Vagus nerve stimulation in medically refractory epilepsy: Adverse effects and clinical correlates

Discussion

This study specifically focused on evaluating the adverse effects of VNS implantation. The patient population included both pediatric and adult patients, with a median age of 14 years. The classification of epilepsy in our patients was based on the ILAE definition of medically intractable epilepsy, with a median duration of approximately 12 years before VNS implantation.¹⁵

Previous studies have demonstrated that VNS is an effective adjunctive treatment for medically refractory epilepsy, particularly in cases where resective surgery is not feasible.^7,16,17 The VNS therapy offers a tolerable treatment option that effectively improves the quality of life for patients.^18,19 Approximately 30% of epilepsy patients are classified as medically intractable.²⁰ In our study, approximately 79% (n=34) of patients had generalized epilepsy, for which VNS served as a palliative treatment option. For the remaining patients with focal epilepsy, VNS was employed as a palliative measure after epilepsy surgery failed to achieve optimal seizure control or when the seizure onset zone was located in the eloquent cortex. Epilepsy surgery was performed on 19 patients with focal and generalized epilepsy. The most frequent procedure was corpus callosotomy, conducted on 15 out of 19 patients (78.95%) to address atonic and tonic seizures. Other procedures, including functional hemispherectomy, frontal disconnection, posterior quadrant disconnection, and temporal lobectomy, were performed in one case each. The VNS was implanted after epilepsy surgery proved insufficient in adequately controlling seizures. The combination of epilepsy surgery followed by VNS might have contributed to higher rates of VNS therapy efficacy in our study population compared to the median seizure reduction reported in other studies.^3,7 Regarding epilepsy etiology within our study population, 21 patients (48.8%) had an idiopathic etiology, followed by seven patients (16.2%) with a genetic etiology, as shown in Table 1 and Table 2. Epilepsy in the patients included in this study was often comorbid with cognitive impairment or GDD (n=30).

Adverse effects are anticipated in patients following VNS therapy; in most cases, these effects are mild and tolerable. Common adverse effects include cough and hoarseness resulting from stimulation of the recurrent laryngeal nerve, which causes vocal fold contraction at therapeutic stimulation parameters.^21,22 In our series of patients, dysphagia was reported more frequently than in other studies, with 17 patients (39.5%) experiencing some difficulty swallowing after implantation. Another notable but less frequent adverse effect was dyspnea, reported by 10 patients (23.2%). These symptoms can significantly impact the quality of life and lead to challenges in feeding and therapy tolerance.

Additionally, 3 patients (6.9%) reported snoring, which can manifest as obstructive sleep apnea in patients with VNS. Adjusting the device settings, specifically by reducing the frequency of the stimulus and decreasing the duty cycle, can help alleviate symptoms of sleep apnea.^8,11 None of the patients were referred for diagnostic polysomnography. However, it is advisable to screen for sleep-disordered breathing (SDB) and confirm the diagnosis through polysomnography to provide specific therapy, such as positive airway pressure, as the incidence of SDB tends to increase following VNS implantation.^23,24 Interestingly, one patient (2.3%) experienced shortness of breath, which improved when the autostim mode was switched off without altering other device parameters. This approach may serve as a successful method prior to lowering the generator normal mode parameters.

Significant adverse effects were observed in 8 patients (18.6%), leading to swallowing assessment, adjustment of settings, device deactivation, or device removal. Within this group, seven patients had GDD, which was significantly associated with severe side effects (p=0.017), and 3 patients had previously undergone corpus callosotomy epilepsy surgery. Most patients had unknown etiology for epilepsy (n=5), while others had genetic etiology, meningoencephalitis, or anoxic ischemic encephalopathy. All patients had symptomatic generalized epilepsy with abnormal EEG findings indicating generalized and multifocal epileptiform discharges. The 8 patients had abnormal brain MRI results, including cerebral atrophy (3 patients), encephalomalacia (2 patients), heterotopia (2 patients), and schizencephaly (one patient). Four patients (50%) had a history of ICU admission due to seizures, and prior ICU admission was significantly associated with severe side effects (p=0.007). Adverse effects were severe enough to deactivate the device in four out of eight patients (50%), with one patient undergoing device removal and one patient experiencing complications leading to death, including aspiration pneumonia. These patients exhibited ≥2 seizure semiologies. The median device output was 1 mA, significantly lower than the typical therapeutic range of 1.5–3 mA,^25,26 Notably, the device parameters did not contribute to the adverse effects as all patients were in the titration phase and had not yet reached optimal stimulation parameters.

Clinical characteristics of patients who developed severe side effects included GDD, symptomatic generalized epilepsy, multiple seizure semiologies, EEG findings of generalized and multifocal epileptiform discharges, abnormal brain MRI, and a history of prior ICU admission. These characteristics can aid in stratifying patients who may experience dysphagia and dyspnea after implantation, which can diminish the tolerability of VNS therapy and quality of life. It underscores the importance of screening for swallowing difficulties before and after implantation to detect and address such issues, thereby reducing the incidence of serious complications like aspiration pneumonia that may require ICU admission. Interestingly, non-invasive VNS through the auricular branch of the vagus nerve in epilepsy patients has shown a reduction in seizure frequency with side effects such as headaches, vertigo, nausea, dizziness, and fatigue, but dysphagia, dyspnea, and obstructive sleep apnea were not reported,^27,28 Similarly, trials of non-invasive VNS at the neck for other indications like migraine did not report obstructive sleep apnea, dysphagia, or dyspnea.^29,30

The implanted device in our study was the latest generation VNS device, the SenTiva Model 1000, and to the best of our knowledge, this study is likely the first to evaluate adverse effects associated with this model. Our study has limitations, including its retrospective nature and reliance on documentation from different providers who may employ different methods of recording information. The small sample size was also a limitation. The frequency of seizures, improvements, and severity of adverse effects were subjective and based on patient or caregiver reports, which may result in underreporting or omission of certain details. Furthermore, during the documented follow-up, changes were made to anti-seizure medications and VNS settings, making evaluation of the efficacy and side effects challenging.