR 0.7 - 11.6 days), proteinuria (1.3 days, IQR -6.2 - 4.5 days), and hypertension (1.0 days, IQR 1.0 - 4.0 days) appeared relatively later in H-TMA. Notably, schistocytes appeared after diagnosis in 25 (5.5%) patients. Furthermore, the absence of schistocytes did not significantly alter NRM rate in either overall H-TMA (35.9 vs. 41.5%, P=0.973) or HR-TMA (39.1 vs. 49.1%, P=0.770), further supporting the concept of TA-TMA without overt schistocytosis. All patients diagnosed with C-TMA were also identified by the harmonized criteria (H-TMA), forming a concordant group of H-TMA+/C-TMA+ patients (n=54). For this group, the harmonized criteria provided a significantly earlier diagnosis than Cho's criteria (median time to diagnosis 2.2 vs. 4.0 months, P=0.007). Patients in the H-TMA+/C-TMA- group exhibited relatively high NRM rate, although lower than H-TMA+/C-TMA+ group (32.8 vs. 50.0%, P=0.007). Of the 402 patients categorized as H-TMA+/C-TMA-, 363 (90.1%) were excluded from Cho's criteria because they never exhibited schistocytes ≥ 2/HPF, which is required for the diagnosis of C-TMA. Among the remaining patients who showed high schistocytes, 35 (8.7%) patients showed abnormal PT or aPTT at the time of schistocyte detection. Conclusions: Harmonization criteria may identify TA-TMA cases missed by conventional criteria through greater flexibility and inclusion of hypertension and proteinuria. Allowing diagnosis with lower or no schistocytes enables earlier detection of this high-risk complication. Adoption of harmonized criteria may facilitate timely intervention and improve risk prediction after allo-HSCT. Finally, the diagnostic and prognostic value of sC5b-9 in harmonization criteria, which was not assessed in this study, warrants further investigation in adult populations.