Real-World Application of Recently Proposed ASTCT/CIBMTR/EBMT/APBMT Consensus Risk Stratification for...

Consensus diagnostic and risk stratification of transplantation-associated thrombotic microangiopathy (TA-TMA) was recently achieved from international transplantation groups. Although the proposed diagnostic criteria have been applied to multiple pediatric cohorts, there are scant data applying the novel risk stratification approach in children with TA-TMA. In this retrospective cohort study, all children undergoing an allogeneic HCT or autologous HCT for neuroblastoma were prospectively screened for TA-TMA, diagnosed, and risk-stratified using the Jodele criteria from August 2019 to October 2023. Our institutional practice during the study period was treat all Jodele intermediate-risk (IR) and high-risk (HR) patients with eculizumab. Harmonization risk stratification criteria were applied retrospectively. All survival analyses were calculated from the day of TA-TMA diagnosis. To identify which specific harmonization high-risk features were the most important predictors for nonrelapse mortality (NRM), full and reduced logistical regression models were tested. The lowest Bayes information criterion and optimal Mallows CP statistic were used to identify the best subset. The analysis was performed with SAS 9.4 (SAS Institute, Cary, NC). Fifty-two children were diagnosed with TA-TMA during the study period, at a median of 37.5 days post-HCT (range, 3 to 735 days). Using Jodele risk stratification, 11 (21%) were SR, 21 (40%) were IR, and 20 (39%) were HR. Forty (77%) were treated with eculizumab. There were no statistically significant differences in NRM among Jodele risk groups, although overall survival (OS) differed significantly. Using the harmonized stratification, 49 children (94%) were stratified as HR and 3 as standard risk (SR), there were no statistically significant differences in NRM or OS between groups. Eight children (15.4%) were classified as SR using Jodele risk stratification but restratified as HR using the harmonization criteria. One child (12.5%) died in the setting of severe GVHD, and the remaining 7 were alive at the last follow-up. In a best subset model, lactate dehydrogenase (LDH) level >2 times the upper limit of normal (ULN) (odds ratio [OR], 6.52, 95% confidence interval [CI], .96 to 44.3; P = .05), grade II-IV acute graft-versus-host disease (GVHD) at the time of TA-TMA diagnosis (OR, 15.4; 95% CI, 2.14 to 110.68; P = .01), and organ dysfunction at the time of TA-TMA (OR, 21.5; 95% CI, 2.96 to 156.37; P = .002) were significantly associated with NRM; elevated sC5b-9, urine protein/creatinine ratio, and viral infections were not significantly associated with NRM. Using these best-fit criteria, 14 patients were classified as SR and 38 were classified as HR, NRM was significantly higher, and OS was significantly lower. In this cohort of children with TA-TMA, retrospective application of the harmonization criteria resulted in more patients stratified as HR compared to use of the previously described Jodele criteria. The intention of the harmonization criteria was to identify those at greatest risk of poor outcomes; while all harmonization SR patients survived, this risk stratification was very sensitive. Previous criticisms of harmonization risk stratification include limited access to sC5b-9 testing. These data suggest that organ dysfuncion, acute GVHD, and LDH >2 times ULN are the most important predictors of NRM in this cohort, allowing risk stratification even in the absence of available sC5b-9 testing. Additional studies are needed to validate these findings.