Is There an Age Cutoff for Intracranial Pressure Monitoring?: A Propensity Score Matched Analysis of the National Trauma Data Bank

Abstract

Background Despite mostly favorable past evidence for use of intracranial pressure monitoring (ICPM), more recent data question not only the indications but also the utility of ICPM. The Fourth Edition Brain Trauma Foundation guidelines offer limited indications for ICPM. Evidence supports ICPM for reducing mortality in patients with severe traumatic brain injury (TBI) and cites decreased survival in elderly patients. Methods All patients ≥ 18 years of age with isolated TBI, head Abbreviated Injury Scale (AIS) ≥ 3, and a Glasgow Coma Scale (GCS) ≤ 8 between 2008 and 2014 were included from the National Trauma Data Bank. Exclusion criteria were head AIS = 6 and death within 24 hours. Patients with and without ICPM were compared using TBI-specific variables. Patients were then matched via propensity-score matching (PSM), and the odds ratio (OR) of death with ICPM was determined using logistic regression modeling for 8 different age strata. Results A total of 23,652 patients with a mean age of 56 years, median head AIS of 4, median GCS of 3, and overall mortality of 29.2% were analyzed. After PSM, ICPM was associated with death beginning at the age stratum of 56-65 years. Intracranial pressure monitoring was associated with survival beginning at the age-group 36-45 years. Discussion Based on a large propensity-matched sample of TBI patients, ICPM was not associated with improved survival for TBI patients above 55 years of age. Until level 1 evidence is available, this age threshold should be considered for further prospective study in determining indications for ICPM.

Intracranial Pressure Monitoring in Patients With Severe Traumatic Brain Injury: Extension of the Recommendations and the Effect on Outcome by Propensity Score Matching

“…To deal with this problem, we found 6 studies that have performed adjusted logistic regression with the inclusion of the PS of receiving ICP monitoring into the model 6,[40][41][42]44,45 and 6 additional studies 8,17,43,[46][47][48][49] that have used PS matching. Even after these adjustments, the beneficial, 8,17,[41][42][43][44][45]47,49 null, 6,46 or detrimental effect 40,48 of ICP monitoring on the outcome measure of interest varied between studies. We noticed some limitations as to their epidemiological basis with the selection of patients being based on International Classification of Diseases coding or AIS head and so omitting important clinical and radiological information that could have compromised the adequate balance of confounding factors.…”

“…The main limitation of previous observational studies 12,13,[15][16][17][18][19] is the insufficient approach to deal with an imbalance in the baseline characteristics because patients undergoing ICP monitoring sustained more severe injuries. To deal with this problem, we found 6 studies that have performed adjusted logistic regression with the inclusion of the PS of receiving ICP monitoring into the model 6,[40][41][42]44,45 and 6 additional studies 8,17,43,[46][47][48][49] that have used PS matching. Even after these adjustments, the beneficial, 8,17,[41][42][43][44][45]47,49 null, 6,46 or detrimental effect 40,48 of ICP monitoring on the outcome measure of interest varied between studies.…”

“…We noticed some limitations as to their epidemiological basis with the selection of patients being based on International Classification of Diseases coding or AIS head and so omitting important clinical and radiological information that could have compromised the adequate balance of confounding factors. In addition, the study design can be controversial because of differences in prehospital care 45 and low rate of compliance with guidelines 41,42,45,46 along with the exclusion of specific patients such as those with extracranial injuries 41,48,49 or those monitored after craniotomy. 42 The key strengths of our study include the large number of patients and covariates recorded, the evaluation of extended indications, and the use of PS matching.…”