Introduction

Epilepsy is one of the most common neurologic disorders and a leading cause of disability worldwide.1 Clinical indices of epilepsy such as seizure frequency do not adequately account for the burden of this disorder, and research indicates that comorbidities such as stigma, anxiety, and depression can be more debilitating than the seizures themselves.2,-,4 Stigma associated with epilepsy can exist in the society at large, or at an individual level, wherein a person with epilepsy holds a sense of shame or fear regarding their own health condition.5 The latter, known as felt stigma, is a common consequence of epilepsy in both developed and developing countries.5,-,7

Felt stigma is shown to significantly restrict access to treatment, health care, and medication adherence and was found to be a key predictor of emergency department visits,8,9 impaired quality of life, and poor psychosocial well-being,7,10,11 therefore demonstrating a direct impact on the clinical outcomes of the epilepsy. To date, there exists no intervention with proven efficacy to mitigate felt stigma in epilepsy, which remains a recalcitrant and inadequately addressed problem for neurologists and caregivers around the globe.

Previously, authors have demonstrated that certain interpersonal factors such as maladaptive coping strategies and poor emotion regulation are foundational in development of felt stigma related to any existing condition, including epilepsy.12,-,14 Yoga practices are shown to improve regulation of emotions.15 In addition, similar mind-body approaches were previously shown to be efficacious in reducing the overall burden of epilepsy.16,-,25 Therefore, it was hypothesized that yoga therapy can effectively reduce felt stigma in persons with epilepsy.

Yoga falls under the umbrella of mind-body therapies which encompass a large group of interventions such as meditation, tai chi, and visual imagery.26 These approaches are hypothesized to influence health by interacting with the autonomic nervous system and inducing a relaxation response.26

We conducted an assessor blinded, sham yoga–controlled randomized clinical trial of an integrated yoga-based intervention vs sham yoga for adults with epilepsy who scored positively on the Kilifi Stigma Scale (KSS). In addition, both groups were administered epilepsy-related psychoeducation over the course of study duration. The following variables were measured as secondary outcomes: seizure frequency, quality of life, anxiety, depression, mindfulness, cognitive impairment, trait rumination, and emotion regulation.

The primary objective of this clinical trial was to assess whether a yoga intervention in conjunction with psychoeducation can reduce the felt stigma in persons with epilepsy as compared with a sham yoga and psychoeducation intervention.

MethodsTrial Design

This was a prospective, 2-armed, assessor-blinded, parallel randomized controlled trial (RCT) with 3 assessment time points—baseline, 3 months, and 6 months (Figure 1, eFigure 1, links.lww.com/WNL/D201). The study was conducted in the neurology unit of All India Institute of Medical Sciences, New Delhi, India, from January 2018 to July 2020 (eAppendix 1).

Figure 1 Consolidated Standards of Reporting Trial Chart of the Clinical TrialStandard Protocol Approvals, Registrations, and Patient Consents

Ethical clearance for this trial was obtained from the Institute Ethics Committee of All India Institute of Medical Sciences, New Delhi, and it was prospectively registered in the Clinical Trials Registry of India (Ref No. CTRI/2017/04/008385). All patients gave written informed consent. This study follows the Consolidated Standards of Reporting Trials (CONSORT) guidelines of reporting for clinical trials.

Patient Selection and Enrollment

Consenting adult patients diagnosed with epilepsy (age 18–60, as per ILAE criterion) who scored ≥4 on KSS were enrolled for this trial. A score greater than the 66th percentile on KSS indicates the presence of high felt stigma, whereas below that indicated low felt stigma.27,28 For determining the 66th percentile of KSS in our settings, data were collected from a group of 30 patients attending the neurology outpatient department before the commencement of the trial. Based on these results, the cutoff for high felt stigma in our population was determined to be 4. Details of this analysis are provided in eAppendix 2 (links.lww.com/WNL/D201).

Exclusion criterion included patients who were nonconsenting or were practicing yoga/meditation for a period of 3 months or more before enrollment. Patients who presented with psychogenic nonepileptic seizures (PNES) alone were excluded, while patients with epilepsy and comorbid PNES were included in the trial.

Randomization and Blinding

Randomization was stratified in a 1:1 ratio based on seizure frequency (patients who were seizure-free for a period of 1 year or more and those who had at least 1 epileptic seizure in the last 1 year were randomized separately). A computer-generated randomization list was created using the block permutation method stratified by seizure frequency. Serially numbered, opaque envelopes containing randomization codes were prepared accordingly for the 2 strata. Both the randomization sequence and the sealed envelopes were prepared by an independent researcher who was not involved in the study. The treatment group to which a patient was allocated (yoga + psychoeducation or sham yoga + psychoeducation) was concealed from patients (and caregivers) themselves, the outcome assessor (R.D.), the analyzing statistician (A.D.U.), and the clinicians (M.T., B.R.) until the study completion by the study investigator (K.K.).

Interventions

The intervention group was administered a yoga protocol, while the control group was administered a sham yoga protocol for the same duration. The yoga and sham intervention modules were prepared and administered in collaboration with the Centre for Integrative Medicine and Research, All India Institute of Medical Sciences, New Delhi.

The yoga module consisted of the following: loosening practices (Sukshma Vyayama), breathing exercises (Prānāyāma), meditation, and positive affirmations. The detailed protocol for the same including the frequency and duration of administration is provided in eAppendix 3.1 (links.lww.com/WNL/D201).

The sham yoga module consisted of exercises that mimic the yoga asanas mentioned above, but the participants were not provided instructions on 2 key components of yoga, namely (1) slow and synchronized breathing and (2) attention to the body movements and sensations during practice. Without these 2 mentioned techniques, the sham intervention was hypothesized to have a metabolic signature equivalent to that of yoga but was devoid of its key components which are assumed to induce a relaxation-based response. Details of the same are provided in eAppendix 3.2 (links.lww.com/WNL/D201).

A psychoeducation module consisting of the following key elements—information about epilepsy, its symptoms, treatment and known causes, guidelines for medication compliance and self-care strategies, psychosocial aspects of epilepsy including myths and superstitions surrounding this disorder, management, and coping strategies to maintain a positive attitude were also delivered to both the groups (eAppendix 3.3, links.lww.com/WNL/D201).

Both the intervention and comparator group were given the same preintervention counseling and reading material after the baseline assessment and before the initiation of intervention. Each group received 7 supervised group sessions lasting about 45–60 minutes which were conducted over a period of 12 weeks (3 months). All supervised sessions of yoga were conducted by a certified yoga therapist. Instructions were given to the patients and their caregivers in the vernacular (Hindi) or English in an easy to understand manner. Patients were asked to practice the respective interventions at home at least 5 times a week for a duration of 30 minutes. In addition, they were given a combined seizure diary and practice log which documents the daily frequency of seizures and duration of practice at home.

Outcomes

The primary outcome measure was a reduction in the scores of KSS for epilepsy (KSSE) in the intervention arm as compared with the control arm. Because there exists no “gold standard” for measuring felt stigma, the Revised Epilepsy Stigma Scale (RESS)29 was also assessed as a secondary outcome in addition to KSSE. Further justification for using this additional scale is given in eAppendix 4 (links.lww.com/WNL/D201).

All scales were translated and validated in vernacular wherever necessary. The WHO guidelines from validation of translations and adaptations of instruments (2016) were used for validating these translations (eAppendix 5, links.lww.com/WNL/D201).

Seizure frequency as monitored by a seizure diary was a secondary outcome of this trial. The research team requested the participants to fill the seizure diary every day, and the following 3 strategies were used to ensure the same—(1) the seizure diaries were cross-checked by the research team at every visit, and any discrepancies were discussed and resolved; (2) patients were sent regular (weekly) reminders to fill the seizure diary and home practice log by the research team; and (3) caregivers were requested to encourage intervention adherence and maintenance of seizure logs in the preintervention counseling. This seizure diary was handed over directly to the outcome assessor during follow-up, and the patients were requested not to discuss their intervention protocol with the outcome assessor.

Frequency of only epileptic seizures was considered for the analysis presented in this article, and all nonepileptic seizures were excluded. Details of the same are provided in eAppendix 6 (links.lww.com/WNL/D201). In addition, depression symptoms (Patient Health Questionnaire-9 [PHQ-9]), anxiety symptoms (Generalized Anxiety Disorder Assessment-7 [GAD-7]), trait rumination (rumination subscale of response style questionnaire), quality of life (Quality of Life in Epilepsy Inventory-10 [QoLIE-10]), cognitive impairment (Montreal Cognitive Assessment [MoCA]), mindfulness (Mindful Attention and Awareness scale), and emotion regulation (Emotion Regulation Questionnaire) were also measured as secondary outcomes. Finally, all recruited patients underwent at least 20 minutes of resting-state high-density EEG recording (using EGI System; Electrical Geodesics, Inc., Eugene, OR) to study the changes in spectral power density in alpha, beta, theta, delta, and gamma bands, the results of which have been described in an independent publication.

Monitoring Protocol Adherence

The recruited patients were given a seizure and home practice log, which was to be maintained for the period of intervention as well as follow-up period. Similarly, the caregivers were also requested to monitor and encourage intervention compliance. All patients were given reminder calls/text messages by the research team 1 day before the session to maintain attendance of supervised sessions.

Adverse Event Monitoring

Any adverse events reported during the duration of the study were monitored at every visit and all follow-up time points by the outcome assessor.

Sample Size

Based on previous literature,30 a difference of >45% in primary outcome was anticipated between the intervention and control group at the end of 6 months. Considering 5% level of significance and 80% power of the study, we get a sample of 63 patients per group. Allowing for 20% dropout and rounding off to the nearest multiple of 10, 80 patients per group or a total of 160 patients were enrolled in this trial.

Statistical Analysis

Data were summarized by frequency (%); quantitative variables were summarized by mean (SD)/median (interquartile range), as appropriate. Categorical variables were compared by the χ2 test/Fisher's exact test and continuous variables by unpaired t test (normal distribution) or the Wilcoxon-Mann-Whitney test (for non-normal distribution). Given the high variability and skewed distribution of seizure frequency at baseline, the difference in seizure frequency after the administration of interventions was calculated for seizure reduction, 50% responder rate, and 100% responder rate. The seizure reduction was tested for statistical significance using analysis of covariance (ANCOVA), while the responder rates were assessed using Fisher's exact test. Intention-to-treat (ITT) analysis was performed for all the recruited patients using the method of last observation carried forward. Patients who adhered to the protocol were included in per-protocol (PP) analysis. ANCOVA and rank ANCOVA were used to test for difference in change between the group by considering the baseline in case of normally and non-normally distributed variables, respectively. In addition, percent change from baseline at varying time points was also calculated.

All statistical analysis was conducted using RStudio (version 4.1).

Data Availability

Data not provided in the article because of space limitations may be shared (anonymized) at the request of any qualified investigator for the purposes of replicating procedure and results.

Results

A total of 456 patients were assessed for eligibility; of this, 296 were excluded because they did not meet the eligibility criterion of the trial (Figure 1). A total of 160 patients were enrolled in the study as per the requirements of the sample size. Sixteen patients from the control group and 11 from the intervention group were lost to follow-up at the end of 6-month time point (Figure 1).

An overview of the screening, recruitment, and follow-up process in accordance with CONSORT guidelines is given in Figure 1.

The demographic baseline and clinical characteristics for both the groups are provided in Table 1. All demographic and clinical variables were balanced between both the groups. The mean (SD) age of the patients in both groups was 29.64 (8.37). Sixty percent of the enrolled patients (54 in the control group and 42 in the intervention group) were classified as having focal onset seizures. The mean number of antiseizure medications prescribed to both the groups was ≥2. Approximately 15% patients in the comparator arm and 19% in the intervention arm reported having comorbid psychogenic seizures (Table 1).

Table 1

Baseline Demographic and Clinical Characteristics of the Recruited Patients

Primary Outcome

The baseline scores and the results for the primary outcome, that is, KSS, are given in Table 2 according to ITT and PP analysis. Significant reduction was found at the end of primary endpoint (6 months) using both methods of analysis in the intervention arm as compared with the comparator arm. The effect sizes at the end of 6-month period was found to be Cohen's d = 0.38, 95% CI 0.04–0.72. The corresponding results for 3 months of follow-up are provided in eTable 1 (links.lww.com/WNL/D201). A violin plot depicting the change scores of KSS from baseline to 6 months in both groups according to ITT analysis is shown in Figure 2.

Table 2

Outcomes on the Kilifi Stigma Scale According to the ITT and PP Analysis at 6 Months of Follow-up

Figure 2 Violin Plots of the Change Score From Baseline to 6 Months

**p < 0.05.

Secondary Outcomes

The patient reported seizure frequency per week in both the groups at baseline and 6-month of follow-up is given in Table 3, and the corresponding results for 3 months are provided in eTable 2 (links.lww.com/WNL/D201). The proportion of patients having >50% seizure reduction and complete seizure remission at the end of 6 months of follow-up was significantly higher in the intervention group (odds ratio [OR] 4.11, 95% CI 1.34–14.69 and OR 7.4, 95% CI 1.75–55.89, respectively) (Table 3, Figure 2). Similar differences were also observed at the end of 3 months (eTable 3).

Table 3

Seizure Frequency (per Week) and Percentage Reduction in Seizure Frequency at 6 Months of Follow-up

The outcomes of other secondary variables with respect to ITT and PP analysis at 6 months of follow-up period are provided in Tables 4 and 5, respectively. The corresponding results for 3 months of follow-up are provided in eTables 4 and 5 (links.lww.com/WNL/D201).

Table 4

Outcomes of Secondary Variables as per ITT Analysis After 6 Months of Follow-up

Table 5

Outcomes of Secondary Variables According to per Protocol Analysis After 6 Months of Follow-up

The intervention group showed significant decrease in anxiety symptoms (GAD-7) compared with the control group at the end of 6 months according to ITT analysis (Table 4). However, similar significant decrease could not be observed with respect to PHQ-9 in the intervention group in either ITT or PP analysis (Tables 4 and 5).

Statistically significant improvement was reported with respect to the quality of life outcome (QoLIE-10) in the intervention group for both ITT and PP analysis at primary end points as compared with the control group (Tables 4 and 5). Similar changes were also observed with respect to trait mindfulness and cognitive impairment (Tables 4 and 5). No significant improvements were observed with respect to train rumination, cognitive reappraisal, and emotion suppression as per both ITT and PP analysis at the end of follow-up period.

On an average, participants in the intervention group attended approximately 60% of the supervised sessions, while those in the control group attended approximately 54% (eTable 6, eFigure 1, links.lww.com/WNL/D201). The self-reported median level of home practice in the intervention and control group was 80% and 60%, respectively. Protocol adherence was observed to be a significant predictor of outcomes on KSS (eAppendix 7). Additional details of protocol adherence, results, and discussion on the same are given in eAppendix 8.

Classification of Evidence

This study provides Class I evidence that yoga reduces felt stigma in adult patients with epilepsy.

Discussion

In our study, favorable changes were reported in stigma scores as measured by the KSS after 6 months of follow-up according to both ITT and PP analysis. This is a relevant outcome because previous studies on felt stigma in epilepsy did not report any positive findings,31,-,34 and to date, there exists no efficacious intervention for the same.

In this study, we did not find any positive change in the RESS, which was measured as a secondary outcome. However, a correlation analysis between RESS and KSSE at the baseline and subsequent time points yielded significant results, suggesting that these scales were measuring common underlying items (eAppendix 9, links.lww.com/WNL/D201).

RESS measures the perceived inferiority, discomfort, and avoidance of the people with epilepsy by the society at large.35 The KSS28 in addition to these items also measures constructs such as perceived discrimination, disappointment, and loneliness in people with epilepsy, which are important elements of the “concept map of felt stigma” in epilepsy as reported by Wei et al.36 This may account for the observed differences in the outcomes of the 2 scales.

In this study, a significant number of patients reported >50% reduction in seizure frequency at 6 months in the intervention arm as compared with the control arm. This measure of >50% seizure reduction (known as 50% responder rate) is also mandated by the European Medicines Agency for regulatory approval of novel pharmaceuticals for epilepsy. The 50% responder rate in our study was approximately 40% in the intervention arm at 6 months. However, because self-reported measures were used for measuring seizure frequency, these estimates may be prone to inaccuracies.37 Second, the natural history of seizure frequency tends to be highly variable, and spontaneous decrease in seizure count can erroneously be attributed to an intervention.38 Therefore, an adequately powered, large-scale trial is required to conclusively determine the effect of yoga-based interventions on seizure frequency.

Previously, several studies have shown the beneficial effects of similar relaxation-based interventions such as meditation, mindfulness-based therapies, and progressive muscle relaxation in epilepsy.16,17,19,25,39 These studies demonstrated a positive impact of such behavioral interventions on seizure frequency as well as the overall quality of life and mental well-being of patients. In addition, a few studies have specifically explored the impact of varied yoga techniques in epilepsy and found similar results.18,21,-,24

However, only a few of these studies discussed above were adequately powered or blinded RCTs. Moreover, only 2 of these studies controlled for the placebo effect by providing sham yoga.21,22 This is an RCT with a large sample size to demonstrate that patients administered yoga-based intervention have greater odds of seizure reduction, although this finding needs further validation in an adequately powered clinical trial.

We postulate that these changes may be driven by the multidimensional effect of the primary intervention on autonomic nervous system and brain network connectivity as has been previously indicated by studies of similar nature.17,40,41

The literature suggests that for our study cohort, >2 point average reduction in PHQ-9 and >2.2 point reduction in GAD-7 are the minimum clinically important difference on these scales.42 These clinically important differences were observed in the median values of PHQ-9 in both the control and intervention group at the end of the follow-up period when compared with the baseline. However, on GAD-7, the same was observed for the intervention group alone. In addition, changes in symptoms of depression and stigma were found to be correlated in the intervention arm at both follow-up time points. A further discussion on the same is provided in eAppendix 10 (links.lww.com/WNL/D201).

Mood disturbances are a common feature of epilepsy,43 and these remain inadequately addressed in clinical settings of developing countries. Yoga may offer a scalable option for addressing the same.

Consistent improvement was also reported in the quality of life of patients recruited in the intervention arm relative to the control arm. This is another significant outcome because a holistic improvement in health-related quality of life, and not seizure remission alone is the primary aim of comprehensive medical care. Quality of life is a multifaceted concept, and many factors such as improvement in mood, mindfulness, or reduction in seizure frequency may contribute to the same.

Interestingly, the decline in stigma scores was not associated with a postintervention change in scores of trait rumination and emotion regulation, although previous literature indicates that these factors may play a key role in the development of a stigmatized identity.12,13

Another interesting finding of this study was the consistently significant increase in trait mindfulness in the intervention group as per both ITT and PP analysis. Mindfulness has been described as the “nonjudgmental attention to experiences in the present moment” and a cognitive skill which can be improved with training.44 A few studies have hypothesized that mindfulness itself may be a strategy for regulating emotions in addition to others such as emotion suppression, reappraisal, and rumination.45 A cross-sectional study46 showed that greater levels of mindfulness predicted lower levels of self-stigma and increased levels of help-seeking behavior. Our study also lends partial support to this idea because regular yoga practice was associated with increased mindfulness and reduction in felt stigma and depression symptom severity. It can be hypothesized that yoga improves neuropsychological outcomes in patients with epilepsy by positively modulating trait mindfulness.

Patients enrolled in the intervention group also showed improvement in cognitive impairment as measured by MoCA. A systematic review by Chiesa et al.47 concluded that stronger evidence is needed to demonstrate the association between various forms of meditation training and improvement in different facets of cognition. This study again lends support to the hypothesis that yoga meditation–based interventions can improve cognitive outcomes.

Finally, it is relevant to note that this trial attempted to control for the placebo effect by administering a sham intervention to the control group, while providing similar levels of treatment care, attention, and psychoeducation. This group was administered exercises which mimic yoga asanas but were devoid of its key components, namely—breath modulation and attention which are hypothesized to induce a relaxation response.48 However, it is possible that patients in this cohort benefited from the placebo effect of the intervention. Thus, the effect of yoga intervention may be more pronounced in a day-to-day clinical setting where most patients do not receive any additional treatment.

A limitation of this study was the lack of a passive control group which would indicate the effect size of the intervention in patients who only receive treatment as usual (TAU). We postulate these effect sizes would be larger when compared with TAU as opposed to sham control, as was the case in this study. Another limitation of this study was the lack of prospective monitoring of seizure frequency before commencement of intervention. This may have biased the estimation of baseline seizure frequency and therefore the change in seizure frequency as an outcome as well.37 Furthermore, there was no statistically significant change in seizures per week across interventions. This may be a result of the high variability in seizure frequency per week at the baseline, and a similar phenomenon was observed in another recent publication by Manral et al.49 In addition, the integrity of treatment blinding was not assessed at the end of the intervention period. For future research, querying the patients about the intervention they were assigned to may be an important step for studies of such nature because these factors may influence adherence and outcomes.

A further limitation of this trial was the less than expected rate of adherence and attendance of supervised sessions in both groups. The median frequency of protocol adherence in our study was 80% and 60% in the intervention and control group, respectively. Although these rates of adherence are comparable or greater than previous studies of similar nature,50 these may still be a source of bias in the analysis. A further discussion on the same is provided in eAppendix 8 (links.lww.com/WNL/D201).

In conclusion, this RCT reported significant improvement for stigma, seizure frequency, cognition, mindfulness, and overall quality of life for patients who enrolled in the intervention group, relative to the control group. Yoga offers an efficacious method to address these comorbidities in a clinical setting. Second, the psychoeducation module developed in this study can easily be delivered by a nurse, medical student, or health care worker. Therefore, the intervention assessed in this trial can be potentially disseminated online or using prerecorded videos using minimal resources.

This sham-controlled trial demonstrated that yoga-based intervention in conjunction with psychoeducation can reduce epilepsy-related stigma, anxiety symptoms, and cognitive impairment in persons with epilepsy. Simultaneously, it may also improve the overall quality of life and trait mindfulness in these groups of patients. This intervention was efficacious in managing neuropsychiatric comorbidities in epilepsy.