Reoperative cardiac valve surgery: a multivariable analysis of risk factors.

From December 1983 to March 1992, of 1650 patients who underwent a cardiac valve procedure 41 (2.5%) underwent reoperation on prosthetic heart valves. The indications for reoperation were: structural degeneration (13 patients, 32%), non-structural dysfunction (nine patients, 22%), prosthetic valve endocarditis (seven patients, 17%), valve thrombosis (four patients, 10%), native valve pathology (three patients, 7%), aortic ascending pathology (five patients, 12%). The hospital mortality rate for mitral valve reoperation was 12% (two of 17) and for reoperation on the aortic prosthesis 26% (five of 19). No deaths occurred for associated mitral and aortic replacement (zero of five). The mortality rate for associated procedures was 24% (four of 17) versus 12% (three of 24) for isolated procedures. Preoperative and operative variables were analysed to determine risk factors for hospital mortality. New York Heart Association functional class, ascites, endocarditis and surgical timing had a univariate influence on operative mortality. The pulmonary capillary wedge pressure and preoperative creatinine level were also higher in non-survivors. Elective reoperation had a mortality rate of 7%, while that for emergency procedures was 38%. Multivariable analysis indicated a longer extracorporeal circulation time and the urgent-emergency status as predictors of mortality. Reoperation before severe haemodynamic impairment occurs is recommended.