Does assessment of pretest probability of disease improve the utility of echocardiography in suspected endocarditis in children?

Abstract

Objective To compare the yield rate (YR) of echocardiography when evaluating children with suspected infectious endocarditis (IE) in both the actual clinical setting and in the hypothetic setting where strict clinical criteria are applied. Study design Medical records of 101 children undergoing echocardiography for suspected IE were reviewed. Echocardiograms with positive findings were identified and the actual diagnostic YR was calculated. With the use of clinical criteria proposed by von Reyn (VR), the probability of IE was retrospectively classified as (1) rejected, (2) possible, or (3) probable. Theoretic YR of echocardiography was calculated for each classification. Results The actual YR of echocardiography was 12% (12/101). The YR of echocardiography by VR class was 0% in rejected, 20% in possible, and 80% in probable cases (χ2 = 55.1, P < .0001). Echocardiography did not change the probability of IE in any patient classified as rejected, but allowed reassignment of disease probability in a significant proportion of patients with possible or probable IE. Conclusions The YR of echocardiography was significant when clinical probability of IE was intermediate-to-high, and low, with marginal clinical utility, when clinical probability was low. Strict pretest assessment of disease probability may lead to more effective utilization of echocardiography in this population. (J Pediatr 2003;142:263-7)