While evoked potentials are sensitive tools for diagnosing multiple sclerosis, little is known about their prognostic value and their role in determining the course of the disease. To validate the visual and motor evoked potentials (VEP and MEP) as measures for the course of multiple sclerosis, we examined prospectively 30 patients with relapsing-remitting or secondary progressive multiple sclerosis. The Expanded Disability Status Scale (EDSS), VEP and MEP were measured at entry and after 6, 12 and 24 months. The Spearman rank correlation was used for statistical analysis. Applying multiple regression in 15 randomized patients allowed derivation of a formula for predicting changes in EDSS score based on changes in MEP and VEP. Validation was done by comparing the predicted with the real changes in EDSS in the other 15 patients. The number of pathological VEP and MEP results correlated at all four measurement points with the EDSS (rho > or = 0.6, P < or = 0.01). When the latencies of VEP and MEP were combined using the sum of their Z scores, correlation with the EDSS was even more significant (rho > or = 0.6, P < 0.001). Changes over time of electrophysiological data and EDSS were also correlated (rho = 0.43, P < 0.05). Moreover, VEP and MEP at baseline correlated with the EDSS after 2 years (rho = 0.43,P = 0.03). Reliable prediction of the course of multiple sclerosis for individual patients is not possible from VEP and MEP data. However, we conclude that, for groups of patients with secondary progressive or relapsing-remitting multiple sclerosis the combined testing of VEP and MEP yields numerical data that allow objective estimation of the course and prognosis of the disease.