Long term results of isolated mitral and aortic valve replacement. A statistical evaluation

589 patients were followed up between 1972 and 1977 after aortic (285) or mitral valve (304) replacement. Three hundred and thirty patients had a ball valve (Starr-Edwards) and 259 a disc valve (Björk-Shiley). The average follow up was 37 months for the mitral patients and 40 months for the aortic. The results were analyzed by statistical methods based on multifactor analysis (cluster) in order to test the homogeneity of the different groups of patients. Survival and embolism were followed by actuarial rates. The survival rate between AVR patients at 72 months was 93 per cent for the Björk and 71 per cent for the Starr valve was a statistically significant difference. The survival of MVR at 48 months was 84 per cent for Starr and 78 per cent for Björk prostheses; no significant difference was found on statistical analysis. Embolic episodes occurred at a rate of 16 per cent for the Björk and 14 per cent for the Starr valves in AVR, and 10 per cent for the Björk and 18 per cent for Starr prostheses in MVR. These differences were not significant. By the cluster analysis based on different parameters (age, valve lesion, NYHA class, duration of symptoms, C/T ratio, size of the left atrium) it was shown that no proper risk factors were found between the parameters under consideration. We conclude that in both aortic and mitral valve replacement, the most important risk factor is the type of prosthesis and the presence of the prosthesis itself.