Institutional members access full text with Ovid®

Share this article on:

Differentiation of systemic infection and congenital obstructive left heart disease in the very young infant

Pediatric Emergency Care: August 1998

Background The differentiation of severe systemic infection, such as sepsis or meningitis, from a congenital obstructive left heart abnormality presents a unique challenge to clinicians responsible for the care of such infants in the first few weeks of life. Clinical findings are very similar in the two populations. Failure to identify the need for specific intervention, such as prostaglandin administration, by the primary care or emergency physician may result in increased morbidity or death in these infants.

Methods We undertook a retrospective review of critically ill infants 0 to 28 days of age presenting with either bacterial sepsis or meningitis or a congenital obstructive left heart syndrome (COLHS), in order to identify historical, physical, or laboratory findings which might differentiate the two groups at presentation. Discriminant analysis was performed using the presence or absence of COLHS as the dependent variable. A COLHS index was derived to determine its sensitivity and specificity for differentiating the two groups.

Results The presence of cardiomegaly predicted COLHS with 85% sensitivity and 95% specificity. Cardiomegaly had a positive predictive value for COLHS of 0.95. Unfortunately, most of the other variables which, individually or in combination, were significantly different between the two groups demonstrated poor sensitivity for prediction of the presence of obstructive left heart disease. Eleven of the predictor variables were chosen for inclusion in the multivariate model, and a COLHS index was developed which correctly classified 62/63 cases (98% sensitivity, 100% specificity).

Conclusions We conclude that while it is very difficult to differentiate these two groups at presentation, early clinical suspicion of COLHS with attention to key clinical parameters identified in this study may expedite appropriate intervention and enhance outcome. The multivariate model derived may provide a template from which further research can elucidate a more clinically useful tool for the clinician.

© Lippincott-Raven Publishers.