Ventriculostomy-associated infections: incidence and risk factors

BACKGROUND: The objective of this study was to assess the incidence of ventriculostomy-associated infections (VAI) and to examine the related risk factors. METHODS: Data on all consecutive patients with ventriculostomy catheters admitted to the intensive care unit (ICU) in a tertiary care center over a 3-year period were identified from the ICU database and from medical records. VAI was documented using a preset definition. The following patient data were documented: demographics, severity of illness measures, indication for the catheter, presence of cerebrospinal fluid (CSF) leak, and length of stay and mortality. The following catheter data were collected: venue of catheter placement (operating room or nonoperating room areas), use of antibiotic irrigation and prophylactic systemic antibiotics, and number of catheter days. The frequency of CSF sampling was documented. RESULTS: In 84 patients, 99 catheters were placed, of which 19% developed VAI. There was a total of 586 catheter days (infection rate, 32 per 1000 catheter days). The risk of VAI increased steadily until catheter day 7 then reached a plateau. Among patients' factors, repeat catheter insertion was associated with a significant increase in VAI. There were no significant associations with age, severity of illness, indication for the catheter, craniatomy, or presence of CSF leak. Among catheter factors, the number of catheter days and repeat catheter insertion emerged as significant independent predictors on multivariate analysis. Placement outside the operating room was associated with a trend toward higher VAI. The use of prophylactic antibiotic or antibiotic irrigation did not significantly alter VAI rates. Routine surveillance cultures of CSF were no more likely to detect infection than cultures obtained when clinically indicated. Gram-negative bacilli were responsible for 50% of the infections, followed by gram-positive cocci (29%) and others (21%). CONCLUSIONS: The risk of VAI increases with increasing duration of catheterization and with repeated insertions. The use of local antibiotic irrigation or systemic antibiotics does not appear to reduce the risk of VAI. Routine surveillance cultures of CSF were no more likely to detect infection than cultures obtained when clinically indicated. These findings need to be considered in infection control policies addressing this important issue.