The aim of this retrospective analysis was to assess the benefit of reduced-intensity conditioning allo SCT (RIC allo-SCT) in a cohort of 32 relapsed multiple myeloma (MM) patients. A total of 19 patients had an HLA-identical sibling donor ('donor' group), while 13 patients had no donor ('no-donor' group). There were no significant differences between these two groups as for prognosis risk factors. Eighteen patients from the 'donor' group could actually proceed to RIC allo-SCT. With a median follow-up of 36 (range, 21-60) months, six patients died from transplant-related toxicity (cumulative incidence, 33% (95% CI, 11-55%)). Only 4 patients from the 18 transplanted patients (22%; 95% CI, 7-48%) progressed after RIC allo-SCT, as compared to 12 (86%; 95% CI, 56-98%; P=0.0003) among the nontransplanted patients. In an 'intention-to-treat' analysis, the Kaplan-Meier estimate of PFS was significantly higher in the 'donor' group as compared to the 'no-donor' group (P=0.01; 46 versus 8% at 3 years). There was no difference in terms of overall survival. However, in multivariate analysis, actual performance of RIC allo-SCT was associated with better PFS (relative risk, 0.35; 95% CI, 0.15-0.82; P=0.01). These data suggest a potential benefit for RIC allo-SCT in the management of relapsed MM warranting further prospective investigations.