7026

Predictors of Ventricular Shunt Infections in Children

Friday, October 16, 2009: 5:45 PM
207A (Washington Convention Center)
Elisabeth J. Ashley, MD1, Amir Kimia, MD1, Joseph R. Madsen, MD2, Lise E. Nigrovic, MD, MPH1 and Mark I. Neuman, MD, MPH1, (1)Emergency Medicine, Children's Hospital of Boston, Boston, MA, (2)Neurosurgery, Children's Hospital of Boston, Boston, MA

Purpose: Data regarding clinical predictors of ventricular shunt infection in children are limited, and no data exist regarding optimal management strategies for children who present to the emergency department (ED) with concern of a shunt infection. Among a population of children with a ventricular shunt presenting to a pediatric ED, and in whom CSF was obtained, we sought to: 1) determine the rate of positive CSF bacterial culture, and 2) identify clinical predictors of ventricular shunt infection.

Methods: : We performed a retrospective cohort study of children 3 months to 21 years of age evaluated in a single pediatric tertiary ED between 1995-2008 for possible ventricular shunt infection. All included children had cerebrospinal fluid (CSF) obtained within 24 hours of presentation to the ED. Clinical and laboratory information were abstracted from the medical record. A shunt infection was defined by growth of bacteria in the CSF of a child who underwent shunt removal or externalization due to concern of shunt infection. Multivariate logistic regression was performed to identify clinical predictors of ventricular shunt infection.

Results: : 773 children met inclusion criteria; median age was 7.0 years (IQR 2.1, 13.7).  101 (13.1%) patients with a ventricular shunt had growth of bacteria in CSF; of which 48 (6.2% of total) underwent shunt removal or externalization for suspected infection. The median time since last shunt revision or replacement was shorter for patients with a shunt infection compared to children without shunt infection (45 days vs. 197 days, p< 0.001).  In multivariate analyses, factors associated with shunt infection include: history of fever (odds ratio [OR] 7.7, 95% CI  4.1-14.6), abdominal pain (OR 5.4, 95% CI 1.9-15.5), and shunt revision within the prior 3 months (OR 3.2, 95% CI 1.7-6.1). The presence of headache, vomiting, seizure, and lethargy were not associated with shunt infection.

Conclusions: :  Among children with a ventricular shunt who have CSF obtained in the ED, 6% have a shunt infection.  The presence of fever, abdominal pain, and recent shunt revision are associated with an increased risk of shunt infection.