Focal encephalomalacia is a common structural brain lesion detected during magnetic resonance imaging (MRI) in patients with pharmacoresistant epilepsy (1, 2). The etiology of encephalomalacia includes brain trauma, perinatal hypoxia, infection, intracranial hematoma, surgical procedures, as well as some unknown factors (3). Although the encephalomalacia alone may not cause seizures, the surrounding scars may interfere with the normal electrophysiological activity of neurons and cause hyperplastic glial dysfunction, which in turn leads to abnormal discharge associated with seizures (4, 5). Patients with pharmacoresistant epilepsy secondary to encephalomalacia are usually resistant to anti-seizure medications, and surgical intervention is another widely accepted treatment option (3, 6). However, in conditions of widely distributed encephalomalacia involved in eloquent brain regions or bilateral hemispheres, patients are not good candidates for resection (5, 7). Thus, for those unsuitable for surgical therapy or with unsatisfactory surgical outcomes, it is urgent to explore novel therapeutic strategies.

Since its first reported use in humans in 1988 and more than 100,000 subsequent implantations, vagus nerve stimulation (VNS) has become a reliable method of treating patients with pharmacoresistant epilepsy who are not good candidates for epilepsy surgery or in whom surgery resulted in no benefit (8). According to the results of randomized controlled trials (9), meta-analyses (10), and retrospective studies (11, 12), ~50–60% of patients achieve a seizure reduction of ≥50% after VNS surgery, with a rate of complete seizure freedom ranging from 6 to 8%. Predictors of VNS effectiveness are a focus of related research at present. Several potential predictors updated recently include brain connectomic profiling (13), heart rate variability (14), and genetic variations of adenosine kinase (15). The effectiveness and safety of VNS are also demonstrated in some specific types of epilepsy, such as tuberous sclerosis complex (16), Lennox–Gastaut syndrome (17), post-encephalitic epilepsy (18), and post-traumatic epilepsy (19). Based on the advantages of VNS therapy, it may shed some light on the therapy of pharmacoresistant epilepsy secondary to encephalomalacia.

Various causes can lead to encephalomalacia in the brain, such as stroke (20, 21), head trauma (19, 22, 23), and encephalitis (18, 24), in which the effectiveness of VNS for epilepsy has been reported, separately. This present study aimed to demonstrate the effectiveness of VNS in 93 patients with pharmacoresistant epilepsy secondary to encephalomalacia under different conditions, as well as to evaluate potential predictors for VNS effectiveness.

Focal encephalomalacia is a common structural brain lesion found in MRI of patients with pharmacoresistant epilepsy. VNS has been used in pharmacoresistant epilepsy for decades. For patients who are unsuitable for resection surgery, VNS may provide better benefits for seizure reduction. However, the long-term seizure outcomes and potential prognostic predictors of VNS in pharmacoresistant epilepsy secondary to encephalomalacia remain unclear. In this study, we first assessed the VNS effectiveness in pharmacoresistant epilepsy secondary to encephalomalacia with a follow-up over 1 year. Out of 93 patients enrolled in this study, 60 (64.5%) patients obtained a reduction of ≥50% in seizure frequency, and seizure freedom occurred in 15 (16.1%) patients. During the follow-up time ranging from 1.0 to 12.0 years, the most frequently reported adverse events included voice hoarse, coughing, and throat pain, while all the side effects above were tolerable and transient. After device implantation, the responder rate, seizure freedom rate, and the median reduction of seizure frequency were all found to gradually increase over time. Those results were consistent with most studies involved VNS effectiveness in pharmacoresistant epilepsy (30, 31), which reported a reduction of or more than 50% in seizure frequency in 45–65% of patients as well as a progressive increase in the overall response to VNS therapy over time. Therefore, VNS therapy was demonstrated to be effective and safe in patients with epilepsy secondary to encephalomalacia. For those who are not suitable for resection surgery, VNS might be a promising therapeutic strategy.

Seizure freedom is generally considered a prominent predictor of life quality in patients with epilepsy. Unfortunately, complete seizure freedom is rarely obtained (6–8%) in patients who underwent VNS surgery (9–12). Among 93 patients with pharmacoresistant epilepsy secondary to encephalomalacia in the present study, 15 (16.1%) patients obtained seizure freedom at the last follow-up, which was higher than that observed in the general population of epilepsy. The relatively high rate of seizure freedom indicated that patients in the small cohort may achieve more improvements in the overall life quality via VNS therapy than those with other types of epilepsy. Further studies with larger sample sizes are expected to focus on this problem in the future.

Although the VNS benefit was found more significant in the present cohort than in other types of pharmacoresistant epilepsy, our results confirmed that the complete seizure freedom rate was still less common with VNS than with resective surgery (3). In our previous study focusing on the surgical outcomes in patients with epilepsy secondary to encephalomalacia who received resective epilepsy surgery, ~75.0% of the patients obtained seizure freedom 5 years after surgery (3). A study involving 17 patients with resection of frontal encephalomalacia for pharmacoresistant epilepsy reported that 12 (70%) patients were seizure-free or had only rare seizures after a median of 3 years of follow-up (6). The phenomenon also occurs in other neuromodulation treatments for pharmacoresistant epilepsy (32, 33). Therefore, current neuromodulation techniques are indeed not a substitute for resection therapy for pharmacoresistant epilepsy. However, epilepsy patients with widely distributed encephalomalacia which is involved in eloquent brain regions or bilateral hemispheres are not good candidates for resection. In such conditions, as a palliative treatment, VNS may help reduce the seizure frequency, as well as improve the overall life quality.

In addition to reducing seizure frequency, VNS may also benefit the quality of a patient's life by improving physical disability and neuropsychological disorders (34, 35). Patients with encephalomalacia on MRI usually have various types of initial etiologies, including brain trauma, perinatal hypoxia, meningoencephalitis, previous surgical procedures, and intracranial hematoma, any of which is associated with different degrees of brain damage (6). Therefore, those patients often suffer from neurological and neuropsychological impairments such as physical disability, depression, or anxiety (36, 37). In this study, 31.2% of patients reported a preoperative neurological deficit. Those deficits included hemiparesis, aphasia, ataxia, and persistent vegetative state. Modification of them was also a crucial step during the overall treatment. Multiple preclinical studies on ischemic stroke models have shown that VNS combined with rehabilitation training can significantly improve the recovery of forelimb motor function compared with rehabilitation training without VNS (34). Stimulation of the vagus nerve accelerates the release of neuromodulators, which can promote neuroplasticity throughout the cortex, such as acetylcholine and norepinephrine (38–40). Besides, it is well demonstrated that VNS therapy has benefits on mood, behavior, and cognition for epilepsy patients, independent of reducing seizures (35, 41). Thus, the potential benefits of VNS on psychological and neurological disorders in patients with pharmacoresistant epilepsy secondary to encephalomalacia cannot be ignored. Unfortunately, the neuropsychological disorders and the effectiveness of VNS for those symptoms were not presented in this study, which deserved further exploration in the future.

In the present study, we first evaluated the prognostic predictor of VNS effectiveness in patients with pharmacoresistant epilepsy secondary to encephalomalacia. After multivariate analysis, the age of the seizure onset >18 years was found to predict better effectiveness. Similar results have also been reported in previous studies. In a study recruiting 5,554 epilepsy patients with VNS therapy, the age of epilepsy onset >12 years was found to predict a higher rate of seizure freedom (10). A retrospective analysis of 158 patients with medically pharmacoresistant epilepsy reported that patients with age at seizure onset ≥15 years were ideal candidates for VNS (42). Thus, patients with seizure onset in adults (>18 years old) demonstrated more likely to benefit from VNS therapy than those who had seizure onset in children (≤18 years old). The potential mechanisms of the finding were still unclear. More studies with larger sample sizes are expected to further confirm the phenomenon and explore the underlying mechanisms in the future.

Among the recruited 93 patients, those with unilateral IEDs were found to have a higher rate of responder and seizure freedom, as well as higher median reduction of seizure frequency compared with those with bilateral IEDs at different follow-ups. The important role of EEG features in the prediction of VNS effectiveness for epilepsy has been demonstrated before (43–46). In our previous studies exploring the VNS effectiveness in 42 patients with pharmacoresistant post-encephalitic epilepsy with a follow-up ranging from 1.00 to 11.83 years, patients with focal IEDs were found to have better seizure outcomes than those with generalized IEDs at the last follow-up (18). A study including 144 patients with pharmacoresistant epilepsy reported a significant association between unilateral IEDs and a higher probability of seizure freedom after VNS surgery (44). Notably, it is also the case in resective surgery of patients with epilepsy. In patients with mesial temporal sclerosis who received surgical treatment, bitemporal IEDs indicated bitemporal epileptogenicity and predicted a worse seizure prognosis than unilateral-temporal spike foci (47, 48). The most recognized reason was probably that the bilateral IEDs represented an enlarged epileptogenic zone or a greater epileptogenicity, as the bilateral IEDs were usually accompanied by a bilateral seizure onset zone, a generalized seizure diffusion, and a greater seizure frequency (49, 50). In addition, the bilateral IEDs arising from an interaction of multiple active foci presented a higher degree of epileptogenicity (51). Thus, whether for VNS effectiveness or resection surgery, the spatial distribution of IEDs could be considered a reliable assessment tool for the prognosis of seizure outcome. Besides, the distribution of encephalomalacia foci on MRI was also found associated with VNS effectiveness in the present study. Similar to bilateral IEDs, bilateral encephalomalacia might contribute to the worse seizure outcome via similar mechanisms, such as widely distributed brain lesions, generalized seizure propagation, and higher epileptogenicity. In addition, we also evaluated the association between the concordance of the IEDs and ictal onset rhythms or the IEDs and MEG findings with VNS effectiveness in the present study. However, no significant results were obtained after statistical analysis. The results might be biased by the small sample size (36 of 93 patients had MEG results). Further studies with larger sample sizes are expected to focus on this problem in the future.

It was important to acknowledge some limitations of the present study. First, the inherent biases and the relatively small sample size of this single-center retrospective study could not be ignored, and more prospective studies with a larger sample size need to be carried out in the future to make the findings more targeted. Second, some factors that may influence the comprehensive curative effect of VNS in the specific cohort were not included, such as the clinical assessments of neuropsychological problems, behavior disorders, and overall life quality. Third, the 1.5-T MRI equipment used in this study may result in an underestimation of the number of patients with mild encephalomalacia, potentially increasing selection bias. In spite of these limitations, this study suggested the effectiveness of VNS in reducing seizure frequency in patients with pharmacoresistant epilepsy secondary to encephalomalacia. In addition, the age of seizure onset, the spatial distribution of IEDs, and the spatial distribution of encephalomalacia foci on MRI might be independent predictors of VNS effectiveness.

The present study indicated that VNS therapy was effective in patients with pharmacoresistant epilepsy secondary to encephalomalacia, with an ideal tolerance in patients over a 1-year-follow-up period. Patients with seizure onset in adults (>18 years old), unilateral IEDs, or unilateral encephalomalacia on MRI were found to have better seizure outcomes after VNS therapy.

The data analyzed in this study is subject to the following licenses/restrictions: Non-identifying data are available from the corresponding author on reasonable request. Data were not placed in a repository due to the risk of re-identification of participants. Requests to access these datasets should be directed to TL, tianfuli@ccmu.edu.cn.

The studies involving human participants were reviewed and approved by the Ethics Committee of Sanbo Brain Hospital, Capital Medical University (SBNK-2017-15-01). Written informed consent to participate in this study was provided by the patients/participants or patients/participants' legal guardian/next of kin.

TL, GL, and MG contributed to the conceptualization. TL, GL, MG, JW, CT, JD, and JZha contributed to the methodology. MG, ZX, and XK contributed to the formal analysis and investigation. MG contributed to the writing—original draft preparation. TL, GL, MG, JW, CT, JD, JZha, XW, XK, YG, JZha, FZ, and ZX contributed to the writing—review and editing. YG, JZha, and FZ contributed to the investigation. All authors contributed to the manuscript revision, read, and approved the submitted version.