Infective Endocarditis - Diagnosis

Diagnosis

In general, a patient should fulfill the Duke criteria in order to establish the diagnosis of endocarditis.

As the Duke criteria rely heavily on the results of echocardiography, research has addressed when to order an echocardiogram by using signs and symptoms to predict occult endocarditis among patients with intravenous drug abuse and among non drug-abusing patients. Unfortunately, this research is over 20 years old and it is possible that changes in the epidemiology of endocarditis and bacteria such as staphylococci make the following estimates incorrect.

Among patients who do not use intravenous drugs and have a fever in the emergency room, there is a less than 5% chance of occult endocarditis. Mellors in 1987 found no cases of endocarditis nor of staphylococcal bacteremia among 135 febrile patients in the emergency room. The upper confidence interval for 0% of 135 is 5%, so for statistical reasons alone, there is up to a 5% chance of endocarditis among these patients. In contrast, Leibovici found that among 113 non-selected adults admitted to the hospital because of fever there were two cases (1.8% with 95%CI: 0% to 7%) of endocarditis.

Among patients who do use intravenous drugs and have a fever in the emergency room, there is about a 10% to 15% prevalence of endocarditis. This estimate is not substantially changed by whether the doctor believes the patient has a trivial explanation for their fever. Weisse found that 13% of 121 patients had endocarditis. Marantz also found a prevalence of endocarditis of 13% among such patients in the emergency room with fever. Samet found a 6% incidence among 283 such patients, but after excluding patients with initially apparent major illness to explain the fever (including 11 cases of manifest endocarditis), there was a 7% prevalence of endocarditis.

Among patients with staphylococcal bacteremia (SAB), one study found a 29% prevalence of endocarditis in community-acquired SAB versus 5% in nosocomial SAB. However, only 2% of strains were resistant to methicillin and so these numbers may be low in areas of higher resistance.