Authors :
Presenting Author: James Chen, MD, PhD – VAGLAHS/UCLA
Cindy Le, BS – VAGLAHS
Kathleen Cui, BA – VAGLAHS
Olga Alexeeva, MD – VAGLAHS/UCLA
Janice Joo, MD – VAGLAHS/UCLA
Julia Bailey, PhD – VAGLAHS/UCLA
Rationale:
Post-traumatic epilepsy (PTE) could be a lifelong complication after mild or moderate traumatic brain injury (TBI). PTE is often encountered in veterans with epilepsy. Despite multiple risk factors, such as severity of TBI, that have been identified to associate with the development of PTE, there is still no mechanistic understanding of the epileptogenesis in PTE. We have proposed that the development of PTE requires the first hit, the genetic predisposition, and the subsequent second hits, the TBIs, to initiate the process of epileptogenesis. Our new cohort data supports this hypothesis. It is plausible that the PTE is a polygenic disorder. Based on this understanding, we have developed a model framework to calculate the odds ratio for the development of PTE.
Methods:
In the VA, we recruited a cohort of PTE patients vs. a control cohort of TBI patients without epilepsy and performed whole-exome sequencing (WES). The demographics of the PTE cohort (N=28) and TBI cohort (N=22) were compatible. The identified 375K variants from each subject were matched with the 15K verified epilepsy variants identified from the ClinVar database. Fisher's exact test was computed to identify variants that are PTE related. The contingency table of each PTE-related variant was computed to obtain the odds ratio of developing PTE and to differentiate the matched variants as PTE-prone or PTE-protective. A ROC curve with the AUC was calculated to evaluate the accuracy of this model for predicting PTE.
Results:
1. The cohort selection was satisfactory. The statistical chance of including a false negative PTE subject in the TBI cohort is exceptionally low.
2. 30 PTE-prone variants and 57 PTE-protective variants were identified, with their respective odds ratios computed. The risk of developing PTE in a subject can be calculated by combining the odds ratios of the PTE-related variants. The odds ratios identified in epidemiological studies can be further incorporated into the risk calculation using this framework.
3. The ROC curve showed an excellent AUC of 94.7%, suggesting that it is promising as a clinically acceptable model for risk estimation.
Conclusions:
1. Our data support the two-hit hypothesis for developing PTE. PTE is conceivably a polygenic disorder.
2. We have identified multiple PTE-related variants in our cohorts. However, our list of variants might not be exhaustive.
3. We have developed a framework using a WES-based tool to estimate the risk of developing PTE. The risk can be calculated either before or after TBI. A translational study with larger cohorts can validate this model and further refine the tool for clinical application.
Funding: