Recurrent infective endocarditis: A multivariate analysis of 21 years of experience

Background. To evaluate which variables predict recurrence of endocarditis after surgical treatment, we reviewed our 21-year experience. Methods. Between January 1979 and May 2000, 308 consecutive valve replaceme nt procedures for infective endocarditis were performed in 271 patients. Un ivariate and multivariate time-related analyses were performed to retrospec tively evaluate the role of the following variables in the development of r ecurrent postoperative endocarditis: gender, site of endocarditis, previous valve disease, drug abuse, diabetes, positive valve/blood cultures, sepsis , perivalvular involvement, previous embolic events, type of replacement de vice, and persistent postoperative fever. Results. Clinical and echocardiographic follow-up was 97.36% complete, mean follow-up time was 53.2 +/- 3.4 months. Recurrent endocarditis developed i n 58 cases (22.5%). Variables predicting recurrence were prosthetic endocar ditis (p = 0.00001), positive valve culture (p 0.0039), and persistence of fever at the seventh postoperative day (p = 0.000001). Conclusions. Correct protocols of antibiotic therapy guided by microbiology may reduce the incidence of recurrent endocarditis to allow for surgery on sterile tissues and to prevent prosthetic infection. Recurrence rate is no t affected by the choice of valve substitute, but can be prevented by compl ete surgical debridement. (C) 2001 by The Society of Thoracic Surgeons.