The combination of melphalan-prednisone-thalidomide (MPT) has been investigated in several clinical studies that differed significantly with regard to patient characteristics and treatment schedules. This prospective trial differs from previous melphalan-prednisone (MP) vs. MPT trials by treatment dosing, duration, routine anticoagulation, and permission for a crossover. Newly diagnosed patients with multiple myeloma (MM) (n = 122) aged greater than 55 yr, not eligible for transplantation were randomized to receive 8 cycles of M (9 mg/m2/d) and P (60 mg/m2/d) for 4 d every 6 wk (n = 62) or MP and thalidomide (100 mg/d) continuously (n = 60). Primary endpoint was treatment response and toxicities following 4 and 8 cycles of therapy. Secondary endpoints were disease-free (DFS) and overall survival (OS). Overall, MPT-treated patients were younger (median 69 yr vs. 72 yr; P = 0.016) and had a higher incidence of renal impairment (RI, 19% vs. 7%, respectively; P = 0.057). After 4 cycles of treatment (n = 115), there were more partial responses or better in the MPT arm than in the MP arm (57.9% vs. 37.5%; P = 0.030). However, DFS and OS were not significantly different between the arms after a median of 23 months follow-up (median OS 26.0 vs. 28.0 months, P = 0.655; DFS 21.0 vs. 14.0 months, P = 0.342, respectively). Crossover to MPT was required in 11 patients, 57% of whom responded to treatment. A higher rate of grade 3-4 infections was observed in the MPT arm compared with the MP arm (22.4% vs. 7.0%; P = 0.033). However, none of these infections were associated with febrile neutropenia. Death within the first 3 months was observed more frequently in the MP arm (n = 8, 14.0%) than in the MPT arm (n = 2, 3.4%; P = 0.053). Long-term discontinuation and dose reduction rates were also analyzed (MPT: 15.5% vs. MP: 5.3%; P = 0.072). Although patients treated with MPT were relatively younger and had more frequent RI, better responses and less early mortality were observed in all age groups despite more frequent discontinuation. This study is registered at http://www.clinicaltrials.gov as #NCT00934154.