Hawryluk GWJ, Nielson JL, Huie JR, et al. Analysis of Normal High-Frequency Intracranial Pressure Values and Treatment Threshold in Neurocritical Care Patients: Insights into Normal Values and a Potential Treatment Threshold. JAMA Neurol. Published online June 15, 2020. doi:10.1001/jamaneurol.2020.1310
Reviewed by Sanjeev Sivakumar, MD
Read the article.*
*You will need to log in to US National Library of Medicine National Institutes of Health to read this article.
The magnitude and duration of intracranial pressure (ICP) elevation closely correlates with clinical outcomes following traumatic brain injury (TBI), but the exact threshold that exerts a detrimental effect remains unclear.
To better delineate the association between ICP thresholds and patient outcomes, the authors used high-frequency physiologic data from a prospective cohort of patients admitted to San Francisco General Hospital from 2004-2010. TBI patients were treated in line with the BTF guidelines, with an ICP treatment threshold of 20 mmHg. Authors used a computer system that automatically collected one-minute physiologic data in patients for the entirety of their ICU stay. If patients received two ICP monitors, the average was analyzed. Mean ICP values were calculated, and authors defined and analyzed ICP’s greater than 79 different thresholds, from 1 to 80 mmHg in 1-mmHg increments using MATLAB software. The number of one-minute epochs with ICP values greater than a specific threshold and between specified periods were calculated. Analysis of Variance (ANOVA) was used to analyze mean ICP values from continues data with post hoc tests adjusting for multiple comparisons, and logistic regression was used for categorical data.
A generalized estimating equation (GEE) was used to estimate parameters of a linear model, with possible unknown correlation between ICP thresholds and outcomes. Principal components analysis (PCA) was used to assess associations between heart rate, BP, PbtO2 and other physiologic variables, and 15 binary ICP threshold categories. Regression was used to assess relative contribution of each ICP threshold (between 15-30 mmHg) in predicting the Glasgow Outcome Score (GOS) at discharge. The ICP threshold identified was then used in another regression model with other independent predictors to determine the strength for each ICP threshold in predicting clinical outcomes.
A total of 4,090,964 one-minute ICP measurements were recorded from 523 patients, and 73% of all patients had TBI. TBI patients were younger and more likely to be male. Of those without TBI, majority (53%) had hemorrhagic stroke, 19.3% had subarachnoid hemorrhage, 5.7% had brain tumor and 4.3% each for ischemic stroke and anoxic brain injury.
There was a significant correlation between poor functional outcomes and older age, lower post-resuscitation GCS scores and longer hospital stays. The most common ICP measurements at two different time frames (from days 3 to 5 and days 1 to 30) were 8 mmHg and 9 mmHg in all patients and in TBI patients, respectively. The mean ICP values over time were higher among TBI patients when compared to other conditions, and higher ICP values on days 2 and 3 were associated with increased mortality risk. For all patients, between days 3 to 5, there was a significant difference in outcome across ICP thresholds. PCA demonstrated that at ICP thresholds of 19 and 24mm, there were distinct changes in the physiome analyzed, including changes in MAP, HR and PbtO2. After controlling for age, sex, type of injury, diffuse vs. focal TBI, and irrespective of craniotomy/craniectomy status, elastic net regression identified the ICP threshold of 19mmHg as significantly predictive of discharge GOS in patients with TBI and other patients.
Using an automated computed collection of high-frequency patient measurements of ICP, this study found that ICP above 19mmHg was strongly associated with outcome and that lower ICP was associated with improved outcome. Moreover, given the number of readings, they were able to add insight into normal ICP values, which generally ranged from 8-9mmHg.
Interestingly, ICP thresholds did not seem to be different across various brain pathology, comparing TBI and non-TBI injury, which goes against some prior data that has argued for disease-specific ICP targets. The authors also identify some patients with poor outcomes where ICP is as low as 10mmHg, suggesting that ICP alone is insufficient to predict outcome. The degree to which altering ICP targets in these patients could improve outcome remains unknown.
The major strength of this study is that unlike prior research that looked at one ICP value associated with poor outcome, this study looked at millions of readings across multiple patients to determine the optimum threshold across a broad range. The limitations of this study are the lack of standardized treatment and the single-center nature of the study. Given the strong relationship between ICP threshold and mortality but not outcome, it remains to be seen how much of high ICP is a reflection of the underlying severity of injury. Future studies that focus on therapeutic goals can further our understanding of the relationship of ICP and outcome.