Moving the goalposts: A comparison of different definitions for primary external ventricular drain infection and its risk factors: A multi-center study of 2575 patients.

External ventricular drainage is the most common procedure performed in daily neurosurgical practice. One devastating complication is ventriculostomy-associated infection, but the establishment of evidence-based management guidelines has been hindered by the lack of an universal definition. There is also limited data with regard to the utility of comorbidity health indices and surgery-related factors in predicting infection. This study aims to compare the incidence of infection according to five commonly used definitions and to identify risk factors for this complication. 2575 patients from seven neurosurgical centers in Hong Kong underwent primary external ventricular drainage. The frequency of infection according to Gozal was 2.2% (n=57), 4.7% (Chi), 0.6% (Lozier), 0.8% (Lyke) and 2.8% (Scheithauer). The commonest pathogen was coagulase negative staphylococcus (39%) and 49% of all microbial isolates were multiple-drug resistant. The mean Charlson comorbidity index was 0.5±1.1. Using Gozal's definition as the primary endpoint, the index was not predictive of infection and no surgical risk factors were identified. The only significant risk factor was the performance of two or more additional neurosurgical procedures within 30days of catheterization (OR: 2.1, 95% CI 1.1-4.5). The rate of infection is relatively low, but considerable disparity exists depending on the definition used. Our data implies that patient factors, in particular the Charlson comorbidity index, and variations in surgical practice are less influential than the strict observance of infection control measures. The high incidence of antibiotic-resistant bacteria is concerning and the routine of exchange of catheters within 30days should be discouraged.