Background: Little is known about the rate at which new abdominal aortic aneurysms (AAAs) develop or whether screening older men for AAA, if undertaken, should be limited to once in a lifetime or repeated at intervals. Methods: A large population of veterans, aged 50 through 79 years, completed a questionnaire and underwent ultrasound screening for AAA. Of these, 5151 without AAA on the initial ultrasound (defined as infrarenal aortic diameter of 3.0 cm or larger) were selected randomly to be invited for a second ultrasound screening after an interval of 4 years. Local records and national databases were searched to identify deaths and AAA diagnoses made during the study interval in subjects who did not attend the rescreening. Results: Of the 5151 subjects selected for a second screening, 598 (11.6%) had died (none due to AAA), and 20 (0.4%) had an interim diagnosis of AAA. A second screening was performed on 2622 (50.9%), of whom 58 (2.2%; 95% confidence interval, 1.6%-2.8%) had new AAA. Three new AAAs were 4.0 to 4.9 cm, 10 were 3.5 to 3.9 cm, and 45 were 3.0 to 3.4 cm. Independent predictors of new AAA at the second screening included current smoker (odds ratio, 3.09; 95% confidence, 1.745.50), coronary artery disease (odds ratio, 1.81; 95% confidence interval, 1.07-3.07), and, in a separate model using a composite variable, any atherosclerosis (odds ratio, 1.97; 95% confidence interval, 1.16-3.35). Adding the interim and rescreening diagnosis rates suggests a 4- year incidence rate of 2.6%. Rescreening only in subjects with infrarenal aortic diameter of 2.5 cm or greater on the initial ultrasound would have missed more than two thirds of the new AAAs. Conclusions: A second screening is of little practical value after 4 years, mainly because the AAAs detected are small. However, the incidence that we observed suggests that a second screening after longer intervals (ie, more than 8 years) may provide yields similar to those seen in initial screening and therefore warrants further study.