Fifty secondary progressive multiple sclerosis (SPMS) patients who had lost one or more EDSS points in the prior two years were selected to receive either cyclophosphamide (25 patients, 13 females, 12 males, F/M = 1.08; mean age: 42.4 years; mean disease duration: 13.3 years; mean EDSS at study entry: 5.7) or mitoxantrone (25 patients, 14 females, 11 males, F/M = 1.27; mean age: 38.2 years; mean disease duration: 11.5 years; mean EDSS at study entry: 5.5). SPMS patients were treated for two years with clinical evaluation (relapse rate, disability progression) every three months and radiological imaging (conventional magnetic resonance imaging) before therapy initiation and at the end of the first and second years of therapy. Safety profile and costs of the two therapeutic protocols were also analysed. In terms of clinical and radiological measures the drugs exerted a quite identical effect on both, and produced a significant reduction in both relapse rate (mitoxantrone Mito): p = 0.001, cyclophosphamide (Cy): p = 0.003) and disability progression (Mito: p = 0.01; Cy: p = 0.01). Subgroups of mitoxantrone- and cyclophosphamide-responding patients were identified (14/25 and 17/25, respectively) and were characterized by a significantly shorter duration of the secondary progressive phase of the disease. In these subgroups, the improvement in the EDSS score at the end of therapy was highly significant (p<0.0001 for Mito, p = 0.0004 for Cy). The safety profiles of both drugs were acceptable; however, the Cy-based therapy protocol was significantly less expensive. We conclude that Cy should be considered as a therapeutic option in rapidly deteriorating SPMS patients.