We describe the clinical and laboratory characteristics of 52 patients with rheumatoid vasculitis whose condition was diagnosed at a tertiary care center between 1974 and 1981, and we report their survival and the factors that were associated with decreased survival. The patients with rheumatoid vasculitis had decreased survival in comparison with an age-, sex-, and region-matched general population. Their survival was also decreased in comparison to that of an incidence cohort of community patients with rheumatoid arthritis. In the latter cohort, decreased survival was confined to those patients with classic but not definite rheumatoid arthritis. After partial correction for referral bias, we found no difference in survival between the cohort with rheumatoid vasculitis and the cohort with classic rheumatoid arthritis. We found that the age at diagnosis of rheumatoid vasculitis, the therapeutic decisions before and at diagnosis, and the referral distance were the best predictors of survival. Abnormal urinary sediment and hypergammaglobulinemia also predicted poor survival, but because of a lack of specificity in a small number of clinically abnormal values, we urge a cautious interpretation of their importance.