Simulating cerebral edema and delayed fatality after traumatic brain injury using triphasic swelling biomechanics.

Objectives: Contemporary finite element (FE) models, like that from the Global Human Body Models Consortium (GHBMC), have been useful for developing safety systems to reduce the severity of injuries in motor vehicle crashes (MVCs), including traumatic brain injury (TBI). However, not all injury occurs during the MVC. Cerebral edema after TBI contributes to mortality by increasing intracranial pressure (ICP) and preventing adequate cerebral blood supply. The focus of this study was to model post-traumatic cerebral edema and subsequent mortality due to increased ICP.Methods: Brain tissue swells in a manner consistent with triphasic biomechanics, which models biological tissues as a charged deformable porous solid matrix (fixed charge density [FCD]), a solvent, and monovalent counter-ions (cerebrospinal fluid). Fluid uptake into the brain is driven by the Gibbs-Donnan osmotic pressure as the FCD is exposed when cells die. Post-TBI edema was simulated in FEBio (febio.org), which includes triphasic material formulations.The GHBMC mesh was imported into FEBio, and each element was assigned a FCD to represent impact-related cell death based on its maximum principal strain (MPS) experienced during the crash-simulation using the stock GHBMC model and LS-DYNA. The ensuing pathophysiology was simulated in FEBio in two steps. First, the brain swelled in response to exposure of FCD, causing some adjacent elements to compress as fluid was redistributed. Biologically, the compression was assumed to reduce blood flow and cause ischemic cell death, represented by additional exposure of FCD, swelling, and increased ICP. Using published prognostic models of clinical outcome, mortality was predicted based on ICP.Results: Post-traumatic volume ratio of elements ranged from less than 30% (compaction) to greater than 200% (swelling). Predicted ICP values for a fatal impact were as high as 8.55 kPa (64.1 mmHg), which is associated with a 99% probability of death.Conclusion: To the best of our knowledge, this is the first study to simulate post-traumatic brain swelling to predict outcome. By incorporating swelling, ischemia, and cell death, our novel approach may improve fidelity of predicting outcome after MVCs. A strength of our approach is relying on the validated GHBMC model to predict brain deformation in the crash-scenario. The main goal of the current study was to demonstrate feasibility of simulating post-injury swelling using triphasic biomechanics. We successfully predicted clinically relevant increases in ICP that suggest a high likelihood of death when simulating a fatal impact scenario, however, more validation of our methodology is needed.

Prediction of probability of fatality due to brain injury in traffic accidents.

Objective: Fatal brain injuries result from physiological changes in brain tissues, subsequent to primary damage caused by head impact. Although efforts have been made in past studies to estimate the probability of brain injury, none of them involved prediction of such physiological changes. The goal of this study was to evaluate the fatality prediction capability of a novel approach that predicts an increase in intracranial pressure (ICP) due to primary head injury to estimate the fatality rate using clinical data that correlate ICP with fatality rate. Methods: A total of 12 sets of head acceleration time histories were used to represent no, severe, and fatal brain injury. They were obtained from the literature presenting head kinematics data in noninjurious volunteer sled tests or from accident reconstruction for severe and fatal injury cases. These were first applied to a Global Human Body Models Consortium (GHBMC) head-brain model to predict nodal displacement time histories of the brain, which were then fed into FEBio to predict ICP. A Weibull distribution was applied to the data for the relationship between fatality rate and ICP obtained from a clinical paper to estimate fatality rate from ICP (procedure A). Fatality rate was also estimated by applying the temporal and spatial maximum value of maximum principal strain (MPSmax) obtained from the GHBMC simulation to an injury probability function for MPSmax (procedure B). Estimated fatality rates were compared between the 2 procedures. Results: Both procedures estimated higher average fatality rate for higher injury severity. The average fatality rate for procedure A without ischemia representation and procedure B was 72.4 and 51.0% for the fatal injury group and 8.2 and 21.7% for the severe injury group, respectively, showing that procedure A provides more distinct classification between fatal and nonfatal brain injury. It was also found that representation of ischemia in procedure A provides results sensitive to injury severity and impact conditions, requiring further validation of the initial estimate for the relationship between brain compression and ischemic cell death. Conclusions: Prediction of the probability of fatality by means of a combination of simulations of the primary brain deformation and subsequent ICP increase was found to be more distinct compared to the prediction of primary injury alone combined with the injury probability function from a past study in the select 12 head impact cases.