Background: The Duke criteria and the use of the transesophageal echocardiogram (TEE) are sensitive and specific tools for diagnosing infective endocarditis (IE). We evaluated how physicians use the Duke criteria in the diagnosis and management of patients with suspected IE and a negative TEE.
Methods: We performed a retrospective analysis of 95 consecutive patients with suspected IE who underwent TEEs with respect to the impact of echocardiograms, blood cultures, and the Duke criteria on diagnosis and treatment.
Results: Of these 95 patients, 28 patients had positive TEEs and 67 had negative TEEs for IE. All of the 28 patients with positive TEEs were treated for IE. Of the 67 patients with negative TEEs, 40 patients were treated for IE, and 27 patients were not. Of the 40 patients treated for IE, 32 patients fulfilled the Duke criteria for definite or possible IE, whereas 8 patients had the diagnosis rejected. Of the 27 patients not treated for IE, 11 patients fulfilled the Duke criteria, whereas 16 patients did not.
Conclusions: There was poor correlation between the Duke criteria and the treatment of IE. Of those patients whose condition was diagnosed by the Duke criteria as definite or possible IE, 26% were not treated for IE. Of those patients whose diagnosis of IE was rejected by the Duke criteria, 33% were treated for IE. The most important factors in influencing physicians to diagnose the condition of patients as IE were isolation of typical organisms from blood cultures and/or a positive TEE. The most important reasons for rejecting the diagnosis of IE were isolation of atypical organisms from blood cultures and identification of alternative sources of bacteremia. The physicians relied on the Duke criteria to a lesser extent for their treatment decisions. Two patients with definite IE diagnosed by the Duke criteria who were not treated as IE, did not receive consultation from the infectious diseases department. We feel that this underscores the value of the infectious diseases consultant in the management of patients with IE.