Abstract
Cytomegalovirus (CMV) infection is a major complication in allogeneic stem cell transplantation. The utility of CMV prophylaxis with letermovir has been reported; however, the specific applications remain unclear. In this study, we retrospectively analyzed large-scale registry data (N = 10 480) to clarify the risk factors for nonrelapse mortality (NRM) in connection with CMV reactivation. First, we identified risk factors for CMV reactivation using multivariate analysis and developed a scoring model. Although the model effectively stratified reactivation risk into 3 groups (43.7% vs 60.9% vs 71.5%; P < .001), the 3-year NRM was significantly higher in patients with CMV reactivation, even in the low (20.9% vs 13.0%, P < .001), intermediate (21.4% vs 15.6%; P < .001), and high (29.3% vs 18.0%; P < .001) reactivation risk groups. Next, survival analysis considering competing risks, time-dependent covariates, and interaction terms for exploring the heterogeneous impact of CMV reactivation on NRM in the training cohort revealed that chronic myeloid leukemia (CML) (hazard ratio [HR], 1.76; 95% confidence interval [CI], 1.05-2.96; P = .033), good performance status (PS) (HR, 1.42; 95% CI, 1.04-1.94; P = .028), HLA-matched donor (HR, 1.34; 95% CI, 1.06-1.70; P = .013), and standard-risk disease (HR, 1.28; 95% CI, 1.04-1.58; P = .022) were associated with increased NRM. In the test cohort, CMV reactivation was significantly associated with increased 3-year NRM among patients with 2 to 4 factors (22.1% vs 13.1%; P < .001) but was comparable among patients with 0 or 1 factor (23.2% vs 20.4%; P = .62). We propose that CMV prophylaxis should be determined based on reactivation risk, as well as these other factors.