gms | German Medical Science

Objective: External ventricular drainages (EVD) are essential in the care of patients with elevated intracranial pressure and with acute hydrocephalus. Traumatic brain injury (TBI) belongs to the rare indications for an EVD. Nevertheless, ventriculitis in these patients is a dreaded life threatening complication. In this retrospective study, we focused on patients with TBI whounderwent a ventriculostomy or an intraparenchymatous monitoring of ICP prior to ventriculostomy. Our objective is to define the risk of developing a ventriculitis in this patient population.

Methods: 1,238 patients, who underwent an external ventriculostomy in our institute in the period of 2003 to 2009 were evaluated. 799 patients met the following inclusion criteria: 1) >15 years, 2) no preexisting EVD or shunt system, and 3) any known infection of CNS/CSF at admission. 61 of 799 patients suffered from a TBI (7.6%). Demographic data, comorbidities, medical complications, procedure-related data and the laboratory values were assessed. Furthermore, analysis of the microbiological parameters, co-infections and pathogens thereof, detection of a particular ventriculitis-pathogen in other body fluids, systemic inflammatory parameters and CSF parameters were acquired. Ventriculitis was defined as at least two positive bacteriological cultures. In comparison, contamination was defined as a single positive culture.

Results: 12 of 61 patients had an open head TBI (19.7%). Of these, 4 patients demonstrated CSF flow. A total number of 78 EVDs were inserted. Catheter days amounted to 660 days with a median duration of 9 days (1–33days). In 6 (9.8%) patients. a ventriculitis was diagnosed, two were openTBI, one with CSF flow. There was a single case of contamination. There was no correlation between infection rate and gender, age, preceding craniotomy or ICP monitoring. Furthermore, there was no correlation with common comorbidities such as pneumonia or diabetes mellitus. However, the incidence of ventriculitis was significantly higher in patients with brain edema (p=0,005).

Conclusions: In the present retrospective series of EVD-related infections in TBI, only severe brain edema showed a significant correlation, most likely due to the higher number of procedures per case in this severely ill patient population. Compared to the total population of 799 EVDs, TBI had a significantly higher ventriculostomy-related risk for infections, namely 9.8% in TBI vs. 3.3% in the remaining cases.