BACKGROUND: Although data are inconsistent, angiotensin-converting enzyme inhibitors (ACE-Is) have been associated with a reduced incidence of abdominal aortic aneurysm (AAA) rupture in analysis of administrative databases. OBJECTIVES: (1) To investigate whether or not the ACE-I perindopril (Coversyl arginine, Servier) reduces small AAA growth rate and (2) to evaluate blood pressure (BP)-independent effects of perindopril on small AAA growth and to compare the repeatability of measurement of internal and external aneurysm diameters. DESIGN: A three-arm, multicentre, single-blind, randomised placebo-controlled trial. SETTING: Fourteen hospitals in England. PARTICIPANTS: Men or women aged ≥ 55 years with an AAA of 3.0-5.4 cm in diameter by internal or external measurement according to ultrasonography and who met the trial eligibility criteria. INTERVENTIONS: Patients were randomised to receive 10 mg of perindopril arginine daily, 5 mg of the calcium channel blocker amlodipine daily or placebo daily. MAIN OUTCOME MEASURES: The primary outcome was AAA diameter growth using external measurements in the longitudinal plane, which in-trial studies suggested was the preferred measure. Secondary outcome measures included AAA rupture, AAA repair, modelling of the time taken for the AAA to reach the threshold for intervention (5.5 cm) or referral for surgery, tolerance of study medication (measured by compliance, adverse events and quality of life) and a comparison of the repeatability of measures of internal and external AAA diameter. Patients were followed up every 3-6 months over 2 years. RESULTS: In total, 227 patients were recruited and randomised into the three groups, which were generally well matched at baseline. Multilevel modelling was used to determine the maximum likelihood estimates for AAA diameter growth. No significant differences in the estimates of annual growth were apparent [1.68 (standard error 0.02) mm, 1.77 (0.02) mm and 1.81 (0.02) mm in the placebo, perindopril and amlodipine groups, respectively]. Similarly, no significant differences in the slopes of modelled growth over time were apparent between perindopril and placebo (p = 0.78) or between perindopril and amlodipine (p = 0.89). The results were essentially unaffected by adjustment for potential confounders. Compliance, measured by pill counts, was good throughout (> 80% at all visit time points). There were no significant in-trial safety concerns. Six patients withdrew because of adverse events attributed to the study medications (n = 2 perindopril, n = 4 amlodipine). No patients ruptured their AAA and 27 underwent elective surgery during the trial (n = 9 placebo, n = 10 perindopril, n = 8 amlodipine). CONCLUSIONS: We were unable to demonstrate a significant impact of perindopril compared with placebo or amlodipine on small AAA growth over a 2-year period. Furthermore, there were no differences in the times to reach a diameter of 5.5 cm or undergo surgery among the three groups. Perindopril and amlodipine were well tolerated by this population. ...