Response to intracranial hypertension treatment as a predictor of death in patients with severe traumatic brain injury.

OBJECT: The normalization of increased intracranial pressure (ICP) in patients with severe traumatic brain injury (TBI) is assumed to limit secondary brain injury and improve outcome. Despite evidence-based recommendations for monitoring and treatment of elevated ICP, there are few studies that show an association between response to ICP-directed therapeutic regimens and adjusted mortality rate. This study utilizes a large prospective database to examine the effect of response to ICP-lowering therapy on risk of death within the first 2 weeks of injury in patients who sustained TBI and are older than 16 years.

METHODS: The current study is based on 1426 patients with severe TBI (Glasgow Coma Scale [GCS] score < 9) of whom 388 were treated for elevated ICP (> 25 mm Hg) between 2000 and 2008 at 22 trauma centers enrolled in a New York State quality improvement program. This prospectively collected database also contains information including age, admission GCS score, pupillary status, CT scanning parameters, and hypotension, which are all known early prognostic indicators of death. Treatment of elevated ICP consisted of administration of mannitol, hypertonic saline, barbiturates, and/or drainage of CSF or decompressive craniectomy. The factors predicting ICP response to treatment and predicting death at 2 weeks were evaluated using logistic regression analyses.

RESULTS: Increasing age and fewer hours of elevated ICP on Day 1 were found to be significant predictors (p = 0.001 and 0.0003, respectively) of a positive response to treatment. Response to ICP-lowering therapy (p = 0.03), younger age (p < 0.0001), fewer hours of elevated ICP (p < 0.0001), and absence of arterial hypotension on Day 1 (p = 0.001) significantly predicted reduced risk of death.

CONCLUSIONS: Patients who responded to ICP-lowering treatment had a 64% lower risk of death at 2 weeks than those who did not respond after adjusting for factors that independently predict risk of death.