Effectiveness of Pharmacological Therapies for Intracranial Hypertension in Children With Severe Traumatic Brain Injury--Results From an Automated Data Collection System Time-Synched to Drug Administration

Abstract

Objectives: To describe acute cerebral hemodynamic effects of medications commonly used to treat intracranial hypertension in children with traumatic brain injury. Currently, data supporting the efficacy of these medications are insufficient.

Design: In this prospective observational study, intracranial hypertension (intracranial pressure ≥ 20 mm Hg for > 5 min) was treated by clinical protocol. Administration times of medications for intracranial hypertension (fentanyl, 3% hypertonic saline, mannitol, and pentobarbital) were prospectively recorded and synchronized with an automated database that collected intracranial pressure and cerebral perfusion pressure every 5 seconds. Intracranial pressure crises confounded by external stimulation or mechanical ventilator adjustments were excluded. Mean intracranial pressure and cerebral perfusion pressure from epochs following drug administration were compared with baseline values using Kruskal-Wallis analysis of variance and Dunn test. Frailty modeling was used to analyze the time to intracranial pressure crisis resolution. Mixed-effect models compared intracranial pressure and cerebral perfusion pressure 5 minutes after the medication versus baseline and rates of treatment failure.

Setting: A tertiary care children's hospital.

Patients: Children with severe traumatic brain injury (Glasgow Coma Scale score ≤ 8).

Interventions: None.

Measurements and main results: We analyzed 196 doses of fentanyl, hypertonic saline, mannitol, and pentobarbital administered to 16 children (median: 12 doses per patient). Overall, intracranial pressure significantly decreased following the administration of fentanyl, hypertonic saline, and pentobarbital. After controlling for administration of multiple medications, intracranial pressure was decreased following hypertonic saline and pentobarbital administration; cerebral perfusion pressure was decreased following fentanyl and was increased following hypertonic saline administration. After adjusting for significant covariates (including age, Glasgow Coma Scale score, and intracranial pressure), hypertonic saline was associated with a two-fold faster resolution of intracranial hypertension than either fentanyl or pentobarbital. Fentanyl was significantly associated with the most frequent treatment failure.

Conclusions: Intracranial pressure decreased after multiple drug administrations, but hypertonic saline may warrant consideration as the first-line drug for treating intracranial hypertension, as it was associated with the most favorable cerebral hemodynamics and fastest resolution of intracranial hypertension.

Figure 1. ICP and CPP before and after initiation of ICP-directed therapy

For each medication administration, mean ICP and CPP (mm Hg) were calculated for each epoch relative to initiation of the bolus. Error bars denote interquartile range. Asterisks denote median values significantly different than the 5 minutes preceding initiation of the bolus by Kruskal-Wallis ANOVA with Dunn's test. Figures A and B include all epochs. Epochs confounded by administration of the subsequent medication were censored from Figures C and D.

Figure 2. ICP and CPP before and after conclusion of ICP-directed therapy

For each medication administration, mean ICP and CPP (mm Hg) were calculated for each epoch relative to conclusion of the bolus. Error bars denote interquartile range. Asterisks denote median values significantly different than the 5 minutes preceding initiation of the bolus by Kruskal-Wallis ANOVA with Dunn's test. Figures A and B include all epochs. Epochs confounded by administration of the subsequent medication were censored from Figures C and D.

Figure 3. Changes in ICP and CPP from baseline to 5min after the medication

The boxes represent the mean difference between the baseline value and the value 5min after the medication. Error bars represent standard error of the mean. Changes in ICP (shaded boxes) are plotted on the left y-axis (with negative values above the x-axis). Changes in CPP (open boxes) are plotted on the right y-axis. Asterisks indicate mean differences that are significantly different than zero (p < 0·05). Figure 3a shows data relative to initiation of the medication; Figure 3b shows data relative to completion of the medication.

Figure 4. Kaplan-Meier curve showing time until resolution of the intracranial hypertension episode

This is a visual representation of the time to ICP crisis resolution for each treatment group. Resolution was defined as ICP < 20 mm Hg for the subsequent 5 minutes.