This paper extends on previous data on prognosis in multiple sclerosis (MS), to encompass the entire course of the disease. The first episode suggestive of MS [the clinically isolated syndrome (CIS)] was included as a starting point, and the speed of secondary progression as an end point. Primary progressive MS was not included. Unique preconditions, with one neurological service covering the Göteborg district, allowed for establishing a strictly population-based, essentially untreated 15-year incidence cohort of 308 MS patients who were followed for 25 years. Survival analysis was performed as Kaplan-Meyer graphs, and independent predictors were ascertained by Cox regression analysis. A matrix of several predictors and end points was created. From CIS, a higher risk of developing clinically definite MS (CDMS), secondary progressive course and Disability Status Scale 7 (DSS7) was predicted by efferent tract lesions. However, less than 25% had reached DSS7 25 years after CIS with pure afferent lesions or other favorable predictors. During the first five years, higher relapse frequency, as well as incomplete remission of early bouts, predicted higher risks of secondary progressive course and DSS7 during follow-up to 25 years. However, these early predictors were unable to predict the rate of progression, which seems to contain an element of the disease process unassociated with its early events. Only late predictors, such as a shorter time from onset to secondary progression (1-10 years) and a higher number of functional systems involved at onset of progression predicted a faster progression rate. Predictors from this study could be used to refine historically controlled trials.