Impact of Adding Melphalan to Fludarabine and a Myeloablative Dose of Intravenous Busulfan for Patients Who Received Allogeneic Hematopoietic Stem Cell Transplantation

Introduction

Fludarabine and a myeloablative dose of intravenous busulfan with or without total body irradiation (Flu/Bu4) is a widely accepted myeloablative conditioning regimen with reduced toxicity for allogeneic hematopoietic stem cell transplantation (HSCT) that has been able to expand the indications for allogeneic HSCT while maintaining the intensity. However, relapse is a major concern especially in patients with high-risk hematological malignancies. Accordingly, a novel conditioning regimen that included Flu/Bu4 plus melphalan (Flu/Bu4/Mel) was used, with excellent outcomes, i.e., the 2-year overall survival (OS) was 54.9% in patients with non-remission acute myeloid leukemia. However, no study has compared Flu/Bu4/Mel and other conditioning regimens. Therefore, this multicenter retrospective observational study aimed to investigate the impact of adding melphalan to Flu/Bu4 by comparing patients who received Flu/Bu4/Mel and those who received Flu/Bu4.

Methods

The present study included 2394 adult patients who received Flu/Bu4/Mel (n=581) or Flu/Bu4 (n=1813) between January 2010 and December 2016 at the Japan Society of Hematopoietic Cell Transplantation. High-risk disease was defined as incomplete remission in patients with acute myeloid leukemia and acute lymphoblastic leukemia as well as refractory anemia with excess blast transformation in patients with myelodysplastic syndrome; standard risk disease was defined as stable disease or disease progression in patients with malignant lymphoma before HSCT.

We set the primary endpoint as the 5-year OS, evaluated by using the log-rank test. To compare the impact of the conditioning regimen on the primary endpoint, we used a multivariable Cox proportional hazards model. The following adjusted covariates were selected clinically based on previous studies: age, sex, disease, disease status at HSCT, hematopoietic cell transplant comorbidity index (HCT-CI), donor source, acute graft versus host disease prophylaxis, and years of HSCT. In addition, we performed propensity score matched analysis to minimize the potential treatment selection bias because the baseline characteristics of patients and disease could have influenced the selection of the conditioning regimen. Propensity score matching with a 1:1 ratio was performed by using the nearest neighbor-matching method with a caliper width fixed at 0.01.

In addition, detailed analysis was performed by dividing patients into the high-risk and standard risk groups because the initial single-center retrospective studies mainly enrolled patients with high-risk disease.

Results

A total of 2394 patients were enrolled and analyzed: 581 patients received Flu/Bu4/Mel and 1813 patients received Flu/Bu4. The median age was 60 years (interquartile range, 53-63 years) and 1489 patients (62.2%) were men. The clinical characteristics of patients treated with Flu/Bu4/Mel and those treated with Flu/Bu4 were different. The Flu/Bu4/Mel group included more patients who were young, had acute myeloid leukemia and malignant lymphoma, had high-risk disease, had higher HCT-CI, received cord blood transplantation, received tacrolimus-based acute graft versus host disease prophylaxis, and had undergone transplantation recently.

Regarding the primary endpoint, the 5-year OS after allogeneic HSCT was 33.7% (95% confidence interval [CI]: 27.5-40.1%) in the Flu/Bu4/Mel group vs 35.4% (95% CI: 32.7-38.1%) in the Flu/Bu4 group (p=0.794). The adjusted hazard ratio was 0.73 (95% CI: 0.62-0.85) for the Flu/Bu4/Mel group compared to the Flu/Bu4 group (p<0.001). After propensity score matching, the 5-year OS after allogeneic HSCT was 35.1% (95% CI: 28.0-42.2%) in the Flu/Bu4/Mel group vs 28.1% (95% CI: 22.4-34.0%) in the Flu/Bu4 group (p=0.008; Figure).

On detailed subgroup analysis, among patients with high-risk disease, the 5-year OS was 29.5% (95% CI: 23.0-36.4%) in the Flu/Bu4/Mel group vs 20.8% (95% CI: 17.3-24.6%) in the Flu/Bu4 group (p<0.001), while among standard-risk patients, it was 46.3% (95% CI: 35.6-56.2%) in the Flu/Bu4/Mel group vs 43.8% (95% CI: 40.3-47.4%) in the Flu/Bu4 group (p=0.745; P for interaction: 0.010).

Conclusion

The results showed that Flu/Bu4/Mel resulted in superior 5-year OS compared to Flu/Bu4, with an adjusted hazard ratio of 0.73. In addition, patients with high-risk disease showed better survival benefit with Flu/Bu4/Mel treatment.

Figure

View largeDownload slide

Figure

View largeDownload slide

Close modal