Intracranial pressure monitoring and inpatient mortality in severe traumatic brain injury: A propensity score-matched analysis

BACKGROUND: Although intracranial pressure (ICP) monitoring in severe traumatic brain injury (TBI) is recommended by the Brain Trauma Foundation, the benefits remain controversial. We sought to determine the impact of ICP monitor placement on inpatient mortality within a regional trauma system after correcting for selection bias through propensity score matching. METHODS: Data were collected on all severe TBI cases presenting to 14 trauma centers during the 2-year study period (2009-2010). Inclusion criteria were as follows: blunt injury, Glasgow Coma Scale (GCS) score of 8 or lower in the emergency department, and abnormal intracranial findings on head computed tomography (CT). Two separate multivariate logistic regression models were used to predict ICP monitor placement and inpatient mortality after controlling for demographics, severity of injury, comorbidities, and TBI-specific variables (GCS score, pupil reactivity, international normalized ratio, and nine specific head CT findings). To account for selection bias, we developed a propensity score-matched model to estimate the "true'' effect of ICP monitoring on in-hospital mortality. RESULT: A total of 844 patients met inclusion criteria; 22 died on arrival to the emergency department. Inpatient mortality was 38.8%; 46.0% of the patients underwent ICP monitor placement. Unadjusted mortality rates were significantly lower in the ICP monitoring group (30.7% vs. 45.7%, p < 0.001). ICP monitor placement was positively associated with CT findings of subdural hematoma, intraparenchymal contusion, and mass effect and negatively associated with age, alcoholism, and elevated international normalized ratio. After adjusting for selection bias via propensity score matching, ICP monitor placement was associated with an 8.3 percentage point reduction in the risk-adjusted mortality rate. CONCLUSION: ICP monitor placement occurred in only 46% of eligible patients but was associated with significantly decreased mortality after adjusting for baseline risk profile and the propensity to undergo monitoring. As the individual impact of ICP monitoring may vary, future efforts must determine who stands to benefit from invasive monitoring techniques. (Copyright (C) 2015 Wolters Kluwer Health, Inc. All rights reserved.)