Background Cytomegalovirus reactivation (CMVR) is a well-known complication contributing to morbidity and mortality in patients undergoing allogeneic hematopoietic stem cell transplantation (HSCT). Previous studies have demonstrated that CMVR is associated with poor clinical outcomes following HSCT. However, some studies have shown that CMVR following HSCT may have a recurrence-suppressing effect on hematological diseases, including acute myeloid leukemia and acute lymphoblastic leukemia. Conversely, data on the relevance between CMVR and relapse in patients with myelodysplastic syndromes (MDS) following HSCT remain limited and inconclusive. This nationwide retrospective study aimed to examine the association between CMVR and MDS relapse following HSCT. Methods This study was conducted by the Adult Myelodysplastic Syndromes Working Group of the Japan Society for Transplantation and Cellular Therapy (JSTCT). Clinical data were collected by the JSTCT and the Japanese Data Center for Hematopoietic Cell Transplantation employing the Transplant Registry Unified Management Program (TRUMP). Overall, 1,082 patients aged ≥16 years who were diagnosed with de novo MDS, underwent initial allogeneic HSCT between 2010 and 2018, and survived until 100 days post-transplant without disease relapse were examined. Patients who received cord blood transplantation, in vivo T-cell depletion, and prophylactic anti-CMV treatment were excluded. CMVR was defined as the initiation of CMV preemptive or definitive therapy within 100 days post-transplant. Patients underwent pp65 antigenemia monitoring commencing at the time of neutrophil engraftment after HSCT. Preemptive therapy was typically instituted upon the detection of a minimum of 2 CMV pp65 antigen-positive cells per 50,000 white blood cells. In analysis, CMVR and acute/chronic graft-versus-host disease (GVHD) were evaluated as time dependent covariates. This retrospective study was approved by the Data Management Committee of the JSTCT and the ethics committee of the Ehime University School of Medicine. Results Of the 1,082 patients, 622 (57.5%) experienced CMVR. The median time from HSCT to the onset of CMVR was 46 (range, 4-97) days. In contrast, reactivation beyond 100 days after HSCT was observed in 20 patients (1.8%). End-organ disease due to CMV infection was detected in 79 (7.3%) patients with a median of 61 (range, 19-385) days. CMV infections included enteritis (n = 65), retinitis (n = 5), pneumonitis (n = 4), hepatitis (n = 3), and others (n = 3). Of the whole cohort, the 5-year cumulative incidence of relapse (CIR), non-relapse mortality (NRM), disease-free survival, and overall survival (OS) were 15.6%, 26.7%, 57.7%, and 60.5%, respectively. The OS of patients with CMVR was significantly lower than that of patients without CMVR (P = 0.047). A total of 423 patients died at the time of analysis, and no difference in the cause of death was found between patients with and without CMVR. Then, the effect of CMVR on relapse after HSCT was analyzed by patient, disease, and transplant characteristics. Interestingly, in patients with <5% bone marrow blast count at the time of HSCT, CMVR was significantly associated with a lower CIR (5-year CIR, 7.6% vs. 16.4%; P = 0.002). On the contrary, no difference in the CIR was found between patients with and without CMVR in the group with ≥5% bone marrow blast count (5-year CIR, 19.4% vs. 18.0%; P = 0.72). Moreover, the OS of patients with CMVR was significantly lower than that of patients without CMVR in the group with ≥5% bone marrow blast count (P = 0.009). However, the OS of patients with <5% bone marrow blast count was similar between with and without CMVR (P = 0.82). In the multivariate analysis, CMVR was significantly associated with a lower CIR in the group with <5% bone marrow blast count (hazard ratio [HR], 0.40; 95% confidence interval [CI], 0.23-0.70; P = 0.001). Furthermore, in the analysis excluding patients with grade II-IV acute GVHD, CMVR was similarly associated with a lower CIR in the group with <5% bone marrow blast count (HR, 0.38; 95% CI, 0.18-0.80; P = 0.011). Conclusion This study presents that CMVR may have different effects on MDS relapse depending on the bone marrow blast count at the time of HSCT. Although more studies are needed to understand the underlying mechanisms of this association, these findings may serve as a basis for devising effective CMV prophylaxis after HSCT.