PURPOSE: The clinical diagnosis of infective endocarditis (IE) can be difficult. A new diagnostic schema for IE (the Duke criteria) has been proposed, utilizing clinical, microbiologic, and echocardiographic data. We evaluated the Duke criteria in a cohort of prospectively enrolled patients suspected of having IE and compared the diagnostic efficiency of these criteria with the previously published criteria of von Reyn. PATIENTS: Sixty-three febrile patients with suspected IE at a nonreferral, municipal hospital were evaluated. All patients had the following parameters defined: the presence and nature of underlying heart disease; recent abuse of intravenous drugs; peripheral stigmata of IE; blood culture results; findings on two-dimensional transthoracic and transesophageal echocardiography (TTE, TEE); and the results of open heart surgery. RESULTS: Twelve of 63 patients underwent open heart surgery, at which time IE was pathologically confirmed in 10 patients and excluded in 2 patients. All 10 patients with pathologically confirmed IE were classified as ''clinically definite'' by Duke criteria, whereas 5 of 10 were rejected by von Reyn criteria (p < 0.05). Among the remaining 51 patients suspected of IE and evaluated by both von Reyn and Duke clinical criteria, significantly more cases were classified as ''definite'' IE by Duke criteria than by von Reyn criteria (p < 10(-5)). Similarly, significantly fewer cases were rejected as IE by the Duke criteria as compared with the von Reyn criteria (p < 10(-6)). Duke criteria were also significantly better at diagnosing IE than von Reyn criteria in the following clinical settings: suspected right-sided IE (p < 0.01); suspected left-sided IE (p = 0.014); suspected culture-negative IE (p < 10(-2)); and IE complicating Staphylococcus aureus or viridans streptococcal bacteremias (p < 10(-5); p < 0.05, respectively). Among 30 cases defined as clinically definite by the Duke criteria, the presence of blood culture positivity and echocardiographically defined vegetations was important in this classification of 77% and 57% of cases, respectively. Among the 17 patients in the clinically definite category with vegetative endocarditis observed by echocardiography, 7 (41%) had vegetations defined only by TEE. CONCLUSION: The Duke criteria are superior to the von Reyn criteria for the clinical diagnosis of IE, predominately reflecting use of two-dimensional echocardiographic demonstration of valvular vegetations in the Duke schema.
