The study aim was to review our experience in surgical treatment of infective mitral valve endocarditis, and to identify predictors of early and late outcome. A total of 91 patients (52 males, 39 females, mean age 55.6 years) underwent surgery for endocarditis of mitral (n = 65), mitral and aortic (n = 25) and mitral, aortic and tricuspid valves (n = 1). Native valve endocarditis (NVE) was present in 60 patients (66%) and prosthetic valve endocarditis (PVE) in 31 (34%). Indications for surgery were heart failure (n = 32), valve dysfunction (n = 23), vegetations (n = 21) and persistent sepsis (n = 11). Eighty-six patients (95%) were in NYHA classes III-IV, and 58 (64%) had active culture-positive endocarditis at surgery. Mechanical valves were implanted in 73 patients and bioprostheses in 13; valves were repaired in five. The impact of 46 parameters on early and late outcome was defined using univariate and multivariate statistical analysis. Mean follow up was 5.5 years (range: 0-23.1 years). Operative mortality rate was 11% (n = 10). Recurrent infection occurred in five patients, and reoperation was required in eight. Freedom from recurrent infection and reoperation at 10 years was 89.1% and 87.8%. There were 22 late deaths, 15 from cardiac causes. Actuarial survival rates for all patients at 5, 10 and 15 years were 73.0%, 62.7% and 58.7% (81.9%, 69.7% and 66.0% for hospital survivors). The following were independent predictors: preoperative pulmonary edema (p = 0.01) for operative mortality; PVE (p = 0.02) for recurrence; younger age (p = 0.02) and PVE (p = 0.02) for reoperation; male gender (p = 0.004) and longer ITU stay for survival (all patients included); male gender (p = 0.01) and myocardial invasion by infection (p = 0.02) for survival (if only hospital survivors included). In conclusion, surgery for infective mitral valve endocarditis carries a high, but acceptable, risk and provides freedom from recurrent infection, reoperation and improved long-term survival. Analysis of these data showed preoperative hemodynamic status to be the major predictor of in-hospital outcome, PVE increased the risk for recurrent infection and reoperation, while male gender and myocardial invasion by the infection reduced the probability of long-term survival. Type of pathogen, activity of infection and involvement of more than one valve did not influence early and/or late outcome.