Abstract
OBJECTIVE: We aimed to establish a model that can predict refractory/recurrent cytomegalovirus (CMV) infection after haploidentical donor (HID) hematopoietic stem cell transplantation (HSCT). METHODS: Consecutive acute leukemia patients receiving HID HSCT were enrolled (n = 289). We randomly selected 60% of the entire population (n = 170) as the training cohort, and the remaining 40% comprised the validation cohort (n = 119). Patients were treated according to the protocol registered at https://clinicaltrials.gov (NCT03756675). RESULTS: The model was as follows: Y = 0.0322 × (age) - 0.0696 × (gender) + 0.5492 × (underlying disease) + 0.0963 × (the cumulative dose of prednisone during pre-engraftment phase) - 0.0771 × (CD34+ cell counts in graft) - 1.2926. The threshold of probability was 0.5243, which helped to separate patients into high- and low-risk groups. In the low- and high-risk groups, the 100-day cumulative incidence of refractory/recurrent CMV was 42.0% [95% confidence interval (CI), 34.7%-49.4%] vs. 63.7% (95% CI, 54.8%-72.6%) (P < 0.001) for total patients and was 50.5% (95% confidence interval (CI), 40.9%-60.1%) vs. 71.0% (95% CI, 59.5%-82.4%) (P = 0.024) for those with acute graft-versus-host disease. It could also predict posttransplant mortality and survival. CONCLUSION: We established a comprehensive model that could predict the refractory/recurrent CMV infection after HID HSCT. CLINICAL TRIAL REGISTRATION: https://clinicaltrials.gov, identifier NCT03756675.