Abstract
Background
The Trial of External trigeminal nerve stimulation (eTNS) for the Acute treatment of Migraine (TEAM) study demonstrated that eTNS use during active migraine resulted in significantly higher rates of resolution of migraine‐associated most bothersome symptom (MBS) compared to sham. However, no previous studies have examined the association between pretreatment MBS subtype and efficacy of eTNS treatment for active migraine.
Objective
We conducted a post hoc analysis examining efficacy of eTNS for different pretreatment MBS subtypes using TEAM study data.
Methods
Pretreatment MBS subtypes included photophobia (n = 345), nausea (n = 109), phonophobia (n = 73), and vomiting (n = 11). We examined MBS sub‐group × treatment group (verum n = 259; sham n = 279) interaction for each post‐treatment outcome to explore differential effects conditional on the total sample. We further explored direct, between treatment group comparisons for each MBS subtype, as well as compared treatment outcomes among all MBS subtypes within the sham, verum, and total sample. Finally, clinical heterogeneity of treatment effect (HTE) was assessed using a 1% absolute treatment effect difference as the clinically important threshold.
Results
Significant sub‐group × treatment interactions were found for resolution of MBS at 2 h (p = 0.008), pain relief at 2 h (p = 0.001), rescue medication between 2 and 24 h (p = 0.012), sustained pain freedom at 24 h (p = 0.033), and sustained pain relief at 24 h (p = 0.003). Significant sub‐group × treatment interactions were not found for pain freedom at 2 h (p = 0.054) or absence of all symptoms at 2 h (p = 0.265). Between treatment group comparisons indicated that pain freedom after 2 h of eTNS was not significantly different between the verum and sham groups for any pretreatment MBS. The verum group had a significantly greater proportion of participants who had resolution of nausea MBS after 2 h of treatment compared to sham (37/55 [67.3%] vs. 25/54 [46.3%], respectively; p = 0.028) and resolution of photophobia MBS compared to sham (85/162 [52.5] vs. 71/183 [38.8%], respectively; p = 0.011). There were no significant differences between treatment groups for phonophobia or vomiting. Pain freedom after 2 h of eTNS was not significantly different among pretreatment MBS groups. Within the sham group and total sample, a greater proportion of participants who had vomiting MBS had resolution of their MBS compared to any other pretreatment MBS (p < 0.05 after Bonferroni adjustment). A greater proportion of participants with nausea MBS used rescue medications between 2 and 24 h after eTNS compared to participants with photophobia or phonophobia MBS within the verum and total sample (p < 0.05 after Bonferroni adjustment). No statistical differences were found among MBS groups for any other treatment outcomes. Clinically important HTE was present in vomiting MBS for resolution of MBS and present in nausea MBS for pain freedom and pain relief after 2 h, need for rescue medication, and sustained pain freedom at 24 h post‐treatment. There was no clinically relevant HTE in the nausea MBS group for resolution of MBS at 2 h, absence of all migraine‐associated symptoms and sustained pain relief at 24 h, or for any endpoint for other MBS subtypes.
Conclusion
Our results suggest the presence of both statistically significant HTE as well as clinically meaningful HTE. Statistical differences were primarily found for photophobia MBS, while clinically meaningful HTE was primarily found for nausea MBS. These findings may be clinically relevant for patients and clinicians when developing a treatment plan for acute treatment of migraine. Further studies are needed to elucidate the underlying pathophysiological differences between MBS subtypes and treatment optimization, particularly for patients with nausea MBS subtypes.
Keywords: external trigeminal nerve stimulation, migraine, migraine‐associated symptoms, most bothersome symptom, neuromodulation
Plain Language Summary
External trigeminal nerve stimulation (eTNS) effectively treats migraine and most bothersome symptom (MBS); however, the effectiveness of eTNS according to MBS subtype has not been studied. We examined differences in treatment effects for each type of MBS (photophobia, nausea, phonophobia, and vomiting) between patients who received eTNS with therapeutic stimulation versus patients who received eTNS with non‐therapeutic stimulation. The therapeutic eTNS consistently improved migraine outcomes for most MBS subtypes, including photophobia, phonophobia, and vomiting; however, for nausea, migraine outcomes were variable suggesting that patients with nausea MBS subtype may require a stratified migraine treatment approach compared to other MBS subtypes and further studies are needed to examine the relationships between MBS subtypes and migraine treatment outcomes.
Abbreviations
- eTNS
external trigeminal nerve stimulation
- HTE
heterogeneity of treatment effect
- ipRGCs
intrinsically photosensitive retinal ganglion cells
- MBS
most bothersome symptom
- TEAM
Trial of eTNS for the Acute treatment of Migraine
INTRODUCTION
Migraine is among the most common and debilitating neurological conditions worldwide. 1 , 2 , 3 Migraine‐associated symptoms are classically sensory or autonomic deficits (i.e., photophobia, phonophobia, nausea, vomiting) that occur during a migraine headache, and their presence is a core feature for the diagnosis and management of migraine. 4 Migraine‐associated symptoms may persist, even in the absence of pain, and substantially contribute to migraine‐associated disability and intolerance to activity. 5 The United States Food and Drug Administration recommends that the migraine‐associated most bothersome symptom (MBS) be included as a primary efficacy endpoint in clinical trials assessing acute treatments for migraine due to its importance as a target symptom for treatment. 6
External trigeminal nerve stimulation (eTNS), manufactured by CEFALY Technology, is a non‐invasive neuromodulation treatment that transcutaneously stimulates the supraorbital and supratrochlear branches of the trigeminal nerve. 7 The eTNS device is cleared by the United States Food and Drug Administration for the acute and preventive treatment of migraine. In a recent prospective, multicenter, randomized, sham controlled trial, eTNS use during an active migraine resulted in significantly higher rates of resolution of migraine‐associated MBS when compared to sham stimulation (56.4% vs. 42.3%, respectively, p = 0.001). 8 However, to our current knowledge, the association between pretreatment MBS subtype and the efficacy of the eTNS treatment for active migraine attacks has not been examined. Enhancing our understanding of eTNS through the lens of MBS subtypes can provide patients and clinicians with the necessary insights to tailor and optimize individualized migraine treatment strategies for people living with migraine. To examine the relationships between eTNS and MBS subtypes, we performed a post hoc analysis of data from the Trial of eTNS for the Acute treatment of Migraine (TEAM) study. 8 We hypothesized that there would be no heterogeneity in the treatment effects found among the different MBS subtypes.
METHODS
Ethics
The TEAM study protocol and informed consent form were reviewed and approved by the ADVARRA (formerly IntegReview) Institutional Review Board, protocol #51401, on February 15, 2018. The study was conducted in accordance with Good Clinical Practice guidelines and the principles of the Declaration of Helsinki. All participants provided written informed consent prior to initiation of any study procedures.
Study design
Detailed methods and procedures for the TEAM study have been previously reported. 8 The TEAM study was a prospective, multicenter, randomized, double‐blind, sham‐controlled trial including adults (N = 538) meeting International Classification of Headache Disorders‐Third Edition criteria for migraine with and without aura and experiencing 2–8 migraine days/month. Enrollment occurred from April 10, 2018, to January 11, 2019. Participants were recruited and randomized in a 1:1 ratio to verum (n = 259) or sham (n = 279) stimulation groups. Study participants and care providers were blinded to intervention assignments. Both device groups used identical appearing devices that delivered transcutaneous neurostimulation via a 30 × 94 mm2 self‐adhesive electrode placed on the forehead and covering the supratrochlear and supraorbital nerves bilaterally (Supplemental Figure S1). For both groups, neurostimulation consisted of rectangular biphasic symmetrical impulses with an electrical mean equal to zero, pulse width of 250 μs. and linearly increasing stimulation intensity of up to 16 mA during the first 14 min, which remained constant for the remainder of the treatment. The verum device produced high‐frequency pulses (100 Hz), and the sham device produced low‐frequency pulses (3 Hz). Participants were trained to operate the eTNS device and then were tested with a 20‐min nociceptive assessment to ensure the capability to self‐apply, operate, and tolerate the device stimulation. 9
Following a 2‐h eTNS treatment, participants self‐reported migraine pain, MBS, and rescue medication use between 2 and 24 h using a paper diary. Participants had to treat at least one qualifying migraine within a 2‐month period, defined as a migraine with moderate or severe headache intensity with at least one migraine‐associated symptom, being photophobia, phonophobia, nausea, and/or vomiting.
Primary and secondary outcomes from the TEAM study are listed in Table 1. Power analysis from the TEAM study indicated that a minimum of 239 patients per group was needed to provide 90% power to detect a 13.4% difference in the primary endpoint, pain freedom at 2 h. This post hoc analysis examined the primary and secondary treatment outcomes by pretreatment migraine‐associated MBS, including photophobia (n = 345), nausea (n = 109), phonophobia (n = 73), and vomiting (n = 11). These subgroups are likely underpowered to detect treatment differences; however, these results may serve as valuable preliminary evidence and generate hypotheses for future investigations focusing on specific types of migraine‐associated MBS.
TABLE 1.
Primary and secondary outcomes from the TEAM study. 8
| Primary outcomes (post‐treatment) |
| Pain freedom at 2 h |
| Resolution of most bothersome migraine associated symptoms (MBS) at 2 h |
| Secondary outcomes (posttreatment) |
| Pain relief at 2 h (reduction of moderate–severe to mild or no pain) |
| Resolution of any migraine‐associated symptoms at 2 h |
| Sustained pain freedom at 24 h |
| Sustained pain relief at 24 h |
| Use of rescue anti‐migraine therapy between 2 and 24 h |
Abbreviation: TEAM, Trial of eTNS for the Acute Treatment of Migraine.
Statistical analysis
There were no missing data for participants. All data were checked for normality of distribution. Normally distributed continuous data are presented as mean and standard deviation, and non‐normally distributed continuous data are presented as median, interquartile range. Categorical data are presented as frequencies and percentages. Patient characteristics were compared between treatment groups for each pretreatment MBS using independent samples t‐test or Mann–Whitney U test for continuous data or chi‐square analysis for categorical data.
In this post hoc analysis, we explored post‐treatment outcomes by pretreatment migraine‐associated MBS. First, we performed logistic regression for each post‐treatment outcome and included MBS sub‐group × treatment arm as an interaction term in the model. This approach explored the potential of a differential effect conditional on the total sample. We further explored direct treatment group comparisons for each MBS sub‐group to extrapolate results for potential pilot data for future clinical trials that may focus on particular MBS sub‐groups. These comparisons were done using chi‐square analysis or Fisher's exact test. We also compared treatment outcomes among all MBS within the sham, the verum, and total sample. The within group comparisons were done using chi‐square analysis with Bonferroni correction for multiple comparisons. Finally, heterogeneity of treatment effect (HTE) was assessed using a 1% absolute treatment effect difference as the clinically relevant threshold based on the device being minimal risk. HTE was assessed following methods from the PATH (Predictive Approaches to Treatment effect Heterogeneity) Statement. 10 Analysis was performed using the Statistical Package for the Social Sciences (SPSS), version 26.0 (IBM Corp., Armonk, NY, USA), with statistical significance established at a two‐tailed p < 0.05.
RESULTS
Patient characteristics by treatment group for each pretreatment MBS are shown in Table 2. The most reported MBS was photophobia (64.1%), followed by nausea (20.3%), phonophobia (13.6%), and vomiting (2.0%). Participants (N = 538) had a mean (standard deviation) age of 41.1 (2.3) years, and the majority (82.3%) were female. There were no statistically significant differences in patient characteristics between the verum and sham groups for any pretreatment MBS.
TABLE 2.
Patient characteristics.
| Nausea (n = 109) | Photophobia (n = 345) | Phonophobia (n = 73) | Vomiting (n = 11) | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Verum (n = 55) | Sham (n = 54) | p | Verum (n = 162) | Sham (n = 183) | p | Verum (n = 38) | Sham (n = 35) | p | Verum (n = 4) | Sham (n = 7) | p | |
| Age, years, mean (SD) | 40.8 (11.2) | 44.3 (11.8) | 0.117 | 39.5 (11.4) | 41.8 (12.3) | 0.072 | 42.1 (12.7) | 39.0 (12.9) | 0.301 | 45.4 (16.4) | 45.4 (11.9) | 0.992 |
| Sex, n (%) | 0.297 | 0.131 | 0.284 | >0.999 | ||||||||
| Female | 45 (81.8) | 48 (88.9) | 140 (86.4) | 147 (80.3) | 25 (65.8) | 27 (77.1) | 4 (100.0) | 7 (100.0) | ||||
| Male | 10 (18.2) | 6 (11.1) | 22 (13.6) | 36 (19.7) | 13 (34.2) | 8 (22.9) | 0 (0.0) | 0 (0.0) | ||||
| Migraine with aura, n (%) | 15 (27.3) | 19 (35.2) | 0.376 | 74 (45.7) | 77 (42.1) | 0.502 | 20 (52.6) | 13 (37.1) | 0.187 | 4 (100.0) | 2 (28.6) | 0.061 |
| Headache pain at baseline, n (%) | 0.263 | 0.423 | 0.963 | >0.999 | ||||||||
| No pain | 0 (0.0) | 0 (0.0) | 1 (0.6) | 0 (0.0) | 0 (0.0) | 0 (0.0) | 0 (0.0) | 0 (0.0) | ||||
| Mild | 0 (0.0) | 0 (0.0) | 1 (0.6) | 0 (0.0) | 0 (0.0) | 0 (0.0) | 0 (0.0) | 0 (0.0) | ||||
| Moderate | 32 (58.2) | 37 (68.5) | 98 (60.5) | 105 (57.4) | 23 (60.5) | 21 (60.0) | 2 (50.0) | 3 (42.9) | ||||
| Severe | 23 (41.8) | 17 (31.5) | 62 (38.3) | 78 (42.6) | 15 (39.5) | 14 (40.0) | 2 (50.0) | 4 (57.1) | ||||
| Treatment duration, min, median (IQR) | 120.0 (0.0) | 120.0 (11.0) | 0.463 | 120.0 (31.0) | 120.0 (41.0) | 0.182 | 120.0 (35.0) | 120.0 (0.0) | 0.112 | 13.5 (86.0) | 119.0 (70.0) | 0.099 |
Abbreviations: IQR, interquartile range; SD, standard deviation.
The plots in Figures 1 and 2 provide a visual depiction of the HTE for eTNS. Significant sub‐group × treatment interactions (Figures 1 and 2, column B) were found for post‐treatment endpoints, including resolution of MBS at 2 h (p = 0.008), pain relief at 2 h (p = 0.001), rescue medication between 2 and 24 h (p = 0.012), sustained pain freedom at 24 h (p = 0.033), and sustained pain relief at 24 h (p = 0.003). Sub‐group × treatment interactions were not significant for pain freedom at 2 h (p = 0.054) or absence of all symptoms at 2 h post‐treatment (p = 0.265).
FIGURE 1.
Plots depicting heterogeneity of treatment effect (HTE) for primary outcomes. MBS, most bothersome symptom. (A) The percentage of patients by pretreatment MBS sub‐group for verum and sham that had the treatment outcome. (B) MBS sub‐group × treatment group statistical interaction and verum versus sham odds ratios for each pretreatment MBS sub‐group. Statistical significance indicates the presence of statistically significant HTE. (C) The absolute treatment effect (verum minus sham) with a 1% absolute treatment effect difference as the clinically relevant threshold. Clinically important HTE is indicated when an absolute effect deviates from other sub‐groups in reference to the clinically important threshold (dotted line). [Colour figure can be viewed at wileyonlinelibrary.com]
FIGURE 2.
Plots depicting heterogeneity of treatment effect (HTE) for secondary outcomes. MBS, most bothersome symptom. (A) The percentage of patients by pretreatment MBS sub‐group for verum and sham that had the treatment outcome. (B) MBS sub‐group × treatment group statistical interaction and verum versus sham odds ratios for each pretreatment MBS sub‐group. Statistical significance indicates the presence of statistically significant HTE. (C) The absolute treatment effect (verum minus sham) with a 1% absolute treatment effect difference as the clinically relevant threshold. Clinically important HTE is indicated when an absolute effect deviates from other sub‐groups in reference to the clinically important threshold (dotted line). [Colour figure can be viewed at wileyonlinelibrary.com]
Clinically important HTE (Figure 1, column C) was present for resolution of MBS at 2 h post‐treatment with vomiting MBS (0%) below the 1% absolute effect threshold and all other MBS subtypes above the threshold. Similarly, nausea MBS was the only MBS subtype below the clinically important absolute effect threshold for pain freedom at 2 h (−4%), pain relief at 2 h (0.8%), and sustained pain freedom at 24 h (−4%) post‐treatment (Figure 2, column C). Nausea MBS was the only MBS subtype with an absolute effect above the clinically important threshold for rescue medication between 2 and 24 h (6.5%; positive value favors sham for this endpoint). There was no presence of clinically important HTE for absence of all symptoms at 2 h and sustained pain relief at 24 h post‐treatment (Figure 2, column C).
Between treatment group analysis
For the primary outcome, pain freedom after 2 h of eTNS treatment, there were no statistically significant differences between the verum and sham groups for any pretreatment MBS (Table 3). For the primary outcome, resolution of MBS after 2 h of eTNS treatment, the verum group had a significantly greater proportion of participants who had resolution of nausea compared to sham and resolution of photophobia compared to sham (Table 3). There were no significant differences between treatment groups for phonophobia or vomiting (Table 3).
TABLE 3.
Outcomes for each migraine‐associated most bothersome symptom: verum versus sham.
| MBS treatment group | Pain freedom at 2 h | Resolution of MBS at 2 h | Pain relief at 2 h | Absence of all symptoms at 2 h | Rescue medication between 2 and 24 h | Sustained pain freedom at 24 h | Sustained pain relief at 24 h |
|---|---|---|---|---|---|---|---|
| Nausea (n = 109), n (%) | |||||||
| Verum (n = 55) | 9 (16.4) | 37 (67.3) | 31 (56.4) | 22 (40.0) | 28 (50.9) | 7 (12.7) | 17 (30.9) |
| Sham (n = 54) | 11 (20.4) | 25 (46.3) | 30 (55.6) | 20 (37.0) | 24 (44.4) | 9 (16.7) | 14 (25.9) |
| p | 0.592 | 0.028 | 0.933 | 0.751 | 0.499 | 0.557 | 0.565 |
| Photophobia (n = 345), n (%) | |||||||
| Verum (n = 162) | 44 (27.2) | 85 (52.5) | 120 (74.1) | 68 (42.0) | 46 (28.4) | 40 (24.7) | 80 (49.4) |
| Sham (n = 183) | 34 (18.6) | 71 (38.8) | 102 (55.7) | 61 (33.3) | 65 (35.5) | 29 (15.8) | 66 (36.1) |
| p | 0.057 | 0.011 | <0.001 | 0.096 | 0.159 | 0.039 | 0.013 |
| Phonophobia (n = 73), n (%) | |||||||
| Verum (n = 38) | 12 (31.6) | 20 (52.6) | 27 (71.1) | 18 (47.4) | 7 (18.4) | 11 (28.9) | 20 (52.6) |
| Sham (n = 35) | 5 (14.3) | 15 (42.9) | 20 (57.1) | 11 (31.4) | 12 (34.4) | 5 (14.3) | 14 (40.0) |
| p | 0.083 | 0.410 | 0.228 | 0.166 | 0.122 | 0.135 | 0.284 |
| Vomiting (n = 11), n (%) | |||||||
| Verum (n = 4) | 1 (25.0) | 4 (100.0) | 2 (50.0) | 2 (50.0) | 1 (25.0) | 1 (25.0) | 2 (50.0) |
| Sham (n = 7) | 1 (14.3) | 7 (100.0) | 2 (28.6) | 3 (42.9) | 4 (57.1) | 1 (14.3) | 2 (28.6) |
| p | >0.999 | >0.999 | 0.576 | >0.999 | 0.545 | >0.999 | 0.576 |
Note: Fisher's exact test used for ‘vomiting’ comparison due to small cell counts.
Abbreviation: MBS, most bothersome symptom.
Comparisons of secondary outcomes between treatment groups for each pretreatment MBS are shown in Table 3. For photophobia, the proportion of patients with pain relief at 2 h, sustained pain freedom at 24 h, and sustained pain relief at 24 h were all significantly greater in the verum group compared to sham group. There were no statistically significant differences between treatment groups for any other pretreatment MBS (Table 3).
Within treatment group analysis
There were no statistically significant differences in pain freedom (primary outcome) after 2 h of eTNS treatment among pretreatment MBS groups (Table 4). Within the sham group and total sample, there was a significantly greater proportion of participants who had vomiting as their pretreatment MBS and had resolution of their MBS (primary outcome) compared to participants with any other pretreatment MBS (Table 4). At 2 h after eTNS treatment, all participants with vomiting as their pretreatment MBS were symptom‐free of vomiting, regardless of their treatment group.
TABLE 4.
Outcomes by migraine‐associated most bothersome symptom (MBS) within each treatment group.
| Group outcome, frequency (%) | Migraine‐associated most bothersome symptoms (MBS) | p | ||||
|---|---|---|---|---|---|---|
| Total sample | Nausea (n = 109) | Photophobia (n = 345) | Phonophobia (n = 73) | Vomiting (n = 11) | Total (N = 538) | |
| Pain freedom at 2 h | 20 (18.3) | 78 (22.6) | 17 (23.3) | 2 (18.2) | 117 (21.7) | 0.783 |
| Resolution of MBS at 2 h | 62 (56.9)b | 156 (45.2)b | 35 (47.9)b | 11 (100.0)a | 264 (49.1) | 0.001 |
| Pain relief at 2 h | 61 (56.0) | 222 (64.3) | 47 (64.4) | 4 (36.4) | 334 (62.1) | 0.125 |
| Absence of all symptoms at 2 h | 42 (38.5) | 129 (37.4) | 29 (39.7) | 5 (45.5) | 205 (38.1) | 0.937 |
| Rescue medication between 2 and 24 h | 52 (47.7)a | 111 (32.2)b | 19 (26.0)b | 5 (45.5)a,b | 187 (34.8) | 0.007 |
| Sustained pain freedom at 24 h | 16 (14.7) | 69 (20.0) | 16 (21.9) | 2 (18.2) | 103 (19.1) | 0.586 |
| Sustained pain relief at 24 h | 31 (28.4) | 146 (42.3) | 34 (46.6) | 4 (36.4) | 215 (40.0) | 0.042 |
| Verum | Nausea (n = 55) | Photophobia (n = 162) | Phonophobia (n = 38) | Vomiting (n = 4) | Total (N = 259) | p |
|---|---|---|---|---|---|---|
| Pain freedom at 2 h | 9 (16.4) | 44 (27.2) | 12 (31.6) | 1 (25.0) | 66 (25.5) | 0.335 |
| Resolution of MBS at 2 h | 37 (67.3) | 85 (52.5) | 20 (52.6) | 4 (100.0) | 146 (56.4) | 0.073 |
| Pain relief at 2 h | 31 (56.4) | 120 (74.1) | 27 (71.1) | 2 (50.0) | 180 (69.5) | 0.077 |
| Absence of all symptoms at 2 h | 22 (40.0) | 68 (42.0) | 18 (47.4) | 2 (50.0) | 110 (42.5) | 0.892 |
| Rescue medication between 2 and 24 h | 28 (50.9)a | 46 (28.4)b | 7 (18.4)b | 1 (25.0)a,b | 82 (31.7) | 0.004 |
| Sustained pain freedom at 24 h | 7 (12.7) | 40 (24.7) | 11 (28.9) | 1 (25.0) | 59 (22.8) | 0.228 |
| Sustained pain relief at 24 h | 17 (30.9) | 80 (49.4) | 20 (52.6) | 2 (50.0) | 119 (45.9) | 0.090 |
| Sham | Nausea (n = 54) | Photophobia (n = 183) | Phonophobia (n = 35) | Vomiting (n = 7) | Total (N = 279) | p |
|---|---|---|---|---|---|---|
| Pain freedom at 2 h | 11 (20.4) | 34 (18.6) | 5 (14.3) | 1 (14.3) | 51 (18.3) | 0.892 |
| Resolution of MBS at 2 h | 25 (46.3)b | 71 (38.8)b | 15 (42.9)b | 7 (100.0)a | 118 (42.3) | 0.013 |
| Pain relief at 2 h | 30 (55.6) | 102 (55.7) | 20 (57.5) | 2 (28.6) | 154 (55.2) | 0.555 |
| Absence of all symptoms at 2 h | 20 (37.0) | 61 (33.3) | 11 (31.4) | 3 (42.9) | 95 (34.1) | 0.895 |
| Rescue medication between 2 and 24 h | 24 (44.4) | 65 (35.5) | 12 (34.4) | 4 (57.1) | 105 (37.6) | 0.437 |
| Sustained pain freedom at 24 h | 9 (16.7) | 29 (15.8) | 5 (14.3) | 1 (14.3) | 44 (15.8) | 0.991 |
| Sustained pain relief at 24 h | 14 (25.9) | 66 (36.1) | 14 (40.0) | 2 (28.6) | 96 (34.4) | 0.469 |
Note: Differing superscript letters denote significant differences between groups after Bonferroni adjustment at p < 0.05. Individual comparisons were not significant after Bonferroni adjustment for sustained pain relief at 24 h in the total sample.
For secondary outcomes (Table 4), a greater proportion of participants with nausea as their pretreatment MBS used rescue medications between 2 and 24 h after eTNS treatment compared to participants with pretreatment MBS of photophobia and phonophobia within the verum and total sample. No differences were found among MBS groups for any other secondary treatment outcomes.
DISCUSSION
The TEAM study demonstrated that eTNS is a safe and effective treatment for acute migraine attacks and resolution of the MBS. 8 The TEAMS study was the first clinical trial to examine the safety and efficacy of eTNS with migraine‐associated symptoms including MBS as an endpoint. While other therapies have examined MBS relief/resolution as an endpoint in clinical trials, 11 , 12 this post hoc analysis is the first among migraine neuromodulation therapies to examine safety and efficacy endpoints based on the pretreatment MBS subtype. Understanding the relationship between migraine neuromodulatory effects on MBS subtype is novel and a clinically relevant tool for patients and clinicians to individualize migraine management. For example, heterogeneity in the efficacy of eTNS based on the pretreatment MBS subtype can guide patients and clinicians in considering which primary and adjunctive therapies are most appropriate for them. Results of this analysis indicate the presence of both statistically significant HTE as well as clinically meaningful HTE (Table 5). Statistical differences were primarily found for photophobia MBS, while clinically meaningful HTE was primarily found for nausea MBS.
TABLE 5.
Summary of heterogeneity of treatment effect (HTE) findings.
| Outcome following eTNS | Statistically significant HTE found? | Clinically important HTE found? |
|---|---|---|
| Pain freedom at 2 h | No | Yes; nausea under 1% |
| Resolution of MBS at 2 h | Yes; ↑ in verum for nausea and photophobia; vomiting ↑ than other MBS types in sham and total sample | Yes; vomiting under 1% |
| Pain relief at 2 h | Yes; verum favored for photophobia | Yes; nausea under 1% |
| Absence of all symptoms at 2 h | No | No |
| Rescue medication between 2 and 24 h | Yes; nausea ↑ than other MBS types in verum and total sample | Yes; nausea above 1% |
| Sustained pain freedom at 24 h | Yes; ↑ in verum for photophobia | Yes; nausea under 1% |
| Sustained pain relief at 24 h | Yes; ↑ in verum for photophobia | No |
Note: Clinically important HTE based on the absolute effect: Verum minus Sham with a 1% clinically important threshold. ↑ indicates significantly greater in verum group compared to sham group.
Abbreviation: eTNS, external trigeminal nerve stimulation.
Patients who received verum stimulation with photophobia as their pretreatment MBS were more likely to experience resolution of their photophobia after 2 h of treatment, pain relief at 2 h, sustained pain freedom at 24 h after treatment, and sustained pain relief at 24 h compared to patients within the sham group with photophobia as their MBS. Despite photophobia being the most common migraine‐associated symptom, reported in up to 80% of patients with migraine, 13 the pathophysiology of migraine‐associated photophobia is complex and incompletely elucidated. Our current understanding of migraine‐associated photophobia involves multiple levels of interconnected visual, trigemino‐vascular, brainstem, subcortical, and autonomic pathways. Recent translational studies have examined intrinsically photosensitive retinal ganglion cells (ipRGCs) that express melanopsin photopigment as a main participant in migraine‐associated photophobia. 14 , 15 Electrophysiological and blood‐oxygen‐level‐dependent functional magnetic resonance imaging have identified anatomical and functional connectivity with ipRGCs and thalamic trigeminovascular neurons via lateral posterior and posterior thalamic nuclei. 16 , 17 The convergence of photic signals from ipRGCs to the trigeminovascular thalamo‐cortical pathways is a proposed mechanism to explain how light intensity may worsen an active migraine headache. 18
The mechanisms underlying how eTNS may resolve photophobia MBS are unknown. One possible explanation is that eTNS reduces trigeminovascular excitation via stimulation through the supraorbital and supratrochlear nerves, thereby indirectly reducing central sensory hypersensitization mediated by thalamo‐cortical networks. 19 Another potential mechanism is that eTNS modulates or reduces brainstem parasympathetic outflow via the superior salivatory nucleus and sphenopalatine ganglion thereby reducing release of nitric oxide, vasoactive intestinal polypeptide and pituitary adenylate cyclase‐activating polypeptide (PACAP), the latter of which may directly be involved in ipRGCs neurotransmission. 20 Future functional and physiological studies are needed to examine the relationship between trigeminal nerve stimulation and photic hypersensitivity responses during ictal migraine.
Our results indicate that patients in the verum group with nausea as their pretreatment MBS were more likely to experience resolution of their MBS after 2 h of treatment compared to patients in the sham group. Comparisons among different pretreatment MBS subtypes within the verum and total sample revealed a significantly greater proportion of participants with pretreatment nausea MBS used rescue medications between 2 and 24 h compared to participants with pretreatment photophobia and phonophobia MBS. The lack of difference in post‐treatment rescue medication use across MBS subtypes in the sham group can be due to the consistently higher need for rescue medications between 2 and 24 h across all pretreatment MBS subtypes within this group. In addition, there were no significant differences in the rates of rescue medication use for each MBS subtype when comparing sham and verum groups.
Nausea is the second most common MBS reported in 28% of patients with migraine. 21 , 22 The pathophysiology of migraine‐associated nausea remains unknown, though the current literature implicates several central brainstem neuroanatomical structures and neurotransmitters/neuropeptides such as dopamine, serotonin, substance P, and neurokinin‐1. 23 While connections between trigeminal neurons and nucleus tractus solitarius were thought to underlie the cause of nausea MBS, nausea can occur as a premonitory symptom prior to the ictal headache. A study in 2014 found significant positron emission tomography H2 15O activation in patients experiencing premonitory nausea without ictal headache in the mesencephalic periaqueductal gray, dorsal motor nucleus of the vagus, nucleus ambiguous and nucleus tractus solitaries. 24 This suggests that nausea may emerge as a migraine‐associated symptom independent of trigeminal nerve activation.
While resolution of nausea MBS was significantly higher in the verum group compared to the sham group, the mechanism by which eTNS may improve or resolve nausea MBS is unknown, and further studies are needed to clarify the neurophysiological relationships between trigeminal nerve stimulation and migraine associated nausea. Our findings suggest that while eTNS is effective in reducing nausea MBS, clinically meaningful HTE was present for nausea MBS for pain relief at 2 h and sustained pain freedom at 2 h, suggesting that the response of migraine pain to eTNS for patients with nausea MBS may be inconsistent, especially relative to other MBS subtypes. Congruent with prior literature, this may provide further support that migraine‐associated nausea is a centrally mediated symptom that may evolve somewhat independently or indirectly from the trigeminovascular system. Further studies are needed to elucidate the underlying pathophysiological mechanisms of migraine pain in patients with nausea MBS subtypes.
Our analysis implies that patients with pretreatment MBS of nausea may also benefit from rescue medication concurrent with or following eTNS treatment when compared to other MBS subtypes. Among patients with nausea as their pretreatment MBS, 52/109 (48%) used rescue medications. The most common rescue medication was triptans (24/52 [46%]), followed by aspirin/paracetamol/caffeine (10/52 [19%]) and non‐steroidal anti‐inflammatory drugs (six of 52 [12%]). Only two of the 52 patients (4%) used antiemetics with or without additional rescue therapy. A full description of rescue medications used in the nausea MBS group is included in Supplemental Table S1.
Most patients with pretreatment nausea MBS used rescue medications, which generally prioritize relief from migraine pain as opposed to resolution of nausea. For example, antiemetics accounted for only 4% (two of 52) of the rescue medications used in this group. Although these data should be interpreted with caution given the unknown access to antiemetic therapies from the patient's medical providers, these data may suggest that the migraine pain among patients with nausea MBS may exhibit some refractory features to eTNS therapy. These findings are consistent with prior evidence, 21 , 25 , 26 and, therefore, patients with migraine‐associated nausea or nausea as their MBS may need to be more robustly treated with a combination of treatments that include an antiemetic. Further studies are needed to elucidate the cost‐efficacy of eTNS in patients who experience migraine with pretreatment nausea MBS.
While the rate of vomiting MBS in the TEAM study was proportionately representative of the prevalence in persons with migraine, 21 the interpretation of vomiting MBS compared to other MBS subtypes should be exercised with caution due to relatively lower sample sizes (Table 4). For example, 100% of the 11 patients with vomiting MBS had resolution of their pretreatment MBS, which resulted in a 0% treatment effect and thus displayed clinically important HTE. In addition, within the sham group and total sample, a greater proportion of participants with vomiting as their pretreatment MBS reported resolution of MBS compared to any other pretreatment MBS. However, this difference in post‐treatment resolution of MBS between MBS subtypes results was not found within the verum group. This finding may be explained by the homogeneity and consistency of higher rates of resolution of MBS found within the verum group across all MBS subtypes.
Study limitations and strengths
The present study has limitations. The study results are from a post hoc analysis and, therefore, should be confirmed in a future study that is adequately powered for each of the four MBS groups. To minimize this limitation, adjustment was made for multiple comparisons to decrease the likelihood of Type I error for the within group comparisons. In addition, the sample size for the group of patients with vomiting as their MBS was small (11 patients), and, therefore, any findings for this group should be interpreted with caution. The prevalence of vomiting in this study accurately represents the prevalence in persons with migraine. 21 Finally, this post hoc analysis explored HTE, which presents several challenges in a post hoc analysis. HTE can be interpreted both statistically and using absolute treatment effects with a clinically meaningful threshold. However, clinically meaningful HTE should not be conflated with interpretations from statistical testing. Furthermore, results from HTE analysis should not be implemented in clinical practice until external validation and calibration of prediction have been conducted. 10 Therefore, HTE results from this post hoc analysis should be further evaluated before being considered valid for clinical practice.
This study also presents with strengths. These data were derived from a rigorously designed, sham randomized controlled trial. In addition, the MBS data being evaluated in this post hoc analysis was the original trial's primary endpoint. To our knowledge, this post hoc analysis is the first to examine the efficacy of eTNS by each pretreatment MBS subtype according to the primary and secondary endpoints established in the original clinical trial. The analysis of MBS subtype and statistical approach demonstrated in this post hoc analysis should serve as a framework for future studies examining migraine safety and efficacy.
CONCLUSION
Our results suggest the presence of both statistically significant HTE, as well as clinically important HTE, when using eTNS therapy for acute treatment of headache. Clinically important HTE was found for patients who have nausea MBS, indicating less of a treatment effect compared to sham relative to other MBS subtypes for pain freedom at 2 h and sustained pain freedom at 24 h. However, for nausea MBS, the treatment effect was homogenous in favor of verum for MBS resolution at 2 h, resolution of migraine‐associated symptoms at 2 h, and pain relief at 24 h post‐treatment. There was no clinically important heterogeneity among outcomes for patients with photophobia, phonophobia, or vomiting MBS.
These findings may be clinically relevant for patients and clinicians when developing a treatment plan for acute treatment of migraine. Although patients with photophobia, phonophobia, and vomiting MBS may experience consistently favorable outcomes when using eTNS alone, patients with nausea MBS may be more likely to have recalcitrant migraine pain and require a stratified, multi‐step treatment approach to include other rescue medications in addition to eTNS. In addition, these findings further support prior evidence that migraine phenotypes and response to treatment may differ among patients with nausea MBS compared to other MBS subtypes. 21
Further studies are needed to elucidate the underlying pathophysiological differences between MBS subtypes and treatment optimization. Future studies should also examine the cost‐efficacy of eTNS in the acute treatment of migraine, especially among patients with nausea MBS. This post hoc analysis may serve as a model for future clinical trials assessing potential heterogeneity among MBS subtypes when optimizing acute treatment of migraine.
AUTHOR CONTRIBUTIONS
Gregory A. Panza: Conceptualization; formal analysis; methodology; writing – original draft; writing – review and editing. Michael A. L. Johnson: Conceptualization; methodology; writing – review and editing. Deena E. Kuruvilla: Conceptualization; methodology; writing – review and editing.
FUNDING INFORMATION
This study was funded by CEFALY Technology.
CONFLICT OF INTEREST STATEMENT
Drs. Gregory A. Panza, Michael A. L. Johnson, and Deena E. Kuruvilla are consultants for Cefaly Technology.
Supporting information
Table S1.
Figure S1.
Panza GA, Johnson MAL, Kuruvilla DE. A post hoc analysis of migraine‐associated symptoms from the phase 3 randomized, double‐blind, sham‐controlled Trial of External trigeminal nerve stimulation for the Acute treatment of Migraine (TEAM) study. Headache. 2025;65:779‐790. doi: 10.1111/head.14860
REFERENCES
- 1. Vos T, Abajobir AA, Abate KH, et al. Global, regional, and national incidence, prevalence, and years lived with disability for 328 diseases and injuries for 195 countries, 1990‐2016: a systematic analysis for the global burden of disease study. Lancet. 2017;390(10100):1211‐1259. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. GBD 2019 Diseases and Injuries Collaborators . Global burden of 369 diseases and injuries in 204 countries and territories, 1990–2019: a systematic analysis for the global burden of disease study 2019. Lancet. 2020;396:1204‐1222. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Steiner TJ, Stovner LJ, Jensen R, Uluduz D, Katsarava Z. Lifting the burden: the global campaign against headache. Migraine remains second among the world's causes of disability, and first among young women: findings from GBD2019. J Headache Pain. 2020;21(1):137. doi: 10.1186/s10194-020-01208-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Headache Classification Committee of the International Headache Society (IHS) . The international classification of headache disorders, 3rd edition. Cephalalgia. 2018;38(1):1‐211. doi: 10.1177/0333102417738202 [DOI] [PubMed] [Google Scholar]
- 5. Lipton RB, Dodick D, Sadovsky R, et al. A self‐administered screener for migraine in primary care: the ID Migraine validation study. Neurology. 2003;61(3):375‐382. doi: 10.1212/01.wnl.0000078940.53438.83 [DOI] [PubMed] [Google Scholar]
- 6. U.S. Department of Health and Human Services , Food and Drug Administration , Center for Drug Evaluation and Research (CDER) . Migraine: Developing Drugs for Acute Treatment Guidance for Industry. Accessed March 31, 2024. https://www.fda.gov/downloads/drugs/guidances/ucm419465.pdf
- 7. Johnson MAJ, Kuruvilla DE. External trigeminal nerve stimulation as a non‐pharmacological option for the prevention and acute treatment of migraine. Reviews in Touch Neurology. 2022;10(1):22. doi: 10.17925/USN.2022.18.1.22 [DOI] [Google Scholar]
- 8. Kuruvilla DE, Mann JI, Tepper SJ, Starling AJ, Panza G, Johnson MAL. Phase 3 randomized, double‐blind, sham‐controlled trial of eTNS for the acute treatment of migraine (TEAM). Sci Rep. 2022;12(1):5110. doi: 10.1038/s41598-022-09071-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Lauritsen CG, Silberstein SD. Rationale for electrical parameter determination in external trigeminal nerve stimulation (eTNS) for migraine: a narrative review. Cephalalgia. 2019;39(6):750‐760. [DOI] [PubMed] [Google Scholar]
- 10. Kent DM, Van Klaveren D, Paulus JK, et al. The predictive approaches to treatment effect heterogeneity (PATH) statement: explanation and elaboration. Ann Intern Med. 2020;172(1):W1‐W25. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Yarnitsky D, Dodick DW, Grosberg BM, et al. Remote electrical neuromodulation (REN) relieves acute migraine: a randomized, double‐blind, placebo‐controlled, multicenter trial. Headache. 2019;59(8):1240‐1252. doi: 10.1111/head.13551 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Tepper SJ, Grosberg B, Daniel O, et al. Migraine treatment with external concurrent occipital and trigeminal neurostimulation‐a randomized controlled trial. Headache. 2022;62(8):989‐1001. doi: 10.1111/head.14350 [DOI] [PubMed] [Google Scholar]
- 13. Digre KB, Brennan KC. Shedding light on photophobia. J Neuroophthalmol. 2021;32(1):68‐81. doi: 10.1097/WNO.0b013e3182474548 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Berson DM, Dunn FA, Takao M. Phototransduction by retinal ganglion cells that set the circadian clock. Science. 2002;295(5557):1070‐1073. doi: 10.1126/science.1067262 [DOI] [PubMed] [Google Scholar]
- 15. Lucas RJ, Hattar S, Takao M, Berson DM, Foster RG, Yau KW. Diminished pupillary light reflex at high irradiances in melanopsin‐knockout mice. Science. 2003;299(5604):245‐247. doi: 10.1126/science.1077293 [DOI] [PubMed] [Google Scholar]
- 16. Noseda R, Jakubowski M, Kainz V, Borsook D, Burstein R. Cortical projections of functionally identified thalamic trigeminovascular neurons: implications for migraine headache and its associated symptoms. J Neurosci. 2011;31:14204‐14217. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Maleki N, Becerra L, Upadhyay J, Burstein R, Borsook D. Direct optic nerve pulvinar connections defined by diffusion MR tractography in humans: implications for photophobia. Hum Brain Mapp. 2011;33:75‐88. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Noseda R, Copenhagen D, Burstein R. Current understanding of photophobia, visual networks and headaches. Cephalalgia. 2019;39(13):1623‐1634. doi: 10.1177/0333102418784750 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Coppola G, Ambrosini A, Di Clemente L, et al. Interictal abnormalities of gamma band activity in visual evoked responses in migraine: an indication of thalamocortical dysrhythmia? Cephalalgia. 2007;27(12):1360‐1367. doi: 10.1111/j.1468-2982.2007.01466.x [DOI] [PubMed] [Google Scholar]
- 20. Hannibal J. Melanopsin is expressed in PACAP‐containing retinal ganglion cells of the human retinohypothalamic tract. Invest Ophthalmol Vis Sci. 2004;45(11):4202‐4209. [DOI] [PubMed] [Google Scholar]
- 21. Munjal S, Sagar P, et al. Most bothersome symptom in persons with migraine: results from the Migraine in America Symptoms and Treatment (MAST) study. Headache. 2020;60(2):416‐442. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Lipton RB, Buse DC, Saiers J, Fanning KM, Serrano D, Reed ML. Frequency and burden of headache‐related nausea: results from the American Migraine Prevalence and Prevention (AMPP) study. Headache. 2013;53(1):93‐103. [DOI] [PubMed] [Google Scholar]
- 23. Villar‐Martinez MD, Goadsby PJ. Pathophysiology and therapy of associated features of migraine. Cells. 2022;11(17):2767. doi: 10.3390/cells11172767 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Maniyar FH, Sprenger T, Schankin C, Goadsby PJ. The origin of nausea in migraine–a PET study. J Headache Pain. 2014;15:84. doi: 10.1186/1129-2377-15-84 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Blumenfeld AM, Lipton RB, Silberstein S, et al. Multimodal migraine management and the pursuit of migraine freedom: a narrative review. Neurol Ther. 2023;12(5):1533‐1551. doi: 10.1007/s40120-023-00529-x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Kuruvilla DE, Bilchik T. Preventing the progression of episodic Migraine to chronic migraine with acute treatment optimization. Curr Pain Headache Rep. 2022;26(3):253‐258. doi: 10.1007/s11916-022-01022-0 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Table S1.
Figure S1.