Applicability of the modified Duke criteria for the diagnosis of infective endocarditis in children with and without heart disease

Abstract

We sought to assess whether children with pre-existing heart disease treated for infective endocarditis (IE) were less likely to fulfill the modified Duke criteria for definite IE than those without heart disease. While the modified Duke criteria are widely accepted in research and clinical diagnosis, their applicability in diagnosing IE in children is not as well studied. We performed a retrospective single-center study of children treated for IE from January 1999-December 2009 and compared the proportion of children with and without heart disease who fulfilled modified Duke criteria for definite IE. We also examined these criteria in children with cyanotic heart disease, hospital-acquired IE, or recent cardiac surgery. Fifty (60%) of 84 children treated for IE met modified Duke criteria for definite IE. The proportion of children with and without heart disease who fulfilled modified Duke criteria for definite IE was not significantly different (38/68, 56%, versus 12/16, 75%, respectively, p = 0.26). Patients with cyanotic heart disease or hospital-associated IE were as likely to meet definite criteria as those without. However, children with early postoperative IE were less likely than those with late postoperative or without cardiac surgery to meet definite IE criteria (12/31, 39%, versus 38/53, 72%, p = 0.005). Only 60% of children treated for IE met modified Duke criteria for definite IE. Children with early postoperative IE were least likely to fulfill these criteria. Future studies should evaluate potential strategies to improve the diagnosis of IE in children.