We have performed an 11-year follow-up study on mortality in rheumatoid arthritis (RA). The population was selected by random sampling and consisted of 293 persons with RA fulfilling two sets of relevant criteria and a contrast group of 453 healthy persons. The mortality in the RA group was found elevated, when compared with both the contrast group and the general Stockholm population. The mortality was correlated with the severity of the disease, as defined by the numbers of criteria for RA fulfilled, the Steinbrocker functional grading, and serological titres. The differences in the distribution of causes of death between RA and contrast groups were not large enough to warrant special consideration. Only 9 out of 84 death certificates in the RA group mentioned rheumatic disease, and 4 of these had RA as underlying cause of death.