Original Article: Harder, Tyler J., et al. "Early Signs of Elevated Intracranial Pressure on Computed Tomography Correlate With Measured Intracranial Pressure in the Intensive Care Unit and Six-Month Outcome After Moderate to Severe Traumatic Brain Injury." Journal of Neurotrauma (2023).

Background/Methods:

Traumatic brain injury (TBI) poses significant challenges in treatment and prognostication due to the diversity of trauma phenotypes and complex physiology. Although the Marshall and Rotterdam scores based on initial CT findings are commonly employed for mortality prediction, they have limited capacity in predicting functional outcomes. As part of the standard care for comatose TBI patients, intracranial pressure (ICP) measurements have become crucial, with an ICP threshold of >20 mmHg being correlated with both functional outcomes and mortality. In this article, the authors explored the potential of enhancing predictive models for 6-month outcomes in TBI by integrating imaging biomarkers of elevated ICP with ICP measurements.

Study data were extracted from the ProTECT III trial which included patients with non-penetrating moderate-to-severe TBI. The primary outcome measures were composite functional outcomes evaluated using the Glasgow Outcome Scale extended (GOS-E) and the Disability Rating Scale (DRS) at 6 months post-TBI. Poor outcome was defined as a DRS score of ≥6 or a GOS-E score of ≤6. Five imaging variables associated with ICP elevation were abstracted: sulcal obliteration, lateral ventricular compression, compression of the third ventricle, presence of midline shift, and visualized herniation. A multivariate logistic regression was used to assess the association between clinical and radiographic variables and the outcome measures. Multiple linear regression was employed to analyze continuous variables. Receiver operating characteristic (ROC) curves were generated to evaluate the accuracy of CT markers of elevated ICP, as well as the Marshall and Rotterdam scores, as predictors of the outcomes under consideration

Results:

A total of 882 patients were included in the study, with a median age of 35 years and the majority being male (79.5%). Among them, 354 patients (40%) underwent invasive intracranial pressure (ICP) monitoring, and 13.4% of these patients experienced ICP transgression. CT scans revealed the presence of various signs associated with elevated ICP: third ventricular compression in 24% of patients, lateral ventricle compression in 23.3%, sulcal obliteration in 22.2%, midline shift in 22%, and herniation in 18%.

Patients without signs of increased ICP on CT were found to have a lower likelihood of mortality (OR 0.183, p < 0.001). There was a positive correlation between the number of CT signs and poor outcomes. Patients exhibiting all five CT signs had a shorter time to death (mean difference -17 days, p=0.015), while those without any CT biomarkers had a longer time to death (mean difference 14.4 days, p=0.016). Having 4 (OR 2.648, p < 0.001) or 5 (OR 5.644, p < 0.001) CT biomarkers was associated with increased odds for poor 6-month outcomes.

Out of the 354 patients with ICP data, mortality was higher with increasing maximum ICP and a higher proportion of hours with positive ICP transgression (OR 64.99, p < 0.001). Patients with all five CT biomarkers had greater odds of experiencing poor functional outcomes (OR 4.44, CI 1.51-14.18, p=0.009) but not mortality (p=0.149). Hemorrhage location in the posterior fossa (OR 3.2, p=0.029) or basal ganglia (OR 3.1, p=0.021), and hematoma volume >30 cc (OR 3.7, p=0.003) were associated with mortality. Poor functional outcomes were observed with frontal lobe hemorrhage (OR 2.95, p=0.013), traumatic subarachnoid hemorrhage (OR 2.2, p=0.034), and intraventricular hemorrhage (OR 2.25, p=0.019). Requiring craniotomy (OR 4.3, p=0.019), hemicraniectomy (OR 2.99, p=0.031), and an increased number of interventions were associated with higher odds of mortality, but time to surgical intervention was not.

The accuracy of CT biomarkers, Marshall and Rotterdam scores in predicting mortality were evaluated with area under the curve (AUC) analyses, with AUC values of 0.714, 0.754, and 0.720, respectively. All three metrics demonstrated better predictive capacity for mortality than for functional outcomes, with AUC values for the latter of 0.624, 0.637, and 0.647, respectively.

Commentary

This study identified several significant predictors of mortality after TBI, including maximum ICP, proportion of time with ICP elevation, hematoma volume greater than 30 ml, and hemorrhage locations in the basal ganglia and posterior fossa. On the other hand, predictors of 6-month outcomes were found to be SAH, IVH, and frontal hemorrhage. Sulcal effacement and third ventricle compression were found to be the best predictors of higher maximum ICP. Additionally, higher average ICP was associated with visualized herniation, lateral ventricular compression, and third ventricular compression. It is worth noting that 67% of the ProTECT III trial cohort did not exhibit any of the five CT signs at baseline. Nevertheless, previous studies have indicated that a combination of baseline head CT findings with sustained ICP elevation is strongly associated with poor outcomes in TBI patients. This study contributes to the existing literature by emphasizing specific CT patterns that hold potential for prognostication and can potentially guide treatment decisions in TBI management; however, their ability to predict functional outcomes remains relatively low.