Abstract

Relapse of AML/MDS after allogeneic SCT is associated with poor outcome. A subset of patients (pts) can be salvaged with donor lymphocyte infusion (DLI) with or without preceding chemotherapy or with a second SCT. It was previously speculated that pts given RIC may have a better chance to be salvaged than pts failing high-dose conditioning. To answer this question we retrospectively analyzed results of 171 SCTs for AML/ MDS with intravenous busulfan (ivBu) -based regimens. 58 pts were eligible for myeloablative conditioning and were given ivBu (12.8 mg/kg) and cyclophosphamide (BuCy). Among all others, 57 were given RIC consisting of fludarabine (F) and ivBu (FB2, 6.4 mg/kg) and 56 were given a modified myeloablative conditioning consisting of F and full-dose ivBu (FB4, 12.8 mg/kg). Median age was 40 (17–64), 60 (43–75) and 52 (18–66) years, respectively (p<0.001). 56% of pts had active leukemia at SCT and 47% had SCT from unrelated or mismatched donors with no difference between the 3 regimens. With a median follow-up of 31 months (4–90), 70 pts relapsed, 20 after BuCy, 25 after FB2, and 25 after FB4, cumulative incidence 38%, 49% and 48%, respectively (p=NS). The kinetics of relapse was similar among the regimens. The median time to relapse was 5.3 (1–59), 3.4 (1–34) and 2.7 (1–30) months after BuCy, FB2 and FB4, respectively (p=NS); 65%,60% and 64% of all relapses occurred within 6 months of SCT, while 20%, 12% and 8% occurred after more than 2 years, respectively. 25 pts were treated with immune-suppression withdrawal and supportive care alone, 12 pts were given DLI with or without preceding low-dose chemotherapy, 33 pts were given intensive ‘salvage’ type chemotherapy followed by donor stem-cells and lymphocytes (n=24) or a second SCT from the same (n=4) or another donor (n=5). Allocation was based on pt status and pt and physician discretion, but was not significantly different among the regimens; 60%, 44% and 40% of pts initially conditioned with BuCy, FB2 and FB4 were treated intensively, respectively (p=NS). With a median follow-up of 15 months from relapse (0.5–52), 12 pts are alive, 7 in remission, 3 with active disease and 2 still early after treatment. The most important predicting factor for survival after relapse was the duration of remission. Pts relapsing > 6 months and < 6 months after SCT had a median survival of 7.2 and 1.4 months, respectively and estimated 2-year overall survival (OS) 22% (95C.I. 2–41) and 2% (95C.I. 0–7), respectively (p< 0.001). The 2-year OS after relapse was 20%, 0% and 4% after BuCy, FB2 and FB4, respectively (p=0.04). SCT in refractory disease was also predictive of poor outcome after relapse in univariant analysis (p<0.001). Multivariable analysis determined short remission after SCT and conditioning with FB regimens as independent factors predicting short survival after relapse, with hazard ratios of 3.3 (1.9–6.3, p< 0.001) and 2.0 (1.0–4.0, p=0.04), respectively. In conclusion the notion that pts given RIC can be salvage more easily if they relapse is not substantiated and should not be a rationale to select RIC over myeloablative conditioning. Although this observation may be explained by inability of pts ineligible for myeloablative conditioning to tolerate further intensive therapies, other mechanisms related to differences in the biology of the different transplants may be possible.

Author notes

Disclosure: No relevant conflicts of interest to declare.