Shift to a New Donor Does Not Improve the Outcome After Second Allogeneic Stem Cell Transplantation (alloSCT) in Acute Leukemia Relapse After a First Allosct – a Risk Factor Analysis by the German Stem Cell Registry (DRST).

Relapse is a major cause of treatment failure after alloSCT against acute leukaemia, and no standard treatment has been established in this challenging situation. The introduction of reduced conditioning regimens, and the broader availability of alternative donors have increased the possibilities to perform a second alloSCT as salvage treatment, using different preparative regimen and/or different stem cell donors.

To evaluate the role of a second alloSCT (tx2) for the treatment of relapse after first alloSCT (tx1), we performed a nationwide retrospective analysis based on the German registry for stem cell transplantation (DRST). Datasets were completed by the reporting centres on request, following a specifically designed questionnaire.

212 patients (69% AML, 31% ALL), from 23 centres were included. Median age at tx1 was 37y. Donor at tx1 were HLA identical siblings (41%), matched unrelated (39%), mismatched family or unrelated (17%) or syngeneic donors (3%). Conditioning intensity at tx1 was standard (SIC, 62%), intermediate (intC, 25%) or reduced (RIC, 13%). Median remission after tx1 was 7 months, median time from relapse to tx2 was 74d. At tx2, patients were aplastic (4%), in CR (20%) or showed active disease (76%). In 59%, the same donor was used for tx1 and tx2, whereas a different donor was chosen in 41%. Conditioning at tx1/tx2 were SIC/SIC (14%), intC/intC (10%), (RIC/RIC (10%), less intensive at tx2 (mostly intC or RIC after SIC, 58%), or more intensive at tx2 (SIC after RIC or intC, 8%).

Following tx2, CR was achieved in 56% of patients, out of which 81% relapsed again. Hence, leukemia was the most frequent cause of death. With a median FU of 23 months after tx2, median OS after tx2 is 117d.

In a univariate analysis (log rank), OS after tx2 depended on stage at tx1 (CR vs. active disease, p<.001), stage at tx2 (CR vs. aplastic/active disease, p=.011) and duration of remission after tx1 (<=6m (1y OS 5%) vs. 6-12m (15%) vs. >12m (31%), p<.001). No significant difference was observed regarding age ( median), AML vs. ALL, family versus unrelated donor, or time point of alloSCT (2002). Shift to an alternative donor did not improve the results either. In a multivariate analysis (Cox Regression Model), time of remission after tx1 was the only significant factor for OS (p<.001, hazard ratio .51, 95%CI .49-.74).

Survival of acute leukemia after second allogeneic SCT is determined by the duration of remission after tx1. Using an alternative donor for tx2 did not improve the results in our series. Further analysis is required to evaluate the role of RIC regimen for tx2.

No relevant conflicts of interest to declare.