Based on data from a multiple sclerosis (MS) prevalence study which had been carried out in 1986 in the canton of Berne, Switzerland, a follow-up was performed 10 years later to gather information on mortality in the original study population. The authors used information on residency and death as gathered from municipal offices and, additionally, by record-linkage with the Swiss cause of death statistics. Slightly more than 80% of the cases were identified unequivocally as of January 1996. Among them, 21% of the cases died during the ten-year period; 70% of them have an MS entry in the cause of death statistics, mostly as the main cause of death. A large proportion of the non-identifiable cases appear to be related to mortality; thus, the findings here do not provide a promising basis for certain further analyses. In conclusion, continuous updating of personal data is the only way to avoid loss to follow-up in the carefully assembled prevalence database.