We have previously found a 3-fold increase in prevalence and a corresponding increase in incidence of multiple sclerosis (MS) from 1963 to 1983 in the county of Hordaland, Norway. When studying the official mortality statistics and the deceased patients in our clinical MS-material, no increase in death rate could be observed in the same period. We also found that official mortality statistics include both over- and underestimates of MS-patients. Of definite MS-patients in the clinical material, 91.7% were registered with MS on the death certificates, either as underlying or contributory cause of death. When studying only underlying cause of death, we found that 59.5% of deceased definite MS patients were registered as MS on the death certificates. This study shows that official mortality statistics reflect a change in incidence of MS both incompletely and with a delay of several decades.