OBJECTIVE It is commonly held that preservation of the annulo-ventricular continuity during mitral valve replacement has a beneficial effect on postoperative ventricular function. This paper presents our eight-year experience with this technique. METHODS From 1986 to December 1992, 120 patients with rheumatic valve disease underwent mitral valve replacement (MVR) combined with preservation of the posterior leaflet of the mitral valve. The mean age was 33.87 years (range 16 to 63). The preoperative New York Heart Association (NYHA) class was III in 64% of the patients and IV in 3%. The early (30 day) mortality rate was 2.5% (3/120). Follow up was 100% complete. The total cumulative follow up was 477 patient-years (pty) with a mean 3.94 years (range: 2-8 years). RESULTS The actuarial survival rate (including hospital mortality) was 87.6% at eight years. Advanced age (p = 0.0457), increased preoperative functional capacity (p = 0.0251), increased preoperative end-systolic diameter (p = 0.0352) and combined tricuspid reconstruction (p = 0.0001) were found to be independent parameters for a lower actuarial survival rate. Six patients (1.25%/pty) developed thromboembolic complications. Two of these were cerebrovascular accidents (0.41%/pty) and four were caused by mechanical valve thrombosis (1.83%/pty). Freedom from thromboembolic complications (including mechanical valve thrombosis) was 89.8% +/- 7.9%. Fifteen patients (3.14%/pty) developed valve failure, four (0.83%/pty) mechanical valve thrombosis and nine (1.88%/pty) bioprosthetic valve failure. Freedom from reoperation was 60.7% +/- 16.1%. One patient with a St. Jude Medical valve (0.2%/pty) suffered from prosthetic valve endocarditis. Three patients (0.62%/pty) died during the follow up period, and, freedom from all valve related complications was 51.4% +/- 16.1% at eight years. CONCLUSIONS Preservation of the posterior leaflet during mitral valve replacement reduces the risk of early mortality and did not cause additional complications to the patients. Despite the beneficial effects of this technique in the long term this technique did not increase the long term survival in patients with associated impaired valvular and/or ventricular function.