Busulfan and Fludarabine Conditioning Regimen Negates the Impact of Comorbidity Score on Nonrelapse Mortality in Patients with AML/MDS

Abstract

BACKGROUND: Comorbidities as measured by newly developed Hematopoietic cell transplantation specific comorbidity index (HCT-CI) increases non relapse mortality in patients with hematological malignancies undergoing allogeneic transplantation. Whether this applies to recently treated patients with newer reduced toxicity regimens is not known. To evaluate this, we studied the factors influencing non relapse mortality in patients with AML/MDS.METHODS: 840 consecutive patients with AML/MDS undergoing allogeneic transplantation between January 1996 and April 2008 from a matched related or unrelated donor at our institution were studied. Information about disease characteristics, treatment, and outcomes was prospectively recorded in the departmental database. Comorbidities were retrospectively scored as previously described (Sorror et al Blood 106: 2912–2919, 2005). Predictors of non-relapse mortality(NRM) at 2 years post SCT were evaluated on univariate analysis using Cox's proportional hazards model and included age, sex, disease status at transplant, donor type, graft type, conditioning regimen (reduced intensity, ablative IV Busulfan(Bu) and Fludarabine(Flu), all other ablative regimens), and HCT-CI comorbidity score. Factors significant at the 0.1 level on univariate analysis were considered for classification and regression tree analysis (CART) to adjust for confounding and interaction effects. Statistical significance was defined at the 0.05 level, and cells including less than 10 patients were considered terminal in the CART analysis.RESULTS: There were 470(56%) males and 370(44%) females with a median age of 50 (range 18–77) years. 21% of patients were older than 60 years. HCT-CI comorbidity scores were as follows: 0 in 16% of patients, 1 in 13%, 2 in 13%, 3 or more in 58%. Donors were matched related for 58% of patients and unrelated for 42%. At the time of transplant, 22% of patients were in first complete remission (CR), 15% in second or third CR, and 63% had active disease. 36% of patients had reduced intensity conditioning regimen; 39% had myeloablative regimen consisting of Flu and Bu, and the remaining 25% had other myeloablative regimens. Cumulative incidence of NRM at 2 years was 26% (23–30). Univariate analysis showed that age > 60 (HR 1.6; p 0.002), disease status beyond first complete remission (HR 2.7; p <0.001), matched unrelated donor (HR 1.6; p 0.001), regimen other than Flu/Bu (HR 2.6; p<0.001), and a comorbidity score greater than 2 (HR 1.4; p 0.03) were associated with higher NRM. CART analysis (fig) showed that the use of Flu/Bu conditioning was the primary predictor of lower NRM. Age older than 60 years was the only additional significant predictor of NRM in patients who received Flu/Bu (HR 3.1; p 0.01). In patients who did not receive Flu/Bu, disease status at transplantation was the primary predictor of NRM with CR1 patients having a significantly lower incidence of NRM (CI =16%). The impact of donor type was only significant in patients who were not in CR1 with recipients of a matched related graft having a significantly lower NRM (CI=29%) than recipients of a matched unrelated graft (CI=46%). Comorbidity score did not significantly predict outcome in this analysis.CONCLUSION: Flu/Bu, a fully ablative reduced toxicity conditioning regimen, results in very low NRM, nullifying the impact of comorbidities. [Display omitted]