Reoperations for valve surgery: perioperative mortality and determinants of risk for 1,000 patients, 1958-1984.

One thousand consecutive cardiac reoperations for valve surgery in 897 patients were reviewed to determine in-hospital mortality and indicators of risk. Subgroups based on the number of previous cardiac procedures and the valve or valves replaced or repaired at reoperation (aortic valve, mitral valve, tricuspid valve, or multiple valves and mortality [deaths/number of procedures (% mortality)]) for those subgroups are as follows: (Table: see text) Predictors of increased risk for a first aortic valve reoperation were advanced age (p = .0002), endocarditis (p = .0018), female sex (p = .014), impaired left ventricular function (p = .039), and number of coronary vessels obstructed by 70% or more (p = .055). For a first mitral valve reoperation, the predictors were advanced age (p less than .0001), preoperative shock or cardiac arrest (p = .01), previous aortic or tricuspid valve operations (p = .02), type of mitral valve procedure (risk for repair of periprosthetic leak was greater than mitral valve replacement which was greater than mitral valve-conserving operation [p = .05]), and impaired left ventricular function (p = .059). For a first multiple valve reoperation, the predictors were diabetes (p = .04) and ascites (p = .02), whereas patients undergoing mitral valve replacement and tricuspid valve operations were at decreased risk (p = .01). Comparison of second reoperations with first reoperations indicates risk increases for multiple operations (p = .01) but not for aortic or mitral valve procedures. Rereplacement of a prosthesis (p = .007), coronary bypass grafting at reoperation (p = .006), and advanced age (p = .06) increased the risk for second reoperations. Age is the most consistent predictor of risk for patients undergoing valve reoperations.