Between 1953 and 1985, 109 consecutive patients (17% with isolated coarctation) younger than 3 years of age underwent resection of aortic coarctation. These patients had nonelective operations because of congestive heart failure or severe systemic hypertension not responding to medical therapy. Special surgical techniques were used to lower the prevalence of restenosis. Hospital mortality was 32% (35 patients). Logistic regression analysis proved that age at operation, pulmonary artery banding, and type of repair were independent predictors of hospital death. Late mortality occurred in 9 patients. Associated cardiac anomalies were an independent prognostic factor for late mortality. The operation rate for recoarctation was low (5.8%). However, the follow-up study revealed that 30 patients (41%) had recoarctation. The Kaplan-Meier estimate of recoarctation is 86% after 30 years' follow-up in patients undergoing classic end-to-end anastomosis with silk sutures (n = 48). None in the group with an "extended" anastomosis and polypropylene sutures (n = 26) had recoarctation. The Cox analysis revealed age at operation under 6 months to be prognostic for recoarctation. Because of the shorter period of follow-up, the extended anastomosis with polypropylene sutures proved not to be a significant prognostic factor for recoarctation. In the late postoperative period (mean follow-up 16.7 years), blood pressure was elevated in 49% of the patients. At last follow-up 27 (36%) of the 74 survivors had aortic valve disease. Patients operated on for coarctation of the aorta under the age of 3 years need lifelong follow-up for detection of restenosis, hypertension, and valvular disease. Reoperation-free rate is not a good criterion to judge the outcome of operation for coarctation. Longer follow-up will be needed to investigate whether or not the use of the extended anastomosis technique with polypropylene sutures lowers the prevalence of recoarctation.