Between May 1980 and December 1984, 1000 Medtronic Hall prostheses (538 mitral and 462 aortic) were implanted in 852 patients (mean age 30 +/- 15 years) of a third-world type of population group. The most common valve lesions were aortic and mitral regurgitation, and 32% of the patients were in New York Heart Association class IV or greater. Double valve (aortic and mitral) operations were performed in 209 patients. The 782 operative survivors were followed up for 2 to 6.5 years (mean 3.5 years) for a cumulative 2676 patient-years. Valve performance was analyzed in each position separately for valve-related complications. Results for the whole group for the following events are expressed in both linearized (percent/patient-year) and actuarial (percent +/- standard error of the mean freedom at 5 years) terms, respectively: valve-related mortality, 2.9 and 92 +/- 2; reoperation, 1.9 and 90 +/- 2; systemic thromboembolism, 3.3 and 85 +/- 2; thrombotic obstruction, 1.2 and 95 +/- 1; prosthetic valve endocarditis, 0.7 and 97 +/- 1; anticoagulant-related hemorrhage, 0.7 and 98 +/- 1; and periprosthetic leak, 0.7 and 97 +/- 1. Corresponding composite figures for valve failure were 4.3 and 85 +/- 2, and for all valve-related morbidity and mortality were 6.7 and 75 +/- 3. Thrombotic obstruction (fatal in 71% of the cases), anticoagulant-related hemorrhage (61% fatal), and prosthetic endocarditis (44% fatal) were the most lethal complications. Although the differences did not reach statistical significance, aortic prostheses appeared to have a higher incidence of complications than the mitral prostheses. When analyzed against the background of an essentially noncompliant population group, the Medtronic Hall prosthesis proved to be reliable and remarkably free from structural failure. The high incidence of thromboembolism and thrombotic obstruction still mandates anticoagulation.