Long-term follow-up after high-dose therapy for high-risk multiple myeloma

Between 1985 and 1990, 133 patients with advanced multiple myeloma (MM) (74% resistance; 41% resistant relapse, RR) were treated with five high-dose therapy (HDT) regimens including: melphalan < or =100 mg/m2 (MEL 100) (46 patients); MEL 100 plus GM-CSF (24 patients); MEL 140 plus autologous bone marrow transplantation (ABMT) (eight patients); MEL 140 plus TBI 850 cGy plus ABMT (37 patients); and thiotepa 750 mg/m2 (THIO 750) + TBI 850 cGy plus ABMT (18 patients). The median follow-up of alive patients as of December 1997 was 9 years. Overall, 17% experienced treatment-related mortality within 60 days (TRM) and 12% achieved stringently defined complete remission (CR) with a median duration of 16 months; four of 16 patients (25%) remain in CR at 10 years. The median durations of event-free survival (EFS)/overall survival (OS) were 6/15 months. Superior EFS/OS were noted with MEL 100 plus GM-CSF and the two TBI-containing regimens (9/24 months among 79 patients) compared to the remaining 54 patients receiving MEL < or =100 or MEL 140 plus ABMT (3/5 months) (P = 0.0001/0.0001, respectively). Multivariate regression analyses (MVA) were performed so that, despite patient heterogeneity among the five treatment groups, potentially relevant disease, host, treatment, and supportive care variables could be identified that were associated with TRM, CR, EFS and OS. TRM was higher with creatinine >2.0 mg/dl, absence of ABMT/GM-CSF support and age >50 years; CR was superior with TBI-containing regimens and < or =12 months of prior therapy; EFS and OS both were longer with B2M < or =2.5 mg/l, age < or =50 years, absence of RR and with ABMT/GM-CSF support. In the presence of >2 favorable variables (32 % of patients), median EFS/OS durations of 18/48 months were observed which progressively declined with 2 and <2 favorable parameters to 6/11 months (28% of patients) to 3/5 months (40% of patients) (P = 0.0001/0.0001). At 10 years, 10 and 20% of patients with >2 favorable variables were event-free and alive, which was also true for the 37 patients receiving MEL 140 plus TBI. To appreciate possible long-term contributions of supportive care or treatment intensity, landmark analyses performed at 1, 2, 4 and 6 months revealed virtually identical ranking orders of prognostically favorable variables to those seen pre-HDT; once supportive care was accounted for, regimen intensity with added TBI did not emerge as an independent favorable feature.