RESUMO
BACKGROUND: Infective endocarditis (IE) is a severe disease. Pathogen isolation is fundamental so as to treat effectively and reduce morbidity and mortality. Blood and valve culture and histopathology (HP) are routinely employed for this purpose. Valve HP is the gold standard for diagnosis. OBJECTIVES: To determine the sensitivity and specificity of clinical criteria for IE (the modified Duke and the St Thomas' minor modifications, STH) of blood and valve culture compared to valve HP, and to evaluate antibiotic treatment duration. METHODS: Prospective case series of patients, from 2006 to 2014 with surgically treated IE. Statistical analysis was done by the R software. RESULTS: There were 136 clinically definite episodes of IE in 133 patients. Mean age ± SD was 43 ± 15.6 years and IE was left sided in 81.6 %. HP was definite in 96 valves examined, which were used as gold standard. Sensitivity of blood culture was 61 % (CI 0.51, 0.71) and of valve culture 15 % (CI 0.07, 0.26). The modified Duke criteria were 65 % (CI 0.55, 0.75) sensitive and 33 % specific, while the STH's sensitivity was 72 % (CI 0.61, 0.80) with similar specificity. In multivariate analysis and logistic regression, the only variable with statistical significance was duration of antibiotic therapy postoperatively. CONCLUSIONS: Valve HP had high sensitivity and valve culture low sensitivity in the diagnosis of IE. The STH's criteria were more sensitive than the modified Duke criteria. Valve HP should guide duration of postoperative antibiotic treatment.