B-mode ultrasound examinations of the abdominal aorta were performed from 1990 to 1992 to evaluate the prevalence of abdominal aortic aneurysm (AAA) in a subgroup of the Pittsburgh cohort (656 participants, aged 65 to 90 years) of the Cardiovascular Health Study (CHS). In this pilot study, we evaluated various definitions of aneurysm and the reproducibility of the measurements. In year 5 (1992 to 1993) of the CHS, the entire cohort (4741 participants) was examined. AAA was defined as an infrarenal aortic diameter of > or= 3.0 cm, or a ratio of infrarenal to suprarenal diameter of > or= 1.2, or a history of AAA repair. For the entire CHS cohort, prevalence of aneurysms was 9.5% (451/4741) overall, with a prevalence among men of 14.2% (278/1956) and prevalence among women of 6.2% (173/2785). Variables significantly related to AAA were older age; male sex; history of angina, coronary heart disease, and myocardial infarction; lower ankle-arm blood pressure ratio; higher maximum carotid stenosis; greater intima-media thickness of the internal carotid artery; higher creatinine; lower HDL levels and higher LDL levels; and cigarette smoking. The study has documented the strong association of cardiovascular risk factors and measures of clinical and subclinical atherosclerosis and cardiovascular disease and prevalence of aneurysms. We used a definition that is more sensitive than previously reported (diameter or ratio), which allowed the detection of smaller aneurysms and possibly those at an earlier stage of development. Follow-up of this cohort may lead to new criteria for determining the risk factors for progression of aneurysms.