OBJECTIVE Antibiotic-sterilised homograft valves stored at 4 degrees C have been implanted in the subcoronary position in this unit since 1973. This study was undertaken in order to assess the long-term function of these valves. METHODS All 249 patients undergoing homograft aortic valve replacement (AVR) at the Wessex Cardiothoracic Centre between April 1973 and December 1994 were studied. Homograft valve sizes ranged from 15 mm to 28 mm internal diameter, 202 (81.1%) varying between 18 mm and 22 mm. The mean patient follow-up was 12.4 years with a total follow-up of 3096 patient-years. There were six early deaths (2.4%). RESULTS On actuarial analysis, survival was 78.5+/-2.7% (1SE) at 10 years, 65.7+/-3.3% at 15 years and 55.0+/-3.9% at 20 years. The freedom from redo AVR was 87.9+/-2.4% at 10 years, 71.7 +/-3.8% at 15 years and 49.7+/-5.6% at 20 years. The freedom from structural degeneration was 85.6+/-2.5% at 10 years, 63.6+/-4.0% at 15 years and 41.9+/-6.4% at 20 years. On multivariate analysis the risk of valve failure was significantly higher in younger patients (P<0.0001) and in those who underwent aortic root tailoring (P = 0.024). The freedom from endocarditis was 98.4+/-0.9% at 10 years, 96.2+/-1.6% at 15 years and 95.1+/-1.9% at 20 years. Of the 249 patients, 218 had an isolated homograft AVR and were not anticoagulated. In this group there were two possible thromboembolic events. CONCLUSIONS As well as the established haemodynamic benefits, this study has shown that homograft AVR with antibiotic-sterilised 4 degrees C stored homograft valves implanted in the subcoronary position, offers good long-term results.