Diagnostic criteria of CNS infection in patients with external ventricular drainage after traumatic brain injury: a pilot study.

Abstract

Ventriculostomy-related infection (VRI) is a common complication in patients with traumatic brain injury (TBI) treated with an external ventricular drain (EVD). The aim of this study was to investigate incidence and characteristics of patients with VRI, and to explore diagnostic criteria to confidently rule out VRI in patients with TBI. This retrospective cohort pilot study included adults with severe TBI who were admitted to the ICU and received an EVD, during a 26-month period. Patients were categorized as having Culture-positive VRI, Culture-negative VRI, or No VRI. Variables that were potentially associated with Culture-positive VRI was analyzed, and predictive values were calculated. 75 of 215 patients with severe TBI (35%) underwent EVD placement; nine of these (12%) were classified as Culture-negative VRI and eight (11%) as Culture-positive VRI. The CSF cell counts that led to VRI treatment were compared with 46 CSF cell counts from No VRI patients. A CSF/plasma glucose ratio below 0.6had a negative predictive value (NPV) for culture-verified VRI of 0.97 (95% CI: 0.85-1), whereas a combination of three CSF-derived biomarkers within the reference limits (white/red blood cell ratio, CSF/plasma glucose ratio, and protein content) ruled out Culture-positive VRI in this cohort (PPV 0, 95% CI: 0-0.14). C-reactive protein did not reliably predict VRI. In this pilot study of patients after severe, a combination of biomarkers within reference limits ruled out VRI (PPV 0, CI: 0-0.14). Hypoglycorrhachia was a sensitive marker of VRI (NPV 0.97, CI: 0.85-1). Systemic signs and markers of infection did not predict VRI.