Ventriculostomy related infection in intensive care unit: Diagnostic criteria and related conditions

Abstract

ObjectiveTo evaluate the usefulness of clinical signs, blood tests, microbiological cultures and cerebrospinal fluid (CSF) analysis to detect ventriculostomy related infections (VRI), and to describe related conditions. MethodsA retrospective study was carried out including all patients with external ventricular drain admitted to intensive care unit from January 2000 to December 2006. Diagnosis of VRI, mortality, demographic and clinical data, time and number of drains, microbiological and biochemical CSF results and blood test were recorded. Difference between infected and uninfected patients was statistically significant at P < 0.05. ResultsThe results revealed 136 drainages in 120 patients with 22 (18.33%) infected (15.39 infections per 1000 days of drainage). This group was on overage older, had more severe systemic response syndrome and a significantly higher number of drains and longer duration of drain insertion. We found statistical differences in proteinorrachia, glycorrhachia, and glycorrachia/glycemia ratio during 8.5-day drain insertion (interquartile range 7–10.25). A total of 31 cultures were positive in patients without VRI and 47 were negative in patients with VRI. Furthermore, 35 patients died (2 belonging to the infected group). Significantly higher risk of VRI in intraventricular fibrinolysis and subarachnoid haemorrhage was observed. We made a multivariate regression model resulting in a prediction rule with 55.7% area under curve (95% CI 0.43–0.70). ConclusionsCSF routine cultures and biochemical studies are not recommended to diagnose VRI. Clinical signs, external ventricular drain manipulation and a drainage insertion over a week justify the routine measurement of proteinorraquia, glycorrhachia and the ratio of glycorrachia/glycemia.