Feasibility and Safety of Elexir, an External Trigeminal Nerve Electrical Stimulator for the Acute Treatment of Migraine

1. Ethics and consent statement

The Institutional Review Board of Seoul National University approved the study protocol (reference number: SNU-2302-070-1405; date of approval: June 27, 2023). The study was conducted in accordance with the ethical guidelines of the Declaration of Helsinki of the World Medical Association. All the participants provided written informed consent prior to participating in the study. As this study enrolled participants who were between the ages of 19 and 65 years, obtaining consent from a parent or guardian was not required.

2. Study design and population

This was a prospective, randomized, double-blind, sham-controlled, post-marketing clinical trial of Elexir for the acute treatment of migraine, conducted across two headache centers in South Korea (Seoul National University Hospital and Uijeongbu Eulji Medical Center).

Participants were considered eligible for enrollment if they: were aged ≥19 and <65 years; met the diagnostic criteria for migraine without aura or migraine with aura according to The International Classification of Headache Disorders, 3rd edition (2018, Sections 1.1 and 1.2);13 patients with typical aura without headache, migraine with brainstem aura, and hemiplegic migraine were excluded; had a history of migraine for at least 1 year; experienced migraine onset before age 50; reported 2–8 migraine attacks per month during the 2 months preceding screening; and provided written informed consent. A migraine attack was defined as a headache lasting 4–72 hours (untreated or unsuccessfully treated) with unilateral throbbing pain of moderate or severe intensity. The exclusion criteria were predefined as follows: inability to distinguish migraine from tension-type headache; ≥15 headache days per month; supraorbital nerve block within 4 months prior to screening; botulinum toxin injection for migraine prevention within 4 months prior to screening; change in preventive migraine therapy within 3 months prior to screening; contraindications to use of the study medical device, including implanted metallic or electronic devices such as deep brain stimulators or cardiac pacemakers (excluding dental implants). Additional exclusion criteria are listed in Supplementary Material (Supplementary Appendix 1, available online).

The sample size was calculated based on a previous study on eTNS.10 A two-sample t-test assuming a two-sided significance level of 5% and 80% power was used. Allowing for a 20% dropout rate, the calculated sample size was increased conservatively.

3. Device information

Participants used a portable eTNS device (Elexir; model ALLIVE2), which delivers transcutaneous electrical stimulation to the supraorbital branch of the trigeminal nerve via forehead electrodes. The device has been approved for two stimulation programs, an acute treatment mode (60 minutes) and a preventive daily-use mode (20 minutes), and is designed as a wearable, home-use neuromodulation system. The acute stimulation mode provides a pulsed current at 100 Hz for 60 minutes, consisting of a ramp-up, stable-intensity, and ramp-down phase. Device output parameters included a maximum current of 16 mA, maximum voltage of 8±1 V (500 Ω load), and 250-µs pulse width. The sham device was identical in appearance and interface to the active device but delivered only a minimal symmetric biphasic pulse (0.5 mA, 50 μs) at 30 seconds interval to monitor electrode-skin contact. This stimulation was below the somatosensory threshold and, therefore, not perceptible to participants and was insufficient to produce any therapeutic effects.14

4. Treatment protocol

The patients were randomized to the verum and sham stimulation groups at a 1:1 ratio. Both investigators and patients were blinded to the randomization process, and a third party was responsible for randomization and concealment. After randomization, all participants underwent in-clinic training on the use of the allocated device, including device application and stimulation. Following the training, participants were instructed to self-apply the allocated device at home for 60 minutes (±5 minutes) at the onset of an eligible migraine attack, defined as: 1) migraine-specific headache with at least one associated symptom (photophobia, phonophobia, nausea, or vomiting); 2) moderate or severe pain intensity within 2 hours of onset; 3) no prior headache within the preceding 48 hours; 4) not spontaneously resolving; and 5) no use of acute migraine medication before device application. Participants were followed for up to 8 weeks after enrollment for the occurrence of an eligible migraine attack. Participants completed study questionnaires assessing changes in pain intensity, resolution of associated symptoms, and rescue medication use at prespecified post-treatment time points, as applicable. After one completed stimulation session, participants were required to visit the study site within 7 days to return the device and submit study-related material, including headache diaries and AE reporting forms. When an eligible migraine attack occurred and the device was applied, the use of any acute migraine medication was prohibited for the first 2 hours after device initiation. After 2 hours, permitted acute migraine treatments (including triptans, ergots, nonsteroidal anti-inflammatory drugs, acetaminophen, combination analgesics, and lasmiditan) were administered as needed.

5. Analysis populations

The intention-to-treat (ITT) population included all randomized participants. The home-use population included participants who used the device at home at least once, regardless of whether the treated migraine attack met the eligibility criteria. The per-protocol (PP) population included participants who used the device at home during at least one eligible migraine attack.

6. Study variables

Baseline demographic and clinical variables included age, sex, disease duration, history of preventive medication use, monthly headache days, monthly migraine headache days, monthly acute medication use days, and response to triptans. Characteristics of the eligible migraine attack included pretreatment headache intensity, the presence of aura, associated symptoms, including nausea, vomiting, photophobia, and phonophobia, and the most bothersome associated symptom. Pretreatment headache intensity was assessed using both a 4-point pain intensity scale (0=no pain, 1=mild, 2=moderate, and 3=severe) and a visual analog scale (VAS; 0–10). Several of these baseline clinical variables were added to the data collection protocol after study initiation and were therefore available only for a subset of participants. Device-log data were used to determine the number and duration of home-use stimulation sessions. For the session-level feasibility analysis, home-use device-log entries with a stimulation duration of ≥1 minute were considered treatment sessions, thereby excluding brief device activations or immediate shutdowns that were unlikely to represent treatment attempts. A completed treatment session was defined as a stimulation duration of 60±5 minutes based on device-logged data. Safety variables included the occurrence of AEs, their relationship to device use (device-related or device-unrelated), serious adverse events (SAEs), and discontinuation owing to AEs. Patient satisfaction with the device was evaluated using a questionnaire based on a 5-point Likert scale (1=very dissatisfied; 2=dissatisfied; 3=neutral; 4=satisfied; 5=very satisfied). Satisfaction was defined as a score of 4 or 5. Efficacy variables included changes in headache intensity, pain freedom, pain relief, resolution of associated symptoms, freedom from the most bothersome associated symptom, and rescue medication use. Post-treatment headache intensity was assessed using the VAS at 1 hour after initiating stimulation and using a 4-point pain intensity scale at 1, 2, and 24 hours after initiating stimulation. Pain freedom was defined as a reduction from moderate or severe intensity (grade 2 or 3) at baseline to no headache (grade 0), whereas pain relief was defined as a reduction from moderate or severe intensity (grade 2 or 3) at baseline to mild or no headache (grade 0 or 1). Both variables were assessed at 1, 2, and 24 hours after initiating stimulation. Resolution of associated symptoms and freedom from the most bothersome associated symptom were assessed using binary (yes/no) questionnaire items at 1 and 2 hours after initiating stimulation. Data on the use of rescue medication between 2 and 24 hours after initiating stimulation were also collected.

7. Outcomes

The primary outcome was the home-use feasibility of eTNS for acute migraine attacks. Home-use feasibility was evaluated in the home-use population using the following measures: 1) appropriate device use, defined as the proportion of participants who used the device during an eligible migraine attack; 2) participant-level treatment completion, defined as the proportion of participants who completed at least one stimulation session within the protocol-defined duration of 60±5 minutes, as determined from device-log data; and 3) session-level treatment completion, defined as the proportion of home-use treatment sessions lasting ≥1 minute that were completed within the protocol-defined duration of 60±5 minutes.

Safety outcomes were evaluated in the ITT population and included 1) the proportion of participants reporting AEs, 2) classification of AEs by device relatedness, and 3) identification of SAEs and discontinuations due to AEs. Patient satisfaction and exploratory efficacy outcomes were evaluated in the PP population. The primary exploratory efficacy outcome was the change in VAS score from before device application to 1 hour after initiating stimulation. Secondary exploratory efficacy outcomes included the proportions of participants reporting pain freedom, pain relief, resolution of associated symptoms, freedom from the most bothersome associated symptom, and the use of rescue medication.

8. Statistical analysis

Continuous variables following a normal distribution were presented as mean±standard deviation, whereas those that did not meet normality assumptions were expressed as median with interquartile range (IQR). Categorical variables were described as counts and percentages. Group comparisons were performed using the independent-samples t-test or Mann-Whitney U-test for continuous variables, and the chi-square test or Fisher’s exact test for categorical variables, depending on distributional characteristics. To assess potential attrition bias, baseline demographic and clinical characteristics were compared between the PP and non-PP populations. For home-use feasibility outcomes, 95% confidence intervals (CIs) for proportions were calculated using the Wilson score method. As a sensitivity analysis of home-use feasibility, participant-level treatment completion was additionally evaluated for the first eligible migraine attack in the PP population. For the analysis of patient satisfaction, responses were dichotomized as satisfied (scores of 4 or 5) and not satisfied (scores of 1–3), and the proportions of satisfied participants were compared between the verum and sham groups using the chi-square test. For exploratory efficacy outcomes, effect estimates with corresponding 95% CIs were calculated. The between-group effect for the change in VAS score was estimated using the Hodges-Lehmann estimator with a corresponding 95% CI. Effect estimates for binary efficacy outcomes were expressed as risk ratios with corresponding 95% CIs, with the sham group as the reference. Statistical significance was determined using a two-sided p-value of <0.05. No interim analyses were conducted, and missing data were analyzed without imputation. All statistical analyses were performed using the IBM SPSS Statistics software (version 29.0; IBM Corp.).

1. Participant disposition and analysis populations

A total of 102 individuals were screened and provided written informed consent between November 2023 and June 2024. Of these, two were excluded before randomization: one did not meet the inclusion criteria and the other withdrew consent before in-clinic device training. The remaining 100 participants were randomized and constituted the ITT population. All participants in the ITT population underwent in-clinic device training before home use. Three participants discontinued the study before home device use: one voluntarily withdrew consent, one withdrew due to an AE during in-clinic device application, and one discontinued due to a SAE unrelated to the device. Of the remaining 97 participants, 18 did not experience an eligible migraine attack during the study period and therefore did not use the device at home, whereas one participant used the device during a non-eligible migraine attack. Thus, the home-use population comprised 79 participants. The PP population consisted of 78 participants, defined as those who used the device at home during an eligible migraine attack. A flowchart of the population disposition is shown in Figure 1.

Figure 1.

Study flowchart.

ITT, intention-to-treat; AE, adverse event; SAE, serious adverse event; PP, per-protocol.

2. Home-use feasibility

Home-use feasibility was evaluated in the home-use population (n=79). Of these participants, 78 (98.7%; 95% CI, 93.2%–99.8%) used the device during an eligible migraine attack, whereas one participant used the device during a non-eligible attack. Seventy-five participants (94.9%; 95% CI, 87.7%–98.0%) had at least one device-log-confirmed completed stimulation session within the protocol-defined duration of 60±5 minutes. Three participants had home-use sessions recorded in the device logs but did not complete a session within the predefined duration, with maximum stimulation durations of 25, 27, and 46 minutes, respectively. In one of these participants, the incomplete session was terminated because of an impedance-error shutdown. For one participant in sham group, home device use for 60 minutes was reported in the study records, but the corresponding home-use session could not be confirmed from the device logs. In a sensitivity analysis restricted to the first eligible migraine attack, 72 of 78 participants (92.3%; 95% CI, 84.2%–96.4%) had a device-log-confirmed completed stimulation session within the protocol-defined duration of 60±5 minutes. After excluding in-clinic training sessions and brief device activations lasting <1 minute, a total of 139 home-use treatment sessions were identified from the device logs. Of these, 103 sessions (74.1%; 95% CI, 66.2%–80.7%) were completed within the protocol-defined duration of 60±5 minutes. Among participants in the verum group who completed a treatment session, the median stimulation current was 5.1 mA (IQR, 3.34–7.23 mA).

3. Safety

In the ITT analysis, 10 participants (10.0%) reported AEs during the study period, including eight in the verum group and two in the sham group, with no significant difference between the two groups (p=0.092). Eight participants (8.0%) experienced device-related AEs, including somnolence (n=2; one in each group), skin eruption and pruritus (n=1; verum), aggravated headache (n=1; verum), pulling sensation in the forehead (n=1; verum), tingling over the forehead or at the patch application site (n=2; verum), and afterimaging (n=1; sham). All device-related AEs were transient and resolved without intervention. The longest-lasting AE was somnolence in the sham group, which persisted for up to 2 weeks. No serious device-related AEs were reported. Two patients in the verum group reported AEs categorized as having a weak or no relationship with device use. One patient reported transient pressure-like discomfort and hyperemia in the left eye, which resolved on the same day and was considered to have a low likelihood of being related to device use. Another participant was diagnosed with papillary thyroid carcinoma, which was determined to be a SAE that was unrelated to the device.

4. Patient satisfaction

In the PP population, 78 participants who used the device during at least one eligible migraine attack completed a usability questionnaire using a 5-point Likert scale. Participant satisfaction with device usability and performance is presented in Table 1. In terms of user acceptability, satisfaction with button placement, user manual, comfort while wearing the device, and device size were reported by 66.7% (52/78), 78.2% (61/78), 61.5% (48/78), and 69.2% (54/78) of participants, respectively. Regarding device performance, satisfaction with the operating conditions, output intensity, and accuracy of output adjustment was reported by 78.2% (61/78), 65.4% (51/78), and 64.1% (50/78), respectively. Overall satisfaction during device use was reported by 50.0% (39/78) of participants, and the median time required to adjust the stimulation output was 2.0 minutes (IQR, 1.0–3.0 minutes).

Table 1.

Satisfaction with device usability and performance in the verum and sham groups

In comparisons between the verum and sham groups, satisfaction with wearing comfort was significantly higher in the verum group than in the sham group (74.4% vs. 48.7%; p=0.020). Satisfaction with the output intensity was higher in the verum group than in the sham group (74.4% vs. 56.4%), although the difference was not statistically significant (p=0.096). Satisfaction with the accuracy of the output adjustment was also significantly higher in the verum group than in the sham group (76.9% vs. 51.3%; p=0.018). Overall satisfaction did not differ significantly between groups (59.0% vs. 41.0%; p=0.113) (Table 1).

5. Efficacy

The efficacy analysis was performed in the PP population (n=78; mean age, 42.9±11.6 years; female, 84.6%), defined as participants who used the device during at least one eligible migraine attack. Baseline demographics and characteristics of the PP population are shown in Table 2. The median disease duration was 8.5 years (IQR, 3.0–15.5 years). Among the treated eligible migraine attacks, all were of moderate or severe intensity and were accompanied by at least one associated symptom. Of the 78 participants, 66 participants (84.6%) had moderate headache intensity, and median VAS score was 6.0 (IQR, 5.0–7.0). The reported associated symptoms included nausea (61.5%), photophobia (53.8%), phonophobia (50.0%), and vomiting (9.0%). The most bothersome associated symptoms were nausea (46.2%), photophobia (29.5%), phonophobia (23.1%), and vomiting (1.3%). To assess potential attrition bias, the available baseline demographic and clinical characteristics were compared between the PP population (n=78) and the non-PP population (n=22). No statistically significant differences were observed between the two populations (Supplementary Table 1, available online).

Table 2.

Demographic and clinical characteristics of the study participants

Exploratory efficacy outcomes are presented in Table 3. The median reduction in VAS score at 1 hour was 2.0 (IQR, 0.0–4.0) in the verum group and 2.0 (IQR, 1.0–4.0) in the sham group (Hodges–Lehmann estimate, 0.0; 95% CI, –1.0 to 1.0; p=0.480). No statistically significant between-group differences were observed in pain freedom or pain relief at 1, 2, or 24 hours after initiating stimulation. Similarly, the resolution of associated symptoms and freedom from the most bothersome associated symptom at 1 and 2 hours did not differ significantly between the groups. Rescue medication use between 2 and 24 hours was reported by 18 of 39 participants (46.2%) in each group. Among participants who achieved pain freedom or pain relief at 24 hours, rescue medication use also did not differ significantly between the groups (Table 3).

Table 3.

Efficacy outcomes in the verum and sham stimulation groups