Clinical course and predictors of death in prosthetic valve endocarditis over a 20-year period

Objective: To compare early and late outcome of patients with prosthetic valve endocarditis treated medically versus surgically and to determine predictors of in-hospital death. We retrospectively reviewed patient's clinical records, including laboratory findings, surgery, and pathologic files, in an acute-care, 1200-bed teaching hospital. Methods: One hundred thirty-three episodes of definite prosthetic valve endocarditis as defined by the Duke University diagnostic criteria occurred in 122 patients from January 1986 to December 2005. Logistic regression model was used to identify prognostic factors of in-hospital mortality. Long-term follow-up was made to assess late prognosis. Results: Bioprostheses were involved in 52% of cases and mechanical valves in 48%. The aortic valve was affected in 45% of patients. Staphylococcus epidermidis was isolated in 23% of cases, Streptococcus spp in 21%, S aureus in 13%, and Enterococcus in 8%. Cultures were negative in 18% of cases. Twenty-six patients were treated medically and 107 with combined antibiotics and valve replacement. The operative mortality was 6.5% and the in-hospital mortality, 29%. Presence of an abscess at echocardiography, urgent surgical treatment, heart failure, thrombocytopenia, and renal failure were significant predictors of in-hospital death. Kaplan-Meier survival at 12 months was 42% in patients treated medically and 71% in those treated surgically (P = .0007). Freedom from endocarditis was 91% at the end of follow-up. Conclusions: Prosthetic valve endocarditis is a serious condition with high mortality. Patients with perivalvular abscess had a worse prognosis, and combined surgical and medical treatment could be the preferred approach to improve outcome. (J Thorac Cardiovasc Surg 2010; 139: 887-93)