Predictive value of cerebrospinal fluid parameters in neonates with intraventricular drainage devices

Abstract

Object: Infection is a common and potentially devastating complication following placement of ventriculoperitoneal (VP) shunts and cerebrospinal fluid (CSF) reservoirs in neonates. The goal of this study was to determine the normal ranges for cell count parameters in neonates with VP shunts and CSF reservoirs, as well as to determine the predictive value of CSF parameters as markers of infection.

Methods: The authors evaluated neonates from 150 different neonatal intensive care units of the Pediatrix Medical Group who had undergone a lumbar puncture, VP shunt insertion, or CSF reservoir placement between 1997 and 2004. Data were collected from 9704 neonates with a mean birthweight of 2573 g and a mean gestational age of 35 weeks. Of these neonates, 181 had VP shunt insertions or CSF reservoir placements.

Results: In neonates with negative CSF cultures, significant differences were found between those with and without VP shunts or CSF reservoirs when comparing red blood cell (RBC) count (620/mm' compared with 155/mm3, p < 0.05), absolute eosinophil count (4/mm3 compared with 2/mm3, p < 0.001), protein levels (179 mg/dl compared with 115 mg/dl, p < 0.001), and glucose levels (27.5 mg/dl compared with 49 mg/dl, p < 0.001). No significant difference was found between white blood cell (WBC) counts in neonates with or without VP shunts who had negative CSF cultures. The sensitivity and specificity of a cutoff value of 20 WBCs/mm3 for diagnosing meningitis in neonates with positive cultures and intraventricular drainage devices were 67% and 62%, respectively.

Conclusions: Although differences exist between CSF parameters found in neonates with or without VP shunts or CSF reservoirs, only the difference in RBC count is large enough to be clinically significant. The authors found that the utility of CSF parameters in neonates with VP shunts or CSF reservoirs was limited due to poor diagnostic sensitivity and specificity.