Of the 2,794 subjects enrolled, 1,073 had a diagnosis of IA at the time of study entry, and 1,721 had no diagnosis of IA. There were 8,495 person-years of follow-up, with the overall mean follow-up time of 3.04+/-1.73 years. There was no significant difference (p=.12) in the length of follow-up between the affected IA (3.11+/-1.68 years) and unaffected (3.00+/-1.76) groups. Age at study entry for affecteds (54.811.7 years) was significantly (p.0001) older than for unaffecteds (48.625.9). A Cox proportional hazards model was utilized taking into account age, race, gender, affected status, rupture status, smoking history, and hypertension. After adjusting for age, the overall mortality rate for affecteds was not significantly different than unaffecteds. However, in affected subjects under the age of 55, the risk of death was 4.3 times that of unaffected subjects in the same age group (95% CI 1.58-11.7, p.004). The annual mortality rate was 13.2 per 1000 for affecteds and 8.5 per 1000 for unaffecteds. In conclusion, all but one of the deaths attributed to ruptured IA occurred shortly after study entry due to the initial rupture, yet the overall causes of death in this high-risk cohort during follow-up were more commonly unrelated to their aneurysms. None of the 1721 family members without known IA at study onset died from a subsequent ruptured IA.