Aortic valve (AV) preservation and repair, although effective, is performed in a limited number of centres. Lack of wider application might be due to challenges in dissemination of tacit surgical knowledge. We examined the learning curve in 2 centres that initiated dedicated programs in AV repair.Prospectively collected data on the first 100 (cohort A) and 150 consecutive patients (cohort B) who underwent AV repair surgery were analyzed. Safety end points included mortality, myocardial infarction or stroke, early AV repeat surgery, re-exploration for bleeding, or pacemaker implantation. Efficiency was assessed according to aortic crossclamp and cardiopulmonary bypass times. Efficacy parameters included residual aortic insufficiency or stenosis. Indices of case complexity included presence of severe aortic insufficiency, nontrileaflet AV, and associated cardiac procedures. Each cohort was divided into 3 equal tertiles (T1, T2, and T3).Early mortality was ≤ 1% in both cohorts. In cohort A, a total of 12 safety events occurred with a significant reduction in incidence over the tertiles (18%, 15%, and 3%, in T1, T2, and T3, respectively; P = 0.05). In cohort B, 20 safety events occurred in 18 patients with a trend toward reduction of incidence over tertiles (20%, 12%, and 8%, in T1, T2, and T3, respectively; P = 0.14). aortic crossclamp and cardiopulmonary bypass times decreased significantly after T2 in cohort A and T1 in cohort B (P < 0.01). Intraoperative procedural efficacy was similar across tertiles in both cohorts.Procedural safety and efficiency improves with experience whereas efficacy is consistent over time. AV repair is reproducible and appears to have a learning curve of approximately 40-60 cases.