Heart valves should not be routinely cultured

Abstract

Heart valve (HV) culture is one of the major Duke criteria for the diagnosis of definite infectious endocarditis (IE). However, previous series suggest that heart valve culture does not have good sensitivity (7.8 to 17.6%) and may be contaminated during manipulation. Our goal was to establish the value of routine cultures of heart valves in patients with and without IE. From 2004 to 2006, resected heart valves were systematically cultured according to standard procedures. The definition and etiology of IE were based on the Duke criteria and on valve PCR of specimens from blood culture-negative patients. Bacterial and fungal broad-range PCR was performed. A total of 1,101 heart valves were studied: 1,030 (93.6%) from patients without IE and 71 (6.4%) from patients with IE (42 patients). Overall, 321 (29.2%) cultures were positive (28/71 [39.4%] IE cases and 293/1,030 [28.4%] non-IE). All IE patients with negative heart valve cultures had received antimicrobial therapy. The yield of culture of heart valves for IE diagnosis was as follows: sensitivity, 25.4%; specificity, 71.6%; positive predictive value (PPV), 5.8%; and negative predictive value, 93.3%. Because of its poor sensitivity and PPV, valve cultures should not be performed for patients without a clinical suspicion of IE. For patients with confirmed IE, heart valve cultures should be interpreted with caution.

FIG. 1.

Results of the culture of 1,101 valves from patients with and without IE. *, a valve culture was considered a false negative when the valve culture of a patient with endocarditis was sterile or yielded a contaminant microorganism.