Abstracts

Rationale: Early seizures after traumatic brain injury (TBI) is a predictor of late seizures, i.e., post traumatic epilepsy (JAMA Neurol. 2022;79(4):334-341). Seizure following immediately after the acute injury may alter long-term seizure susceptibility and the progression to spontaneous seizures and thus epilepsy (Epilepsia. 2007;48(S2):65-74). However, human trials of antiseizure medication used after TBI (phenytoin, valproate, and carbamazepine) proved effective only against early seizures, but they did not reduce late seizures. (Epilepsia. 2001;42(4):515-524) Large register-based studies may be better powered to identify smaller effects. We propose using causal mediation analysis to study whether and to what extent the effect of TBI on epilepsy may operate through early seizures.

Methods: We performed a matched (5:1) cohort study including Danish residents registered with a moderate-to-severe TBI diagnosis in the Danish Nation Patient Register from 1995 to 2016. Diagnostic information from the register was used to identify moderate to severe TBI (ICD-10: S061-S069) and epilepsy (ICD-10: G40). Early seizures were characterized as seizures (ICD-10: R568, G40, G41) within two weeks of TBI (or index date among references). Study participants were followed from 14 days after TBI until onset of epilepsy, emigration, death, or the end of follow-up, whichever came first. Regression-based causal mediation analysis was applied to estimate the proportion of the effect of TBI on epilepsy that was mediated by early seizures. Adjustments were made for sex, age, calendar year, psychiatric and somatic disorders, and for the extent of cerebral lesions (focal, extracerebral, and diffuse injury).

Results: In the study period, 39,917 persons sustained a moderate or severe TBI and 188,450 were matched as references. Table 1 shows the characteristics of the study population; persons with TBI and references were comparable according to sex, age and entry date, but somatic and psychiatric conditions were more common among persons with TBI than references. During follow up, 802 (2.12%) persons with TBI had an early seizure, and 2,130 (5.62%) developed epilepsy. Among references, only 11 (0.01%) had an early seizure diagnoses, and 1,437 (0.76%) developed epilepsy. The total hazard ratio of epilepsy after moderate to severe TBI was 8.3 (95% CI, 7.6-9.1), of which the proportion mediated was 12.5% (95% CI, 11.8%-13.2%). A larger reference population is needed to conduct subgroup analyses.

Conclusions: Our results suggest that early seizures may be implicated in the development of epilepsy following TBI. Caution is required and more analyses are needed to substantiate the validity of the findings, i.e., the risk of residual confounding require further scrutiny, and sensitivity analyses must be conducted to ensure robustness.

Funding: This study was supported by the Novo Nordisk Foundation (grant numbers: NNF16OC0 019126 and NNF17OC0029860), the Central Denmark Region, and the Danish Epilepsy Association.