Background Myelodysplastic syndrome (MDS) is associated with dismal outcomes after relapse following hematopoietic stem cell transplantation (HSCT), a situation that remains a clinical challenge. The median overall survival (OS) for patients who relapsed after HSCT has been reported to be only 4.7 months, emphasizing the need to improve post-relapse survival. A second HSCT is a potentially curative approach for these patients; however, its outcomes have been limited so far. In a study based on the Japanese registry between 1990 and 2015, patients who underwent a second HSCT for MDS had a 5-year OS rate of only 25%, with a 5-year non-relapse mortality (NRM) rate of 34%. Following the publication of this study, several developments have occurred. For example, the introduction of post-transplant cyclophosphamide for graft-versus-host disease prophylaxis has expanded the use of haploidentical donors. In addition, the outcomes of cord blood transplantation have improved in Japan. Therefore, we need to re-evaluate the outcomes and risk factors after second HSCT in the context of these recent developments.

Methods We conducted a retrospective analysis of Japanese patients with MDS aged 16 years or older who were registered in the Japanese Data Center for Hematopoietic Cell Transplantation and underwent a second HSCT due to disease relapse or progression between January 2016 and December 2023. We excluded patients who received a second HSCT for graft failure or secondary MDS, or as a planned second transplantation. The primary endpoint was overall survival from the second HSCT. We estimated the 2- and 5-year OS using the Kaplan-Meier method and identified prognostic factors using a multivariable Cox proportional hazards model. The candidate risk factors included age (<50 vs ≥50 years), sex, performance status (0-1 vs 2-4), karyotype risk (poor or very poor vs others), the interval from the first HSCT to relapse or progression (≤18 months vs >18 months), donor source (related donor vs unrelated donor vs cord blood vs haploidentical donor), and conditioning regimen (reduced-intensity vs myeloablative conditioning) at the second HSCT. We also evaluated the relapse rate and NRM considering competing risks.

Result A total of 122 patients were included in this study. The median age at the second HSCT was 56 years (range, 19–72 years), and 89 patients (73%) were 50 years or older. Eighty-four patients (69%) were male, and 24 (20%) had a performance status of 2 or higher. Furthermore, 78 patients (64%) had poor or very poor karyotype risk. The median interval between the first and second HSCT was 290 days (range: 49–2,836 days). Relapse within 18 months was observed in 102 patients (84%). Cord blood was used in 65 patients (54%), and haploidentical donors were used in 13 patients (11%).

The 2-year OS rate was 23.2% (95% confidence interval [CI]: 16.3–32.8%), and the 5-year OS rate was 19.3% (95% CI: 12.8–29.0%). The 2-year cumulative incidence of relapse was 49.5% (95% CI: 40.3–58.6%), whereas the 2-year NRM was 32.8% (95% CI: 24.3–41.4%). Multivariable analysis revealed that poor performance status (hazard ratio [HR]: 3.21, 95% CI: 1.87–5.48; p < 0.001) and poor or very poor karyotype risk (HR: 1.92, 95% CI: 1.18–3.11; p = 0.009) were associated with inferior OS, and female sex was associated with better OS (HR: 0.57, 95% CI: 0.34–0.96; p = 0.03). Other factors were not significantly associated with OS.

Conclusion We revealed that the 2-year OS was 23.2% and the 5-year OS was 19.3% in patients with relapsed MDS who received the second HSCT. This prognosis was similar to previously reported outcomes.

This content is only available as a PDF.
2025
Sign in via your Institution