In this retrospective cohort analysis, 71% of patients with epilepsy secondary to non-accidental trauma achieved response to VNS by one year, compared to only 48% of those with other epilepsy etiologies. A non-significantly larger proportion of NAT patients had predominantly focal motor seizures at baseline (50% versus 25%), which may have contributed to this study’s observed effect size.
In the United States alone, epilepsy affects approximately 500,000 children [12]. Among these pediatric patients, 10–20% develop epilepsy secondary to trauma [13]. NAT is the suspected cause of over 50% of traumatic injury and mortality in infants alone [14]. It is also important to note that pediatric patients with NAT, in addition to higher mortality rates, have longer hospital stays and higher rates of long-term disability compared to non-NAT patients [15]. Paul et al. demonstrated that the risk of NAT increases with lower age [16]. In concordance with these findings, our results unsurprisingly demonstrated that the age at seizure onset for patients with NAT was significantly lower than their peers without NAT (0.3 vs. 3.3 years). Many mechanisms likely explain this age difference. Infants’ heads weigh more in relation to the rest of the body, their skulls are not fully developed, and their brains contain higher water content, making force-related injury more likely [17]. Furthermore, shearing forces are more likely to disrupt fragile bridging veins in infants than older children, increasing the likelihood of subdural hematoma formation and subsequent post-traumatic seizures [4, 18]. Interestingly, the age at seizure onset among NAT patients in this analysis coincides with a time just after the peak age for infantile colic [19]. Indeed, infantile colic has been associated with abusive head trauma [20], which may also explain the earlier age of seizure onset among NAT patients.
Posttraumatic epilepsy caused by nonaccidental trauma has been associated with higher resistance to ASMs [21] and a lower likelihood of successful surgical resection [22]. In such cases, VNS is often considered as a potential next-line treatment option. Surprisingly, despite the earlier age of seizure onset among NAT patients, both NAT and non-NAT patients underwent VNS implantation at similar ages. Evidence suggests that longer duration of epilepsy before VNS insertion is associated with worse VNS response rates [23]. This may highlight a need for earlier consideration of VNS for NAT patients with non-resectable, DRE.
A non-significantly higher proportion of patients with NAT had predominantly focal motor seizures at baseline (50% versus 25%). Interestingly, the prognosis for drug-resistant and unresectable focal seizures is mixed. From the VNS literature, many studies have demonstrated that focal seizures are more response to VNS [23]. Moreover, VNS is currently only FDA-approved for the treatment of focal seizures. However, Englot et al. recently demonstrated that pediatric patients with generalized seizures had 1.36 times the odds of seizure freedom than those with non-generalized seizures [24]. Englot’s group also found similar results in a prior analysis [25]. The reasons for these mixed results are unclear, though is possible that the focal seizures which ultimately go on to receive VNS are the most treatment-resistant cases (those not controlled by ASMs and those which cannot be surgically-resected), while even patients with mild primarily generalized epilepsy often cannot undergo surgical resection. Therefore, there may be a skewed distribution in the baseline severity of epilepsy among VNS patients in some studies. In our sample of NAT patients, most had focal motor seizures (50%) and generalized motor seizures (37%), with 71% of the cohort developing a good response to VNS by one year.
It is worth noting that the difference in VNS response rates between our NAT and non-NAT cohorts was not significant. Still, the effect size was large: we demonstrated a 23% difference in response rates between groups. It is likely that the small sample size of the NAT cohort limited the ability for our analysis to detect this effect size. We performed a post-hoc power analysis, which revealed that in order to have detected this effect size (with alpha = 0.05 and power = 0.8), we would have needed at least 37 NAT patients.
With that said, the overall results of our analysis are consistent with existing literature. Patients with posttraumatic epilepsy have been evidenced to respond better to VNS than those with other seizure etiologies. Englot et al. demonstrated that 78% of posttraumatic epilepsy patients experienced ≥ 50% reduction in seizure frequency after 24 months of VNS treatment compared to only 61% of nontraumatic epilepsy patients [26]. These results underscore the need for future, longer-term and larger-scale prospective trials, informed by pre-hoc power calculations, to more reliably demonstrate the effectiveness of VNS for patients with NAT-related epilepsy.
Limitations
There are several limitations of our analysis which ought to be carefully considered. As a retrospective cohort study, our analysis is limited by the same factors which limit all observational retrospective studies. Primarily, information bias and recall bias are notable limitations of this analysis. All seizure frequency data was based on parent report, which is, by nature, subjective and can vary from parent-to-parent. To minimize this observer-related bias, we utilized the primary treating neurologists’ opinion whenever possible for more consistent reporting of seizure frequency. Another major limitation of this analysis is the small sample size of our NAT cohort. For this reason, the power of our study was limited, and we were unable to detect even large effect sizes. Future, larger-scale, prospective studies with more NAT patients should be performed to validate the results of our findings and similar results in the literature.