The impact of external trigeminal nerve stimulator (e-TNS) on prevention and acute treatment of episodic and chronic migraine: A systematic review

Open AccessPublished:February 06, 2020DOI:https://doi.org/10.1016/j.jns.2020.116725

      Highlights

      • E-TNS has the potential to prevent and improve migraine symptoms.
      • No serious adverse events occurred in any of the studies.
      • The quality of evidence was low to very low.
      • Studies with larger sample sizes and relevant comparators needed.

      Abstract

      Objective

      The aim of this systematic review was to analyze the effectiveness and safety of the external trigeminal nerve stimulator (eTNS) for the prevention and acute treatment of migraine attacks in episodic and chronic migraine patients.

      Methods

      We have conducted a systematic literature search in four databases that yielded 433 citations and additional seven citations were found via hand-search. Two randomised placebo-controlled trials and five prospective case series were included in the analysis.

      Results

      Concerning prevention, statistically significant differences were found with respect to reduction of migraine attacks (0.67 less migraine attacks per month), migraine days (1.74 less migraine days per month), headache days (2.28 less headache days per month), and acute antimigraine drug intake (4.24 less instances of acute drug intake per month).
      Concerning acute treatment, statistically significant differences were found with respect to pain reduction on a visual analogue scale at 1/2/24 h post-acute treatment (1.68/1.02/1.08 improvement, respectively).
      No serious adverse events occurred in any of the studies.

      Conclusions

      While e-TNS has the potential to improve migraine symptoms, for its establishment in the standard practice, high quality comparative data, studies with larger sample sizes, and studies with standard and relevant primary outcome parameters are needed.

      Keywords

      1. Introduction

      1.1 Migraine

      Migraine is a primary headache disorder, characterized by headache attacks with moderate to severe headache, accompanying symptoms like nausea, vomiting or sensitivity to light and noise, and in some patients an aura with transient focal neurological symptoms [
      • Headache Classification Committee of the International Headache Society
      The international classification of headache disorders 3rd edition.
      ]. In episodic migraine (EM), the number of migraine days does not exceed 14 days per month, whereas chronic migraine (CM) is characterized by at least 15 headache days per month of which at least eight days must fulfil migraine criteria [
      • Headache Classification Committee of the International Headache Society
      The international classification of headache disorders 3rd edition.
      ].
      According to Global Burden of Disease 2015, migraine (both EM and CM) was ranked to be the third-highest cause of disability worldwide in both males and females under the age of 50 years [
      • Steiner T.J.
      • Stovner L.J.
      • Vos T.
      GBD 2015: migraine is the third cause of disability in under 50s.
      ]. EM affects more than 10% of the population [
      • Lipton R.B.
      • Silberstein S.D.
      Episodic and chronic migraine headache: breaking down barriers to optimal treatment and prevention.
      ]. The progression from EM to CM develops at a rate of about 2.5% per year, yet migraine patients show a great intra-individual variability and so also move back from CM to EM [
      • Serrano D.
      • Lipton R.B.
      • Scher A.I.
      • Reed M.L.
      • Stewart W.B.F.
      • Adams A.M.
      • et al.
      Fluctuations in episodic and chronic migraine status over the course of 1 year: implications for diagnosis, treatment and clinical trial design.
      ].
      Migraine prophylaxis is indicated in patients with three to four or more migraine days per month or with severe, long-lasting attacks that do not respond well to acute treatments [
      • Diener H.C.
      • Gaul C.
      • Kropp P.
      Leitlinie der Deutschen Gesellschaft für Neurologie (DGN) in Zusammenarbeit mit der Deutschen Migräne- und Kopfschmerzgesellschaft (DMKG).
      ]. Migraine prevention, however, remains to have several drawbacks. Until recently, drugs that were not especially developed for treating migraine were used for migraine prophylaxis (i.e. betablockers, calcium-channel blockers, antiepileptics, antidepressants). This is associated with the need of explaining the rationale of such treatment to the patients and with limiting adverse events. In 2018, the first migraine specific preventions were marketed, namely monoclonal antibodies against calcitonin-gene-related-peptide. In the search for a better migraine prevention or acute treatment, several devices for neurostimulation have been developed.

      1.2 External trigeminal nerve stimulator (e-TNS)

      The e-TNS (Cefaly®) is an intervention that stimulates the supraorbital branch of the trigeminal nerve with the aim of reducing the frequency and length of migraine attacks [
      • Riederer F.
      • Penning S.
      • Schoenen J.
      Transcutaneous supraorbital nerve stimulation (t-SNS) with the Cefaly((R)) device for migraine prevention: a review of the available data.
      ]. The supraorbital nerve ends at the vertex of the scalp, thus providing sensory innervation to the forehead, upper eyelid, and anterior scalp. The self-adhesive e-TNS electrode is placed on the forehead targeting the supraorbital nerve [
      • Anonymous
      A transcutaneous electrical nerve stimulation device (Cefaly) for migraine prevention.
      ]. The battery-operated electrical pulse generator connects magnetically to the electrode from where it conducts electrical micro-pulses [
      • Riederer F.
      • Penning S.
      • Schoenen J.
      Transcutaneous supraorbital nerve stimulation (t-SNS) with the Cefaly((R)) device for migraine prevention: a review of the available data.
      ]. The e-TNS is either used for the prevention of migraine attacks through daily 20-minute sessions, or it is used for acute treatment as a 60 or 120 min long intervention during the migraine attack. The pulse width of 250 μs and maximum intensity of 16 mA is the same for both modes, but the preventive mode has the pulse frequency of 60 Hz, while the acute treatment mode has 100 Hz [
      • Riederer F.
      • Penning S.
      • Schoenen J.
      Transcutaneous supraorbital nerve stimulation (t-SNS) with the Cefaly((R)) device for migraine prevention: a review of the available data.
      ] (see Table 1 for the characteristics of both treatment as well as sham devices).
      Table 1Technical characteristics of the e-TNS (Cefaly®) devices.
      Preventive e-TNS device [
      • Goadsby P.J.
      • Holland P.R.
      • Martins-Oliveira M.
      • Hoffmann J.
      • Schankin C.
      • Akerman S.
      Pathophysiology of migraine: a disorder of sensory processing.
      ,
      • Didier H.A.
      • Di Fiore P.
      • Marchetti C.
      • Tullo V.
      • Frediani F.
      • Arlotti M.
      • et al.
      Electromyography data in chronic migraine patients by using neurostimulation with the Cefaly device.
      ,
      • Magis D.
      • D’Ostilio K.
      • Thibaut A.
      • De Pasqua V.
      • Gerard P.
      • Hustinx R.
      • et al.
      Cerebral metabolism before and after external trigeminal nerve stimulation in episodic migraine.
      ,
      • Vecchio E.
      • Gentile E.
      • Franco G.
      • Ricci K.
      • de Tommaso M.
      Effects of external trigeminal nerve stimulation (eTNS) on laser evoked cortical potentials (LEP): a pilot study in migraine patients and controls.
      ]
      Preventive e-TNS sham device [
      • Magis D.
      • D’Ostilio K.
      • Thibaut A.
      • De Pasqua V.
      • Gerard P.
      • Hustinx R.
      • et al.
      Cerebral metabolism before and after external trigeminal nerve stimulation in episodic migraine.
      ]
      Acute treatment e-TNS device [
      • Institute of Health Economics (IHE)
      Quality Appraisal of Case Series Studies Checklist.
      ,
      • Higgins J.
      • Altman D.G.
      • Gøtzsche P.C.
      • Jüni P.
      • Moher D.
      • Oxman A.D.
      • et al.
      The cochrane collaboration’s tool for assessing risk of bias in randomised trials.
      ,
      • Guyatt G.
      • Oxman A.D.
      • Akl E.A.
      • Kunz R.
      • Vist G.
      • JEA B.
      GRADE guidelines: 1. Introduction-GRADE evidence profiles and summary of findings tables.
      ]
      Acute treatment e-TNS sham device [
      • Higgins J.
      • Altman D.G.
      • Gøtzsche P.C.
      • Jüni P.
      • Moher D.
      • Oxman A.D.
      • et al.
      The cochrane collaboration’s tool for assessing risk of bias in randomised trials.
      ]
      Pulse frequency60 Hz1 Hz100 Hz3 Hz
      Pulse width250 μs30 μs250 μs250 μs
      Maximum intensity16 mA1 mA16 mANA
      Length of treatment20 min/day20 min/day60–120 min60 min
      There is a lack of clarity about the pathophysiology of migraine and hence also about the exact mechanism of action of e-TNS. Migraine is thought to be a disease of the brain involving the nucleus caudalis of the trigeminal nerve, sensory inputs from the meninges, head, and neck areas innervated by the trigeminal nerve, and pain relevant structures like thalamus, hypothalamus, cingulum, periaqueductal grey, and pons [
      • Goadsby P.J.
      • Holland P.R.
      • Martins-Oliveira M.
      • Hoffmann J.
      • Schankin C.
      • Akerman S.
      Pathophysiology of migraine: a disorder of sensory processing.
      ]. With respect to e-TNS, it is assumed by the manufacturer that migraine attacks are triggered by the physiological communication between first cervical spinal nerve roots and the spinal trigeminal tract [
      • Didier H.A.
      • Di Fiore P.
      • Marchetti C.
      • Tullo V.
      • Frediani F.
      • Arlotti M.
      • et al.
      Electromyography data in chronic migraine patients by using neurostimulation with the Cefaly device.
      ]. The application of e-TNS to the supraorbital nerve is then supposed to use this nerve pathway to spread the impulse from the frontalis muscle to peripheral muscles, which may make it recorded in other muscles far from the application area [
      • Didier H.A.
      • Di Fiore P.
      • Marchetti C.
      • Tullo V.
      • Frediani F.
      • Arlotti M.
      • et al.
      Electromyography data in chronic migraine patients by using neurostimulation with the Cefaly device.
      ]. In this way, the e-TNS could act therapeutically on the inhibitory circuit in the spinal cord causing a neuromuscular facilitation and a reduction in contractions of frontalis muscles [
      • Didier H.A.
      • Di Fiore P.
      • Marchetti C.
      • Tullo V.
      • Frediani F.
      • Arlotti M.
      • et al.
      Electromyography data in chronic migraine patients by using neurostimulation with the Cefaly device.
      ]. Alternatively, another assumption is that e-TNS exerts its beneficial effects via slow neuromodulation of central pain-controlling areas [
      • Magis D.
      • D’Ostilio K.
      • Thibaut A.
      • De Pasqua V.
      • Gerard P.
      • Hustinx R.
      • et al.
      Cerebral metabolism before and after external trigeminal nerve stimulation in episodic migraine.
      ]. It is assumed that it reduces the hypometabolism of the orbitofrontal cortex as well as the rostral parts of the anterior cingulate cortex as the metabolic activity in these areas is decreased in migraine [
      • Magis D.
      • D’Ostilio K.
      • Thibaut A.
      • De Pasqua V.
      • Gerard P.
      • Hustinx R.
      • et al.
      Cerebral metabolism before and after external trigeminal nerve stimulation in episodic migraine.
      ]. Furthermore, e-TNS may interfere with the threshold and the extent of trigeminal system activation, thus resolving or preventing migraine attacks [
      • Vecchio E.
      • Gentile E.
      • Franco G.
      • Ricci K.
      • de Tommaso M.
      Effects of external trigeminal nerve stimulation (eTNS) on laser evoked cortical potentials (LEP): a pilot study in migraine patients and controls.
      ].
      The claimed benefit of e-TNS is the reduction of the frequency and length of migraine attacks. E-TNS claims to be less invasive, have a better effectiveness-safety ratio, less side effects, no serious side effects, and fewer contraindications than its comparators. Furthermore, it claims to reduce the acute anti-migraine drug intake, thus preventing medication overuse headache.

      1.2.1 Aims

      The aim of this systematic review was to assess whether e-TNS, as a preventive or acute treatment, is more or equally effective concerning satisfaction, change in pain score, reduction of monthly migraine days and antimigraine drug intake, and safer regarding side effects in patients with EM and CM. This was assessed against the comparators of sham or standard therapy that include triptans, NSAIDs/paracetamol, or their combination for acute treatment, or topiramate and propranolol for prevention. For the full list of endpoints assessed, see Tables 2 and 3.
      Table 2Results from randomised controlled trials and prospective observational studies for prevention of EM and CM with e-TNS.
      Schoenen et al. [
      • Magis D.
      • D’Ostilio K.
      • Thibaut A.
      • De Pasqua V.
      • Gerard P.
      • Hustinx R.
      • et al.
      Cerebral metabolism before and after external trigeminal nerve stimulation in episodic migraine.
      ]
      Russo et al. [
      • Didier H.A.
      • Di Fiore P.
      • Marchetti C.
      • Tullo V.
      • Frediani F.
      • Arlotti M.
      • et al.
      Electromyography data in chronic migraine patients by using neurostimulation with the Cefaly device.
      ]
      DiFiore et al. [
      • Goadsby P.J.
      • Holland P.R.
      • Martins-Oliveira M.
      • Hoffmann J.
      • Schankin C.
      • Akerman S.
      Pathophysiology of migraine: a disorder of sensory processing.
      ]
      Vikelis et al. [
      • Vecchio E.
      • Gentile E.
      • Franco G.
      • Ricci K.
      • de Tommaso M.
      Effects of external trigeminal nerve stimulation (eTNS) on laser evoked cortical potentials (LEP): a pilot study in migraine patients and controls.
      ]
      CountryBelgiumItalyItalyGreece
      SponsorWalloon Region
      The manufacturer STX Med provided the Cefaly® devices.
      Unclear
      STX Med provided the Cefaly® devices, but it is stated that the study was not industry sponsored.
      Unclear
      Mean duration of the chronic phase was 10.7 (8.7).
      STX Med-Cefaly® Technology and Brain Therapeutics Greece
      Study designMulti-centre, prospective, double-blinded, randomised, sham-controlled trialProspective single-arm, interventional trialProspective single-arm, interventional, open label preliminary trialMulti-centre, prospective, single-arm, interventional trial (NCT03125525)
      Conducted in09/2009–09/201101/2013–10/201404/2014–12/2014NA
      IndicationPreventive treatment in pts. with episodic migraine with and without auraPreventive treatment in pts. with episodic migraine without auraPreventive treatment in pts. with chronic migraine with or without medication overusePreventive treatment in pts. with episodic or chronic migraine refractory or intolerant to Topiramate phrophylaxis
      Intervention (I)e-TNS (pulse frequency 60 Hz, pulse width 250 μs, max intensity 16 mA, 20 min/day)e-TNS (pulse frequency 60 Hz, pulse width 250 μs, max intensity 16 mA, 20 min/day)e-TNS (pulse frequency 60 Hz, pulse width 250 μs, max intensity 16 mA, 20 min/day)e-TNS (20 min/day)
      Comparator (C)Sham Cefaly® device (pulse frequency 1 Hz, pulse width 30 μs, max intensity 1 mA, 20 min/day)NoneNoneNone
      Number of pts. (I vs. C)34 vs. 3324
      Baseline data on 20 pts.
      2337
      2 pts dropped out before using e-TNS device. Baseline data on 35 pts.
      Inclusion criteriaPts 18–65 yrs., migraine with or without aura meeting ICHD-II code 1.2.1 or 1.1 criteria, episodic migraine of ≥ 2 attacks per moMigraine without aura meeting

      ICHD-III criteria,

      episodic migraine of ≤ 5 attacks per mo
      Pts 18+ yrs., chronic migraine with or without medication overuse headache+ meeting ICHD-III criteria, chronic migraine for 1+ yrs., not part of withdrawal program to stop medication overuse, normal neurological exam, normal neuroimaging findingsPts refractory/intolerant to topiramate, episodic/chronic (≥15 days headache/mo) migraine according to ICHD-III criteria, stop topiramate 3 mos prior
      Exclusion criteriaPreventive antimigraine treatment in prior 3 mos, failure of ≥3 antimigraine medications, medication overuse headache, frequent/chronic tension type headache, severe neurologic or psychiatric disordersOther type of headache, somatic and psychosomatic conditions, daily medication intake, migraine drug-naïve pts., no structural brain abnormality confirmed by MRIPregnancy, major neurological, systemic or psychiatric illnessNA
      Primary outcome measureChange in monthly migraine days and 50% responder rateChange in monthly migraine days and migraine attacks, % of pts. having ≥50% reduction of monthly migraine attacks and migraine days50% or more reduction in both headache days per month, and in consumption of acute headache relief medications per monthChange in total headache days and days with acute medication use
      Secondary outcome measureChange in monthly frequency of any headache, change in mean headache severity per migraine day (4 points scale), change in monthly acute antimigraine drug use, percentage of patients stating satisfactionAverage of pain intensity during migraine attacks (VAS scale), intake of rescue medication during migraine attacks, satisfaction, compliance, HIT-6 scoreNANA
      Baseline patient characteristics (I vs. C)
      Mean age, yrs. (SD)34.59 (11.01) vs. 39.06 (9.87)32.9 (2.3)43.7 (13.6)45 (median), 22–62 (range)
      Sex, female:male, n31:3 vs. 30:315:518:531:4
      Migraine with aura, n (%)10 (29.4) vs. 10 (30.3)
      While all pts had migraine without aura, 20 pts had occasional visual aura preceding the attack.
      0 (0)NANA
      Migraine without aura, n (%)24 (70.6) vs. 23 (69.7)24 (100)NANA
      Migraine duration, mean n of yrs. (SD)14.71 (9.39) vs. 18.17 (11.68)

      NA
      8.3 (1.7)26.4 (12.8)
      Mean duration of the chronic phase was 10.7 (8.7).
      NA
      Migraine attack duration in hrs, median (IQR)NANANANA
      Follow-up time, days906012090
      Loss to follow-up, n (%)8 (11.9)4 (16.7)
      Pts excluded from the analysis for non-compliance (compliance defined in the study as ≥2/3 of total 60 treatment days).
      4 (17.4)
      1 pt due to keratoconjunctivitis, 3 pts due to inability to tolerate e-TNS. Efficacy reported for 19 pts who fulfilled 4-months treatment schedule.
      8 (21.6)
      Efficacy (I vs. C)
      Reduction in monthly migraine attacks, mean n (± SD)
      • Baseline
      4.37 (1.87) vs 4.04 (1.52)4.6 (baseline)NANA
      • 90 days post interevention
      3.55 (2.94) vs 3.89 (1.89)2.1
      Translated from figure, as values were not explicitly reported on in the text. Unclear if mean or median.
      (60 days)
      NANA
      Difference, mean (p-value)NA (p = .044)2.5 (p < .001)NANA
      Reduction in monthly migraine days, mean n (±SD)
      • Baseline
      6.94 (3.04) vs. 6.54 (2.61)6.6 (baseline)20.7 (baseline)NA
      • 90 days post interevention
      4.88 (3.46) vs 6.22 (2.99)3.1 l(60 days)14.3
      Analysis based on 19 pts.
      (120 days)
      NA
      Difference, mean n (p-value)NA (p = .054)3.5 (p < .001)31.0% (NA)NA
      Reduction in monthly headache days, mean (SD)
      • Baseline
      7.78 (4.00) vs. 6.72 (2.63)NANA8.9 (4.7)
      • 90 days post interevention
      5.27 (3.55) vs 6.49 (3.2)NANA6.3 (3.5)
      Difference, mean n (p-value)NA (p = .041)NANA2 (p < .001)
      HIT 6 score
      • Baseline
      NA62.3 ± 1.4NANA
      • 90 days post interevention
      NA53.1 ± 1.4NANA
      Difference, mean n (p-value)NA9.2 (p < .001)NANA
      Reduction in monthly acute antimigraine drug intake, (n of times per month), mean (SD)
      • Baseline
      11.45 (8.35) vs 9.24 (4.75)5.6 (0.4) (baseline)20.2 (baseline)8.2 (4.6)
      • 90 days post interevention
      7.25 (7.31) vs 9.28 (5.69)2.2 (0.3)
      Unclear if it refers to monthly intake because the text states “total intake”.
      (60 days)
      10.2
      4 pts (16.7%) were considered non-compliant as they did not finish with ≥ 800 min of e-TNS treatment in the 60 days trial period.
      (120 days)
      4.4 (3.3)
      Difference, mean n [%] (p-value)NA [NA](p = .0072)3.4 [40.7] (p < .001)NA [49](NA)NA [46.3] (NA)
      Responder rate, reduction in migraine days, baseline vs. follow-up, %12 (40) vs. 4 (13.8)75 (NA)NANA
      QoLNANANANA
      Satisfaction, n (%)29.4 (10)/41.2 (14)/21.2 (7)/8.8 (3) vs. 18.2 (6)/21.2 (7)/51.5 (17)/9.1 (3)20 (83.3)
      Not measured, but defined by willingness to continue using e-TNS. Patients excluded due to non-compliance are part of this analysis.
      NA23 (65.7)
      Very/moderately/not at all satisfied/NA
      Compliance, n (%)55.54 (61.7) vs. 49 (54.4)20 (83.3)
      4 pts (16.7%) were considered non-compliant as they did not finish with ≥ 800 min of e-TNS treatment in the 60 days trial period.
      19 (82.6)27 (81.8)
      Safety
      SADEs, n0000
      Pain/intolerance to paresthesia (burning sensation), n (%)00012 (34.3)
      Arousal changes (insomnia, sleepiness/fatigue), n (%)00 (0)0 (0)NA
      Headaches after stimulation, n (%)00 (0)1 (4.3)NA
      Nausea after stimulation, n (%)00 (0)0 (0)NA
      Skin allergy/irritation, n (%)00 (0)0 (0)NA
      Neck tension, n (%)NANA2 (8.7)NA
      C – comparator, e-TNS – external trigeminal nerve stimulation, HIT 6 – Headache Impact Test score, ICD – implantable cardioverter defibrillator, ICHD – International Classification of Headache Disorders, IHS – International Headache Society, I – intervention, IQR – inter-quartile range, hrs – hours, mITT – modified intention-to-treat, mos – months, min – minutes, MRI – magnetic resonance imaging, NA – data not available, pts – patients, QoL – quality of life, SADEs – serious adverse device effects, SD – standard deviation, yrs – years, VAS – visual analogue scale, WCD – wearable cardioverter defibrillator.
      a The manufacturer STX Med provided the Cefaly® devices.
      b STX Med provided the Cefaly® devices, but it is stated that the study was not industry sponsored.
      c Baseline data on 20 pts.
      d 2 pts dropped out before using e-TNS device. Baseline data on 35 pts.
      e While all pts had migraine without aura, 20 pts had occasional visual aura preceding the attack.
      f Mean duration of the chronic phase was 10.7 (8.7).
      g Pts excluded from the analysis for non-compliance (compliance defined in the study as ≥2/3 of total 60 treatment days).
      h 1 pt due to keratoconjunctivitis, 3 pts due to inability to tolerate e-TNS. Efficacy reported for 19 pts who fulfilled 4-months treatment schedule.
      i Translated from figure, as values were not explicitly reported on in the text. Unclear if mean or median.
      j Analysis based on 19 pts.
      k Unclear if it refers to monthly intake because the text states “total intake”.
      l Not measured, but defined by willingness to continue using e-TNS. Patients excluded due to non-compliance are part of this analysis.
      m 4 pts (16.7%) were considered non-compliant as they did not finish with ≥ 800 min of e-TNS treatment in the 60 days trial period.
      Table 3Results from randomised controlled trials and prospective observational studies for acute treatment of EM and CM.
      Chou et al. [
      • Higgins J.
      • Altman D.G.
      • Gøtzsche P.C.
      • Jüni P.
      • Moher D.
      • Oxman A.D.
      • et al.
      The cochrane collaboration’s tool for assessing risk of bias in randomised trials.
      ]
      Chou et al. [
      • Institute of Health Economics (IHE)
      Quality Appraisal of Case Series Studies Checklist.
      ]
      Mann et al. [
      • Guyatt G.
      • Oxman A.D.
      • Akl E.A.
      • Kunz R.
      • Vist G.
      • JEA B.
      GRADE guidelines: 1. Introduction-GRADE evidence profiles and summary of findings tables.
      ]
      CountryUnited StatesUnited StatesUnited States
      SponsorSTX Med – Cefaly® TechnologySTX Med -Cefaly® TechnologySTX Med -Cefaly® Technology
      Study designMulti-centre, prospective, double-blinded, randomised, sham-controlled trial (NCT02590939)Prospective single-arm, interventional, open label trial (NCT02411513)Single center, prospective, open-label, phase 1 trial (NCT03217968).
      Conducted in02/2016–03/201704/2015–10/201508/2017–01/2018
      IndicationAcute treatment in pts. with acute migraine (episodic or chronic) with or without auraAcute treatment in pts. with acute migraine attack (episodic or chronic) with or without auraAcute treatment in pts. with single moderate or severe migraine attack (Grade 2 or 3) at home, pts. with episodic migraine
      Intervention (I)e-TNS (pulse frequency 100 Hz, pulse width 250 μs, max intensity 16 mA, 60 min)e-TNS (pulse frequency 100 Hz, pulse width 250 μs,

      max intensity 16 mA, 60 min)
      e-TNS (pulse frequency 100 Hz, pulse width 250 μs, max intensity 16 mA,

      120 min)
      Comparator (C)Sham Cefaly® device (pulse frequency 3 Hz, pulse width 250 μs, max intensity NA, 60 min)NoneNone
      Number of pts. (I vs. C)52 vs. 5435
      Baseline data on 30 pts.
      60
      mITT analysis with 48 pts.
      Inclusion criteriaPts 18–65 yrs., episodic/chronic migraine with or without aura meeting ICHD-III criteria, 3 h long attack, pain intensity stabilized for 1 h, frontal-retro-peri-orbital headachePts 18–65 yrs., episodic/chronic migraine with or without aura meeting ICHD-III criteria, 3+ hrs long attack, pain intensity stabilized for 1 h, frontal-retro-peri-orbital headachePts 18–65 yrs., ≥ 1-year history of episodic migraine with or without aura meeting ICHD-III criteria, migraine onset before 50 yrs. of age, 2–8 moderate-severe migraine attacks/mo in each of the 2 mos prior to screening, pts.' consent
      Exclusion criteriaPregnancy, Botox in past 4 mos, supra-orbital nerve blocks in past 4 mos, other primary/secondary headache (except medication overuse), temporal/occipital headache pts., opioid medication use, migraine abortive medications in past 3 h, allodynia (oversensitivity to pain), metal or electric implants in head, cardiac pacemaker, ICD, WCD, previous Cefaly® pts., pts. with complicated migraine (hemiphlegic, basilar-type, ophthalmoplegic, migranous infarction)Pregnancy, Botox in the head in past 4 mos, supra-orbital nerve blocks in past 4 mos, other primary/secondary headache (except medication overuse), temporal/occipital headache pts., opioid medication use, migraine abortive medications in past 3 h, allodynia (oversensitivity to pain), metal or electric implants in head, cardiac pacemaker, ICD, WCD, pts. with complicated migraine (hemiphlegic, basilar-type, ophthalmoplegic, migranous infarction)Pts' difficulty distinguishing migraine from tension-type headache, >15 headaches per month (chronic migraine pts), migraine aura without headache, hemiplegic migraine and brainstem aura migraine, pts. with supraorbital nerve blocks or Botox in the head in the prior 4 mos, migraine prophylaxis modification in prior 3 mos, other primary/secondary headache disorders (medication overuse), pts. with opioid, alcohol or illicit drugs abuse, metallic or electric device in head, cardiac pacemaker, ICD, WCD, prior experience with Cefaly®, participation in other study in past 30 days, pts. unable to self-service or bear the e-TNS stimulation
      Primary outcome measureMean change in pain score at 1 h after interventionMean change in pain intensity after one-hour acute treatment compared to baselineFreedom from pain and from most bothersome migraine-associated symptoms (photophobia, phonophobia, nausea, vomiting) at 2 h post-acute treatment with e-TNS
      Secondary outcome measureMean change in pain score at 2 h/24 h compared to baseline (VAS scale), pts.' used rescue medication after 2/24 hChange in pain intensity after two-hour acute treatment compared to baseline (VAS scale), percentage of pts. not requiring rescue medication at 2 h/24 hReduction of moderate to severe migraine headache and percentage of pts. with absence of photophobia, phonophobia, nausea, vomiting at 2 h from baseline. Measured at baseline, 2 h, 24 h, on the scale: 0 = no pain; 1 = mild pain; 2 = moderate pain; 3 = severe pain
      Baseline patient characteristics (I vs. C)
      Mean age, yrs. (SD)39.71 (13.62) vs. 40.09 (12.65)39.4 (12.5)46.85 (10.2)
      Sex, female:male, n43:9 vs. 49:524:643:5
      Migraine with aura, n (%)12 (23.1) vs. 5 (9.3)NA15 (25)
      Migraine without aura, n (%)40 (76.9) vs. 49 (90.7)NA33 (55)
      Migraine duration, yrsNANA>1
      Not specified.
      Migraine attack duration in hrs, median (IQR)7 (4–48) vs. 6 (4.63–20.75)NANA
      Pts on medication, mean (SD)

      Other baseline symptoms: (n (%)
      17 vs. 14NANA
      • Nausea
      NANA11 (18.3)
      • Vomiting
      NANA1 (1.66)
      • Sensitivity to light
      NANA27 (45)
      • Sensitivity to sound
      NANA9 (15)
      Follow-up time, days111
      Loss to follow-up, %5 (9.6) vs. 2 (3.7)
      In the IG and the CG, 3 and 1 subjects withdrew and 2 and 1 subjects failed nociceptive test, respectively.
      5
      1 pt due to opioid use in past 3 mos, 4 pts due to inability to tolerate e-TNS.
      12
      1 pt failed the training test, 4 pts withdrew from the study, 1 pt was lost to follow-up and 6 pts did follow the study protocol.
      Efficacy
      QoLNANANA
      Satisfaction, n (%)NANANA
      Headache pain free pts., n (%)
      • 2 h post treatment
      NANA17 (35.4)
      • 24 h post treatment
      NANA12 (25)
      Freedom from nausea, vomiting, sensitivity to light and sound at 2 h, n (%)NANA22 (36.7)
      From the results document it is unclear if 22 or 29 patients were free from the most bothersome symptoms.
      Reduction in pain score units on VAS scale, mean n (SD)
      • Baseline
      .92 (1.68) vs 6.17 (1.81)Change in pain score:NA
      • 1 h post treatment
      2.46 (2.23) vs 4.39 (2.44)−3.22 (2.4)NA
      • P value
      p = .0001p < .001)NA
      Reduction in pain score units on VAS scale, mean n (SD)
      • Baseline
      5.92 (1.68) vs 6.17 (1.81)Change in pain score:
      • 2 h post treatment
      3.06 (2.30) vs 4.31 (2.51)−2.98 (2.31)
      • P value
      p = .028p < .00134 (NA)
      Number (%) of pts with pain relief at 2 hrs, not measured on VAS scale.
      Reduction in pain score units on VAS scale, mean n (SD)
      • Baseline
      5.92 (1.68) vs 6.17 (1.81)NANA
      • 24 h post treatment
      2.46 (2.27) vs 3.79 (2.74)NANA
      • P value
      p = 0.062NANA
      Pts on rescue medication, n (%)
      • 2 h post treatment
      3 (5.8) vs. 2 (3.7)0 (0)NA/
      • 24 h post treatment
      18 (34.6) v s21 (38.9)
      The data on the use of rescue medication was not available for 9.6%/13.5% vs. 3.7%/5.6% of pts at 2/24 hrs.
      17 (48.6)
      34.6% out of 26 pts as 4 pts were lost to follow-up (not reachable at 24hrs).
      24 (40)
      Out of 48 pts as 12 were lost to follow-up.
      • P value
      p = .66NANA
      Compliance, n (%)NANA49 (81.7)
      Defined as withdrawal, loss to follow-up, or violation of protocol.
      Safety
      SADEs, n (%)0 vs. 00 (0)0 (0)
      Pain/intolerance to paresthesia (burning sensation), n (%)3 (5.8) vs. 1 (1.9)
      Before completion of the 1 h of e-TNS session.
      0 (0)7 (11.9)
      Arousal changes (insomnia, sleepiness/fatigue, drowsiness), n (%)NA0 (0)1 (1.7)
      ADEs reported out of 59 pts.
      Headaches after stimulation, n (%)NA0 (0)NA
      Nausea after stimulation, n (%)1 (1.9) vs. 00 (0)2 (3.5)
      ADEs reported out of 59 pts.
      Dizziness, n (%)NANA1 (1.7)
      Vomiting, n (%)NANA1 (1.7)
      Pain in the jaw, n (%)NANA1 (1.7)
      Discomfort in teeth, n (%)NANA1 (1.7)
      Pain in eyes, n (%)NANA1 (1.7)
      Cold feet, n (%)NANA1 (1.7)
      Skin allergy/irritation, n (%)NANA1 (1.7)
      Neck tension, n (%)NA0 (0)11 (18.3)
      NA0 (0)NA
      C – comparator, e-TNS – external trigeminal nerve stimulation, ICD – implantable cardioverter defibrillator, ICHD – International Classification of Headache Disorders, IHS – International Headache Society, I – intervention, IQR – inter-quartile range, hrs – hours, mITT – modified intention-to-treat, mos – months, min – minutes, MRI – magnetic resonance imaging, NA – data not available, pts – patients, QoL – quality of life, SADEs – serious adverse device effects, SD – standard deviation, yrs – years, VAS – visual analogue scale, WCD – wearable cardioverter defibrillator.
      a Baseline data on 30 pts.
      b mITT analysis with 48 pts.
      c Not specified.
      d In the IG and the CG, 3 and 1 subjects withdrew and 2 and 1 subjects failed nociceptive test, respectively.
      e 1 pt due to opioid use in past 3 mos, 4 pts due to inability to tolerate e-TNS.
      f 1 pt failed the training test, 4 pts withdrew from the study, 1 pt was lost to follow-up and 6 pts did follow the study protocol.
      g From the results document it is unclear if 22 or 29 patients were free from the most bothersome symptoms.
      h Number (%) of pts with pain relief at 2 hrs, not measured on VAS scale.
      i The data on the use of rescue medication was not available for 9.6%/13.5% vs. 3.7%/5.6% of pts at 2/24 hrs.
      j 34.6% out of 26 pts as 4 pts were lost to follow-up (not reachable at 24hrs).
      k Out of 48 pts as 12 were lost to follow-up.
      l Defined as withdrawal, loss to follow-up, or violation of protocol.
      m Before completion of the 1 h of e-TNS session.
      n ADEs reported out of 59 pts.

      1.3 Materials and methods

      1.3.1 Search strategy

      The systematic search was conducted between 3-7th of May 2018 in four databases (Medline via Ovid, Embase, the Cochrane library, and CRD) and it was not limited to a year of publication. The search was limited to prospective studies and articles published in English or German. The specific search strategies (including syntaxes, parameters, and results) are available in the supplementary material.

      1.3.2 Study selection

      In the review process, three researchers independently screened abstracts and full-text articles for eligibility based on the a priori defined inclusion criteria according to the PICO (population, intervention, comparator, and outcome) question (see PICO: Inclusion criteria in the Supplementary material). Discrepancies were resolved by consensus.
      The PICO inclusion and exclusion criteria were the following. In terms of population, all studies with adult patients with episodic and/or chronic migraine who could benefit from e-TNS preventive or acute therapy were included. In terms of intervention, all patients that used external trigeminal nerve stimulator or supraorbital transcutaneous nerve stimulator (original name changed after obtaining a GMDN code), produced by Cefaly® were included. In terms of comparators, the following were included for the acute treatment use of e-TNS: sham device, NSAIDs/paracetamol, triptans, and triptans in combination with NSAIDs/paracetamol. For the preventive use of e-TNS, the following comparators were included: sham device, topiramate, propranolol, and all other guideline-directed agents such as metoprolol, flunarizin, valproate, and others. For the exhaustive list of outcomes, see the PICO question in the Supplementary material. In terms of study design, prospective non-randomised controlled trials as well as randomised controlled trials were included in the assessment of clinical effectiveness. For the assessment of safety, prospective case series were also included.
      Following the above inclusion criteria, 396 articles were excluded at the stage of abstract screening and so 44 full-text articles were left to be assessed for eligibility. Of those, seven studies were included for qualitative analysis (see Fig. 1: PRISMA Flow Diagram).
      Fig. 1
      Fig. 1Flow chart of study selection for e-TNS (PRISMA Flow Diagram).

      1.3.3 Data extraction

      One researcher extracted data and the other two checked the data extraction. Data were recorded in a structured manner as follows: sample characteristics, study characteristics, primary outcome measures, and secondary outcome measures.

      1.3.4 Quality assessment

      Authors assessed the quality and risk of bias of the included studies using the Institute of Health Economics' quality appraisal checklist for case series [
      • Institute of Health Economics (IHE)
      Quality Appraisal of Case Series Studies Checklist.
      ] and the Cochrane Collaboration's tool for randomised trials [
      • Higgins J.
      • Altman D.G.
      • Gøtzsche P.C.
      • Jüni P.
      • Moher D.
      • Oxman A.D.
      • et al.
      The cochrane collaboration’s tool for assessing risk of bias in randomised trials.
      ]. All included studies were scored independently by two authors and overseen by a third. Based on the data extraction tables (Tables 2 and 3), the data on each selected outcome category were synthesized across studies according to GRADE (Grading of Recommendations Assessment, Development, and Evaluation) methodology [
      • Guyatt G.
      • Oxman A.D.
      • Akl E.A.
      • Kunz R.
      • Vist G.
      • JEA B.
      GRADE guidelines: 1. Introduction-GRADE evidence profiles and summary of findings tables.
      ].

      2. Results

      2.1 Available evidence

      In the systematic literature search for e-TNS, we found 433 citations through electronic databases and additional seven citations through hand-search. After deduplication, 440 citations were included. The only manufacturer of e-TNS (Cefaly®) submitted 19 publications of which 4 were unpublished at the time of writing of the report and hence were confidential and not included. No new citations were identified. 44 full text articles were assessed for eligibility of which seven studies were included in the qualitative synthesis (see Fig. 1). Prospective controlled trials were included for the assessment of clinical effectiveness and for the assessment of safety, prospective case series were also included.

      2.2 Characteristics of included studies

      For the assessment of clinical effectiveness, two studies met the inclusion criteria. One randomised controlled trial (RCT) for the preventive use of e-TNS [
      • Schoenen J.
      • Vandersmissen B.
      • Jeangette S.
      • Herroelen L.
      • Vandenheede M.
      • Gerard P.
      • et al.
      Migraine prevention with a supraorbital transcutaneous stimulator: a randomized controlled trial.
      ] and one RCT for the acute treatment use of e-TNS [
      • Chou E.D.
      Acute Treatment of Migraine With e-TNS (ACME).
      ]. Both compared the e-TNS (Cefaly®) to a sham (Cefaly®) device. The latter was not in the form of a peer-reviewed publication but in the form of a study protocol and study results published at clinicaltrials.gov [
      • Chou E.D.
      Acute Treatment of Migraine With e-TNS (ACME).
      ]. The peer-reviewed publication was provided by the manufacturer after completion of the review [
      • Chou D.E.
      • Shnayderman Y.M.
      • Winegarner D.
      • Rowe V.
      • Kuruvilla D.
      • Schoenen J.
      Acute migraine therapy with external trigeminal neurostimulation (ACME): a randomized controlled trial.
      ]. In the preventive study, the length of follow-up was 90 days [
      • Schoenen J.
      • Vandersmissen B.
      • Jeangette S.
      • Herroelen L.
      • Vandenheede M.
      • Gerard P.
      • et al.
      Migraine prevention with a supraorbital transcutaneous stimulator: a randomized controlled trial.
      ], while in the acute treatment study, the follow-up was 24 h [
      • Chou E.D.
      Acute Treatment of Migraine With e-TNS (ACME).
      ].
      For the assessment of safety, seven studies met the inclusion criteria: two RCTs described above [
      • Schoenen J.
      • Vandersmissen B.
      • Jeangette S.
      • Herroelen L.
      • Vandenheede M.
      • Gerard P.
      • et al.
      Migraine prevention with a supraorbital transcutaneous stimulator: a randomized controlled trial.
      ,
      • Chou E.D.
      Acute Treatment of Migraine With e-TNS (ACME).
      ] and five prospective case series. One of the five case series publications was not in the form of a peer-reviewed publication but in the form of a study protocol and study results published at clinicaltrials.gov [
      • Mann J.
      Abortive Treatment of Migraine With the Cefaly® Abortive Program Device.
      ]. In the preventive case series, the length of follow-up ranged from 60 to 120 days [
      • Di Fiore P.
      • Bussone G.
      • Galli A.
      • Didier H.
      • Peccarisi C.
      • D’Amico D.
      • et al.
      Transcutaneous supraorbital neurostimulation for the prevention of chronic migraine: a prospective, open-label preliminary trial.
      ,
      • Russo A.
      • Tessitore A.
      • Conte F.
      • Marcuccio L.
      • Giordano A.
      • Tedeschi G.
      Transcutaneous supraorbital neurostimulation in “de novo” patients with migraine without aura: the first Italian experience.
      ,
      • Vikelis M.
      • Dermitzakis E.V.
      • Spingos K.C.
      • Vasiliadis G.G.
      • Vlachos G.S.
      • Kararizou E.
      Clinical experience with transcutaneous supraorbital nerve stimulation in patients with refractory migraine or with migraine and intolerance to topiramate: a prospective exploratory clinical study.
      ], while in the acute treatment case series studies, the follow-up was 24 h [
      • Mann J.
      Abortive Treatment of Migraine With the Cefaly® Abortive Program Device.
      ,
      • Chou D.E.
      • Gross G.J.
      • Casadei C.H.
      • Yugrakh M.S.
      External trigeminal nerve stimulation for the acute treatment of migraine: open-label trial on safety and efficacy.
      ]. The baseline characteristics of each study are presented in Tables 2 and 3.

      2.3 Patient characteristics

      For the assessment of clinical effectiveness, the preventive study included 67 patients, of which 34 were in the intervention group (IG) and 33 in the control group (CG) [
      • Schoenen J.
      • Vandersmissen B.
      • Jeangette S.
      • Herroelen L.
      • Vandenheede M.
      • Gerard P.
      • et al.
      Migraine prevention with a supraorbital transcutaneous stimulator: a randomized controlled trial.
      ]. Of the 67 patients, 91% were women and eight (11.9%) patients were lost to follow-up. The acute treatment study included 106 patients, of which 52 were in the IG and 54 in the CG [
      • Chou E.D.
      Acute Treatment of Migraine With e-TNS (ACME).
      ]. Of the 106 patients, 86.8% were women and five (9.6%) of IG patients and two (3.7%) of the CG patients were lost to follow-up. The mean age of patients was similar in both studies and ranged between 34.59 and 40.09 years.
      For the assessment of safety (apart from the RCTs described above), the three preventive case series studies included 24, 23, and 37 patients, respectively (total of 84 patients) [
      • Di Fiore P.
      • Bussone G.
      • Galli A.
      • Didier H.
      • Peccarisi C.
      • D’Amico D.
      • et al.
      Transcutaneous supraorbital neurostimulation for the prevention of chronic migraine: a prospective, open-label preliminary trial.
      ,
      • Russo A.
      • Tessitore A.
      • Conte F.
      • Marcuccio L.
      • Giordano A.
      • Tedeschi G.
      Transcutaneous supraorbital neurostimulation in “de novo” patients with migraine without aura: the first Italian experience.
      ,
      • Vikelis M.
      • Dermitzakis E.V.
      • Spingos K.C.
      • Vasiliadis G.G.
      • Vlachos G.S.
      • Kararizou E.
      Clinical experience with transcutaneous supraorbital nerve stimulation in patients with refractory migraine or with migraine and intolerance to topiramate: a prospective exploratory clinical study.
      ]. 70–88.6% were women and a total of 16 (19.05%) patients were lost to follow-up. The mean age varied between 32.9 and 45 years. The two acute treatment case series studies included 35 and 60 patients, respectively (total of 95 patients) [
      • Mann J.
      Abortive Treatment of Migraine With the Cefaly® Abortive Program Device.
      ,
      • Chou D.E.
      • Gross G.J.
      • Casadei C.H.
      • Yugrakh M.S.
      External trigeminal nerve stimulation for the acute treatment of migraine: open-label trial on safety and efficacy.
      ]. 80–89.6% were women and a total of 17 (17.9%) patients were lost to follow-up. The mean age varied between 39.4 and 46.9 years. The baseline characteristics of each study are presented in Tables 2 and 3.

      2.4 Clinical effectiveness

      2.4.1 Prevention

      In terms of the preventive effect of e-TNS, the symptoms of monthly migraine attacks, monthly headache days, and monthly migraine days and responder rate were measured. In EM, patients in the IG had a mean reduction in monthly migraine attacks of 0.82, while patients in the CG of 0.15 days. The difference (net gain in the IG compared to baseline) was 0.67 attacks (p = .044) [
      • Schoenen J.
      • Vandersmissen B.
      • Jeangette S.
      • Herroelen L.
      • Vandenheede M.
      • Gerard P.
      • et al.
      Migraine prevention with a supraorbital transcutaneous stimulator: a randomized controlled trial.
      ]. With respect to headache days, patients in the IG had a mean reduction of 2.51 days, while patients in the CG of 0.23 days. The difference (net gain in the IG compared to baseline) was 2.28 day (p = .041) [
      • Schoenen J.
      • Vandersmissen B.
      • Jeangette S.
      • Herroelen L.
      • Vandenheede M.
      • Gerard P.
      • et al.
      Migraine prevention with a supraorbital transcutaneous stimulator: a randomized controlled trial.
      ]. With respect to migraine days, patients in the IG had a mean reduction of 2.06 days, while patients in the CG of 0.32 days. The difference (net gain in the IG compared to baseline) was 1.74 days (p = .054) [
      • Schoenen J.
      • Vandersmissen B.
      • Jeangette S.
      • Herroelen L.
      • Vandenheede M.
      • Gerard P.
      • et al.
      Migraine prevention with a supraorbital transcutaneous stimulator: a randomized controlled trial.
      ]. The 50% responder rate, the number of patients with a 50% or greater reduction in frequency of migraine days at the end of the study vs. the baseline, was 40% (12 patients) in the IG as opposed to 13.8% (4 patients) in the CG [
      • Schoenen J.
      • Vandersmissen B.
      • Jeangette S.
      • Herroelen L.
      • Vandenheede M.
      • Gerard P.
      • et al.
      Migraine prevention with a supraorbital transcutaneous stimulator: a randomized controlled trial.
      ].
      Furthermore, with respect to the mean reduction in monthly acute antimigraine drug intake, patients in the IG had a reduction of 4.2 instances (decrease from 11.45 ± 8.35 at baseline to 7.25 ± 7.3) at 90 days follow-up with p = .0057). Patients in the CG had an increase of 0.04 instances (increase from 9.24 ± 4.75 to 9.28 ± 5.69 with p = .822). The difference (net gain in the IG) was 4.24 instances (p = .0072) [
      • Schoenen J.
      • Vandersmissen B.
      • Jeangette S.
      • Herroelen L.
      • Vandenheede M.
      • Gerard P.
      • et al.
      Migraine prevention with a supraorbital transcutaneous stimulator: a randomized controlled trial.
      ]. No evidence was found to answer the research question of quality of life (QoL) and regarding patient satisfaction, 70.6% (IG), as opposed to 39.4% (CG) of patients, reported moderate to high satisfaction (7.5% difference). Compliance was 61.7% in the IG and 54.4% in the CG group [
      • Schoenen J.
      • Vandersmissen B.
      • Jeangette S.
      • Herroelen L.
      • Vandenheede M.
      • Gerard P.
      • et al.
      Migraine prevention with a supraorbital transcutaneous stimulator: a randomized controlled trial.
      ] and it was assessed by a built-in electronic system in the e-TNS device that recorded if the use of e-TNS stimulators followed the study recommendations (20 min/day) [
      • Russo A.
      • Tessitore A.
      • Conte F.
      • Marcuccio L.
      • Giordano A.
      • Tedeschi G.
      Transcutaneous supraorbital neurostimulation in “de novo” patients with migraine without aura: the first Italian experience.
      ].

      2.4.2 Acute treatment

      In terms of the acute treatment effect of e-TNS during the migraine attack, the intensity of pain was measured on a VAS scale (0-no pain to 10-maximum pain), immediately after one hour treatment session, and at two and 24 h after the beginning of the treatment initiation. In both EM and CM, patients in the IG had a mean reduction of 3.46 points (SD = 2.32) at one hour post intervention, while patients in the CG of 1.78 points (SD = 1.89). The difference (net gain in the IG compared to baseline) was 1.68 points (p = .0001) [
      • Chou E.D.
      Acute Treatment of Migraine With e-TNS (ACME).
      ]. At two hours post intervention, patients in the IG had a mean reduction of 2.87 points (SD = 2.24), while patients in the CG of 1.85 points (SD = 1.96). The difference (net gain in the IG compared to baseline) was 1.02 points (p = .028) [
      • Chou E.D.
      Acute Treatment of Migraine With e-TNS (ACME).
      ]. At 24 h post intervention, patients in the IG had a mean reduction of 3.46 points (SD = 2.65), while patients in the CG had a reduction of 2.38 points (SD = 2.27). The difference (net gain in the IG compared to baseline) was 1.08 points (p = .062) [
      • Chou E.D.
      Acute Treatment of Migraine With e-TNS (ACME).
      ].
      Furthermore, with respect to acute antimigraine medication two hours post treatment in EM and CM, there were three patients (5.8%) on acute medication in the IG and two patients (3.7%) in the CG. The difference (net loss in the IG compared to baseline) was 1 patient (p = .666) [
      • Chou E.D.
      Acute Treatment of Migraine With e-TNS (ACME).
      ]. At 24 h post intervention, there were 18 patients in the IG (34.6%) on acute antimigraine medication, while in the CG, there were 21 patients (38.9%). The difference (net gain in the IG compared to baseline) was 3 patients (3.3%) (p = 1) [
      • Chou E.D.
      Acute Treatment of Migraine With e-TNS (ACME).
      ]. No evidence was found to answer the research questions of QoL, patient satisfaction, compliance, freedom from headache pain at 2/24 h post-acute treatment, and freedom from nausea, vomiting, sensitivity to light and sound at 2 h post-acute treatment. For further details, see Tables 2 and 3.

      2.5 Safety

      2.5.1 Prevention

      Concerning the preventive use of e-TNS, no serious adverse device effects (SADEs) occurred. In terms of adverse device effects (ADEs), two studies reported that there were none [
      • Schoenen J.
      • Vandersmissen B.
      • Jeangette S.
      • Herroelen L.
      • Vandenheede M.
      • Gerard P.
      • et al.
      Migraine prevention with a supraorbital transcutaneous stimulator: a randomized controlled trial.
      ,
      • Russo A.
      • Tessitore A.
      • Conte F.
      • Marcuccio L.
      • Giordano A.
      • Tedeschi G.
      Transcutaneous supraorbital neurostimulation in “de novo” patients with migraine without aura: the first Italian experience.
      ], while intolerance to paraesthesia (burning sensation) was reported in 34.3% of patients in one study [
      • Vikelis M.
      • Dermitzakis E.V.
      • Spingos K.C.
      • Vasiliadis G.G.
      • Vlachos G.S.
      • Kararizou E.
      Clinical experience with transcutaneous supraorbital nerve stimulation in patients with refractory migraine or with migraine and intolerance to topiramate: a prospective exploratory clinical study.
      ]. Furthermore, headache after stimulation as well as neck tension were reported in one study [
      • Di Fiore P.
      • Bussone G.
      • Galli A.
      • Didier H.
      • Peccarisi C.
      • D’Amico D.
      • et al.
      Transcutaneous supraorbital neurostimulation for the prevention of chronic migraine: a prospective, open-label preliminary trial.
      ]: headache occurred in 8.7% of patients, while neck tension in 4.3%. In one study, 10.8% of patients experienced technical issues with the e-TNS device [
      • Vikelis M.
      • Dermitzakis E.V.
      • Spingos K.C.
      • Vasiliadis G.G.
      • Vlachos G.S.
      • Kararizou E.
      Clinical experience with transcutaneous supraorbital nerve stimulation in patients with refractory migraine or with migraine and intolerance to topiramate: a prospective exploratory clinical study.
      ].

      2.5.2 Acute treatment

      Concerning the acute treatment use of e-TNS, no SADEs occurred. In terms of ADEs, one study reported that there were none [
      • Chou D.E.
      • Gross G.J.
      • Casadei C.H.
      • Yugrakh M.S.
      External trigeminal nerve stimulation for the acute treatment of migraine: open-label trial on safety and efficacy.
      ]. Intolerance to paraesthesia was documented in two acute treatment studies in 5.8% (IG n = 52) vs. 1.9% (CG n = 54) [
      • Chou E.D.
      Acute Treatment of Migraine With e-TNS (ACME).
      ] and in 11.9% of patients [
      • Mann J.
      Abortive Treatment of Migraine With the Cefaly® Abortive Program Device.
      ]. Nausea after stimulation was reported in two studies in 1.9% (IG n = 52) vs. 0% (CG n = 54) [
      • Chou E.D.
      Acute Treatment of Migraine With e-TNS (ACME).
      ] and in 3.5% of patients [
      • Mann J.
      Abortive Treatment of Migraine With the Cefaly® Abortive Program Device.
      ]. Furthermore, arousal changes (insomnia, sleepiness/fatigue, drowsiness), dizziness, vomiting, pain in the jaw, discomfort in teeth, pain in eyes, and cold feet occurred in one study all in 1.7% of patients [
      • Mann J.
      Abortive Treatment of Migraine With the Cefaly® Abortive Program Device.
      ]. In the same study, 18.3% of patients reported skin allergy/irritation. For further details, see Tables 2 and 3.

      3. Discussion

      To our knowledge, this is the first systematic review on the preventive and acute treatment use of e-TNS in EM as well as CM patients. We included two RCTs for the analysis of clinical effectiveness and additional five prospective case series complementing the analysis of safety.
      Concerning prevention, the results from the RCT show statistically significant differences between e-TNS and sham treatment in EM patients with respect to reduction of migraine attacks, migraine days, headache days, and acute antimigraine drug intake. With respect to responder rate, 12 (40%) patients in the IG responded as opposed to 4 (13.8%) patients in the CG. And with respect to satisfaction, 70.6% in the IG reported moderate to high satisfaction as opposed to 39.4% in the CG.
      Efficacy data from the three preventive prospective case series suggest a better efficacy profile than the preventive RCT [
      • Schoenen J.
      • Vandersmissen B.
      • Jeangette S.
      • Herroelen L.
      • Vandenheede M.
      • Gerard P.
      • et al.
      Migraine prevention with a supraorbital transcutaneous stimulator: a randomized controlled trial.
      ], namely that e-TNS reduced the number of monthly migraine attacks by 2.5 instances [
      • Russo A.
      • Tessitore A.
      • Conte F.
      • Marcuccio L.
      • Giordano A.
      • Tedeschi G.
      Transcutaneous supraorbital neurostimulation in “de novo” patients with migraine without aura: the first Italian experience.
      ], monthly migraine days by 3.5 days [
      • Russo A.
      • Tessitore A.
      • Conte F.
      • Marcuccio L.
      • Giordano A.
      • Tedeschi G.
      Transcutaneous supraorbital neurostimulation in “de novo” patients with migraine without aura: the first Italian experience.
      ] and by 31% [
      • Di Fiore P.
      • Bussone G.
      • Galli A.
      • Didier H.
      • Peccarisi C.
      • D’Amico D.
      • et al.
      Transcutaneous supraorbital neurostimulation for the prevention of chronic migraine: a prospective, open-label preliminary trial.
      ], and monthly headache days by two days [
      • Vikelis M.
      • Dermitzakis E.V.
      • Spingos K.C.
      • Vasiliadis G.G.
      • Vlachos G.S.
      • Kararizou E.
      Clinical experience with transcutaneous supraorbital nerve stimulation in patients with refractory migraine or with migraine and intolerance to topiramate: a prospective exploratory clinical study.
      ]. Also, responder rate in migraine days was 75% in one study [
      • Russo A.
      • Tessitore A.
      • Conte F.
      • Marcuccio L.
      • Giordano A.
      • Tedeschi G.
      Transcutaneous supraorbital neurostimulation in “de novo” patients with migraine without aura: the first Italian experience.
      ]. Furthermore, monthly acute antimigraine drug intake was reduced by 40.7%, 49.6%, and 46.3%, respectively [
      • Di Fiore P.
      • Bussone G.
      • Galli A.
      • Didier H.
      • Peccarisi C.
      • D’Amico D.
      • et al.
      Transcutaneous supraorbital neurostimulation for the prevention of chronic migraine: a prospective, open-label preliminary trial.
      ,
      • Russo A.
      • Tessitore A.
      • Conte F.
      • Marcuccio L.
      • Giordano A.
      • Tedeschi G.
      Transcutaneous supraorbital neurostimulation in “de novo” patients with migraine without aura: the first Italian experience.
      ,
      • Vikelis M.
      • Dermitzakis E.V.
      • Spingos K.C.
      • Vasiliadis G.G.
      • Vlachos G.S.
      • Kararizou E.
      Clinical experience with transcutaneous supraorbital nerve stimulation in patients with refractory migraine or with migraine and intolerance to topiramate: a prospective exploratory clinical study.
      ]. In terms of satisfaction, 83.3% [
      • Russo A.
      • Tessitore A.
      • Conte F.
      • Marcuccio L.
      • Giordano A.
      • Tedeschi G.
      Transcutaneous supraorbital neurostimulation in “de novo” patients with migraine without aura: the first Italian experience.
      ] and 65.7% [
      • Vikelis M.
      • Dermitzakis E.V.
      • Spingos K.C.
      • Vasiliadis G.G.
      • Vlachos G.S.
      • Kararizou E.
      Clinical experience with transcutaneous supraorbital nerve stimulation in patients with refractory migraine or with migraine and intolerance to topiramate: a prospective exploratory clinical study.
      ] of patients were satisfied.
      Concerning acute treatment, the RCT shows statistically significant differences between e-TNS and sham treatment suggesting more improvement in pain reduction than sham on a VAS scale (out of 11 points) at 1/2/24 h post-acute treatment. The size of the difference measured on the VAS scale, however, is of questionable clinical relevance as concerning the minimal clinically important difference (MCID), the results oscillate around the lower end of the clinically meaningful benefit threshold [
      • Olsen M.F.
      • Bjerre E.
      • Damkjaer Hanse M.
      • Hilden J.
      • Landler N.E.
      • Tendal B.
      • et al.
      Pain relief that matters to patients: systematic review of empirical studies assessing the minimum clinically important difference in acute pain.
      ]. There was also an increase of one IG patient in acute antimigraine drug intake at two hours post-acute treatment compared to CG, but a decrease of three IG patients compared to CG at 24 h.
      Efficacy data from the two acute treatment prospective case series also suggest a better efficacy profile than the acute treatment RCT, namely that 35.4%/25% of patients were headache pain free at 2/24 h post-acute treatment, respectively, and 36.7% were free from nausea, vomiting, and sensitivity to light and sound at 2 h [
      • Mann J.
      Abortive Treatment of Migraine With the Cefaly® Abortive Program Device.
      ]. On the VAS scale, there was a 3.22/2.98 decrease of pain at 1/2 h, respectively [
      • Chou D.E.
      • Gross G.J.
      • Casadei C.H.
      • Yugrakh M.S.
      External trigeminal nerve stimulation for the acute treatment of migraine: open-label trial on safety and efficacy.
      ], and 34.6% and 50% of patients were on acute medication at 24 h post-acute treatment, respectively [
      • Mann J.
      Abortive Treatment of Migraine With the Cefaly® Abortive Program Device.
      ,
      • Chou D.E.
      • Gross G.J.
      • Casadei C.H.
      • Yugrakh M.S.
      External trigeminal nerve stimulation for the acute treatment of migraine: open-label trial on safety and efficacy.
      ].
      Concerning safety, no SADEs occurred neither in the RCTs nor in the case series. The reporting of AEDs, however, remains to be a point of concern as the largest treatment case series study with 60 patients reports several AEDs that are unreported in all the remaining studies (dizziness, vomiting, pain in the jaw, discomfort in teeth, and pain in eyes). Each occurred only in 1/60 patients.

      3.1 Internal and external validity

      Concerning the effectiveness (RCTs) of prevention and acute treatment with e-TNS, the quality of evidence was low to very low. The main reasons were the small sample size, uncertainty about sufficient reporting of AEDs, and the use of an inappropriate comparator (see Tables 6 and 7). For the comparison of e-TNS with the standard practice comparators such as triptans, NSAIDs/paracetamol for acute treatment, and propranolol and topiramate (or others) for prevention, no evidence was found. Concerning safety, the quality of evidence ranged from high, moderate, to low and very low (see Tables 4 and 5). Furthermore, two studies (case-series) were judged to have a high risk of confounding as co-interventions were either not clearly described [
      • Vikelis M.
      • Dermitzakis E.V.
      • Spingos K.C.
      • Vasiliadis G.G.
      • Vlachos G.S.
      • Kararizou E.
      Clinical experience with transcutaneous supraorbital nerve stimulation in patients with refractory migraine or with migraine and intolerance to topiramate: a prospective exploratory clinical study.
      ], or it was clearly stated that preventive as well as acute treatments for CM were not changed during the study [
      • Di Fiore P.
      • Bussone G.
      • Galli A.
      • Didier H.
      • Peccarisi C.
      • D’Amico D.
      • et al.
      Transcutaneous supraorbital neurostimulation for the prevention of chronic migraine: a prospective, open-label preliminary trial.
      ].
      Table 4Risk of bias – study level (randomised studies).
      TrialAdequate generation of randomisation sequenceAdequate allocation concealmentBlindingSelective outcome

      reporting unlikely
      No other aspects which increase the risk of biasRisk of bias – study level
      PatientTreating Physician
      Schoenen et al. [
      • Magis D.
      • D’Ostilio K.
      • Thibaut A.
      • De Pasqua V.
      • Gerard P.
      • Hustinx R.
      • et al.
      Cerebral metabolism before and after external trigeminal nerve stimulation in episodic migraine.
      ] (2013)
      Unclear
      Insufficient information about sequence generation process.
      YesYesYesUnclear
      Even though it was reported that no adverse events or side effects occurred during the trial in both treatment arms, it seems unlikely.
      YesModerate
      Chou et al. [
      • Higgins J.
      • Altman D.G.
      • Gøtzsche P.C.
      • Jüni P.
      • Moher D.
      • Oxman A.D.
      • et al.
      The cochrane collaboration’s tool for assessing risk of bias in randomised trials.
      ]
      Unclear
      Even though it was reported that no adverse events or side effects occurred during the trial in both treatment arms, it seems unlikely.
      Unclear
      Method of concealment is not described to allow a definite judgement.
      YesYesUnclear
      Considering the sample size, there are reasons to think that not all ADEs s and SADEs were reported.
      No
      Conflict of interest: study sponsored by the manufacturer.
      High
      a Insufficient information about sequence generation process.
      b Even though it was reported that no adverse events or side effects occurred during the trial in both treatment arms, it seems unlikely.
      c Method of concealment is not described to allow a definite judgement.
      d Considering the sample size, there are reasons to think that not all ADEs s and SADEs were reported.
      e Conflict of interest: study sponsored by the manufacturer.
      Table 5Risk of bias – study level (case series).
      Study reference/IDRusso et al. [
      • Didier H.A.
      • Di Fiore P.
      • Marchetti C.
      • Tullo V.
      • Frediani F.
      • Arlotti M.
      • et al.
      Electromyography data in chronic migraine patients by using neurostimulation with the Cefaly device.
      ]
      DiFiore et al. [
      • Goadsby P.J.
      • Holland P.R.
      • Martins-Oliveira M.
      • Hoffmann J.
      • Schankin C.
      • Akerman S.
      Pathophysiology of migraine: a disorder of sensory processing.
      ]
      Vikelis et al. [
      • Vecchio E.
      • Gentile E.
      • Franco G.
      • Ricci K.
      • de Tommaso M.
      Effects of external trigeminal nerve stimulation (eTNS) on laser evoked cortical potentials (LEP): a pilot study in migraine patients and controls.
      ]
      Chou et al. [
      • Institute of Health Economics (IHE)
      Quality Appraisal of Case Series Studies Checklist.
      ]
      Mann et al. [
      • Guyatt G.
      • Oxman A.D.
      • Akl E.A.
      • Kunz R.
      • Vist G.
      • JEA B.
      GRADE guidelines: 1. Introduction-GRADE evidence profiles and summary of findings tables.
      ]
      Study objective
      Was the hypothesis/aim/objective of the study clearly stated?YesYesYesYesYes
      Study design
      Was the study conducted prospectively?YesYesYesYesYes
      Were the cases collected in more than one centre?NoYesYesNoNo
      Were patients recruited consecutively?YesYesUnclearUnclearUnclear
      Study population
      Were the characteristics of the participants included in the study described?YesYesPartialPartialYes
      Were the eligibility criteria (inclusion and exclusion criteria) for entry into the studyclearly stated?Partial
      Only the exclusion criteria were clearly stated in the study.
      Partial
      Exclusion criteria were not explicitly mention in the study.
      Partial
      Information on pulse frequency, pulse width and maximal intensity was missing.
      YesYes
      Did participants enter the study at similar point in the disease?YesYesNoNoYes
      Intervention and co-intervention
      Was the intervention clearly described?YesYesNo
      Information on pulse frequency, pulse width and maximal intensity was missing.
      YesYes
      Were additional interventions (co-interventions) clearly described?YesNo
      “Preexisting preventive and acute treatments for CM were not changed” was an insufficient explanation of co-interventions present.
      NoYesYes
      Outcome measure
      Were relevant outcome measures established a priori?YesPartial
      Only the primary endpoints were mentioned a priori.
      PartialYesYes
      Were outcome assessors blinded to the intervention that patients received?Unclear
      No information is given if the outcome assessors were blinded to the intervention.
      Unclear
      However, the validity of used methods to measure outcomes was unclear.
      Unclear
      However, the validity of used methods to measure outcomes was unclear.
      Unclear
      However, the validity of used methods to measure outcomes was unclear.
      Unclear
      However, the validity of used methods to measure outcomes was unclear.
      Were the relevant outcomes measured using appropriate objective/subjective methods?Yes
      However, the validity of used methods to measure outcomes was unclear.
      PartialPartialPartialPartial
      Were the relevant outcomes measured before and after intervention?Yes
      Satisfaction was not measure before the intervention in any of the studies.
      Yes
      It is not stated at what point or where parametric (paired t-test) and non-parametric (Wilcoxon sign-rank test) measures have been applied.
      Yes
      It is not stated at what point or where parametric (paired t-test) and non-parametric (Wilcoxon sign-rank test) measures have been applied.
      Yes
      It is not stated at what point or where parametric (paired t-test) and non-parametric (Wilcoxon sign-rank test) measures have been applied.
      Yes
      It is not stated at what point or where parametric (paired t-test) and non-parametric (Wilcoxon sign-rank test) measures have been applied.
      Statistical analysis
      Were the statistical tests used to assess the relevant outcomes appropriate?Yes
      It is not stated at what point or where parametric (paired t-test) and non-parametric (Wilcoxon sign-rank test) measures have been applied.
      No
      Only descriptive statistics were used.
      No
      Fisher's exact test and the Mann-Whitney test are mentioned in the methods section, but the results fail to present any results from these tests.
      YesNo
      Only descriptive statistics were used.
      Results and conclusions
      Was follow-up long enough for important events and outcomes to occur?No
      Unclear if of follow-up is enough for the effect of preventive treatment. It is taken for granted that TENs machines should have minor ADEs and SADEs even if applied onto the head.
      No
      It was reported that no adverse events occurred in the study population, yet it is assumed otherwise.
      No
      It was reported that no adverse events occurred in the study population, yet it is assumed otherwise.
      No
      It was reported that no adverse events occurred in the study population, yet it is assumed otherwise.
      No
      It was reported that no adverse events occurred in the study population, yet it is assumed otherwise.
      Was the loss to follow-up reported?YesYesYesYesYes
      Did the study provide estimates of random variability in the data analysis of relevant outcomes?YesPartialNoPartialNo
      Were adverse events reported?Yes
      It was reported that no adverse events occurred in the study population, yet it is assumed otherwise.
      YesYesPartialYes
      Were the conclusions of the study supported by results?No
      The study design cannot meet the conclusions about effectiveness.
      No
      The study design cannot meet the conclusions about effectiveness and the conclusions are only made on the basis of a subgroup of patients from the results.
      No
      The study design cannot meet the conclusions about effectiveness and the conclusions are only made on the basis of a subgroup of patients from the results.
      No
      The study design cannot meet the conclusions about effectiveness and the conclusions are only made on the basis of a subgroup of patients from the results.
      No
      The study design cannot meet the conclusions about effectiveness and the conclusions are only made on the basis of a subgroup of patients from the results.
      Study not published yet, data available only at clinicaltrials.gov.
      Competing interest and source of support
      Were both competing interest and source of support for the study reported?YesYesYesPartial
      Source of support for the study is unclear.
      Partialr
      Overall Risk of biasLowModerateHighModerateModerate
      a Only the exclusion criteria were clearly stated in the study.
      b Exclusion criteria were not explicitly mention in the study.
      c Information on pulse frequency, pulse width and maximal intensity was missing.
      d “Preexisting preventive and acute treatments for CM were not changed” was an insufficient explanation of co-interventions present.
      e Only the primary endpoints were mentioned a priori.
      f No information is given if the outcome assessors were blinded to the intervention.
      g However, the validity of used methods to measure outcomes was unclear.
      h Satisfaction was not measure before the intervention in any of the studies.
      i It is not stated at what point or where parametric (paired t-test) and non-parametric (Wilcoxon sign-rank test) measures have been applied.
      j Only descriptive statistics were used.
      k Fisher's exact test and the Mann-Whitney test are mentioned in the methods section, but the results fail to present any results from these tests.
      l Only descriptive statistics were used.
      m Unclear if of follow-up is enough for the effect of preventive treatment. It is taken for granted that TENs machines should have minor ADEs and SADEs even if applied onto the head.
      n It was reported that no adverse events occurred in the study population, yet it is assumed otherwise.
      o The study design cannot meet the conclusions about effectiveness.
      p The study design cannot meet the conclusions about effectiveness and the conclusions are only made on the basis of a subgroup of patients from the results.
      q Study not published yet, data available only at clinicaltrials.gov.
      r Source of support for the study is unclear.
      Table 6Evidence profile table: efficacy and safety of e-TNS for the prevention of EM and CM.
      Certainty assessmentEstimate of effectCertaintyImportance
      № of studiesStudy designRisk of biasInconsistencyIndirectnessImprecisionOther considerations
      Reduction in monthly migraine attacks (assessed with: mean number of days)
      1Randomised controlled trialNot seriousNot seriousSerious
      Inappropriate comparator.
      Serious
      Small sample size.
      NoneThe mean reduction in the number of monthly migraine attacks in the IG (n = 34) was 0.67 more than in the CG (n = 33). Difference between IG and CG was p = .044.⨁⨁◯◯ LowImportant
      Reduction in monthly migraine days (assessed with: mean)
      1Randomised controlled trialNot seriousNot seriousSerious
      Inappropriate comparator.
      Serious
      Small sample size.
      NoneThe mean reduction in the number of monthly migraine days in the IG (n = 34) was 1.74 more than in the CG (n = 33). Difference between IG and CG was p = .054.⨁⨁◯◯ LowCritical
      Reduction of monthly headache days (assessed with: mean)
      1Randomised controlled trialNot seriousNot seriousSerious
      Inappropriate comparator.
      Serious
      Small sample size.
      NoneThe mean reduction in the number of monthly headache days in the IG (n = 34) was 2.28 more than in the CG (n = 33). Difference between IG and CG was p = .041.⨁⨁◯◯ LowImportant
      Reduction in the number of monthly acute antimigraine drug intake
      1Randomised controlled trialNot seriousNot seriousSerious
      Inappropriate comparator.
      Serious
      Small sample size.
      NoneThe mean reduction in the number of monthly acute antimigraine drug intake in the IG (n = 34) was 4.24 more than in the CG (n = 33). Difference between IG and CG was p = .072.⨁⨁◯◯ LowCritical
      High or moderate satisfaction
      1Randomised controlled trialSerious
      Uncertainty about sufficient reporting of adverse events.
      Not seriousSerious
      Inappropriate comparator.
      Serious
      Small sample size.
      None70.6% (IG, n = 34) vs.39.4% (CG, n = 33)⨁◯◯◯ Very lowCritical
      Serious Adverse device effects
      1Randomised controlled trialSerious
      Uncertainty about sufficient reporting of adverse events.
      Not seriousSerious
      Inappropriate comparator.
      Serious
      Small sample size.
      None0/34 (IG) vs. 0/33(CG)⨁◯◯◯ Very lowCritical
      Adverse device effects
      1Randomised controlled trialSerious
      Uncertainty about sufficient reporting of adverse events.
      Not seriousSerious
      Inappropriate comparator.
      Serious
      Small sample size.
      None0/34 (IG) vs. 0/33 (CG)⨁◯◯◯ Very lowImportant
      Abbreviations: CG = Control group, CI = Confidence interval, IG = Interventional group, n = number, pts = Patients, RCT = Randomised controlled trial, VAS = Visual Analogue Scale.
      GRADE Working Group grades of evidence: High certainty: We are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect. Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect.
      a Inappropriate comparator.
      b Small sample size.
      c Uncertainty about sufficient reporting of adverse events.
      Table 7Evidence profile table: efficacy and safety of e-TNS for the acute treatment of EM and CM.
      Certainty assessment№ of patientsEffectCertaintyImportance
      № of studiesStudy designRisk of biasInconsistencyIndirectnessImprecisionOther considerationse-TNSShamRelative (95% CI)Absolute (95% CI)
      Reductionin pain score units on scale compared to baseline at 1 h
      1Randomised controlled trialSerious
      Funded by the manufacturer.
      Insufficient information about sequence generation process.
      Method of concealment is not described to allow a definite judgement.
      Not seriousSerious
      Inappropriate comparator.
      Serious
      Small sample size.
      NoneThe mean improvement of IG (n = 52) over CG (n = 54) in reduction in pain score units on VAS scale compared to baseline at 1 h was 1.68 (pGroup = 0.0001).⨁◯◯◯ Very lowCritical
      Reduction in pain score units on scale compared to baseline at 24 h
      1Randomised controlled trialSerious
      Funded by the manufacturer.
      Insufficient information about sequence generation process.
      Method of concealment is not described to allow a definite judgement.
      Not seriousSerious
      Inappropriate comparator.
      Serious
      Small sample size.
      NoneThe mean improvement of IG (n = 52) over CG (n = 54) in reduction in pain score units on VAS scale compared to baseline at 1 h was 1.02 (pGroup = 0.028).⨁◯◯◯ Very lowImportant
      Reduction in pain score units on scale compared to baseline at 24 h
      1Randomised controlled trialSerious
      Funded by the manufacturer.
      Insufficient information about sequence generation process.
      Method of concealment is not described to allow a definite judgement.
      Not seriousSerious
      Inappropriate comparator.
      Serious
      Small sample size.
      NoneThe mean improvement of IG (n = 52) over CG (n = 54) in reduction in pain score units on VAS scale compared to baseline at 1 h was 1.08 (pGroup = 0.062).⨁◯◯◯ Very lowCritical
      Patients on acute medication at 2 h
      1Randomised controlled trialSerious
      Funded by the manufacturer.
      Insufficient information about sequence generation process.
      Method of concealment is not described to allow a definite judgement.
      Not seriousSerious
      Inappropriate comparator.
      Serious
      Small sample size.
      None3/52 (IG) vs. 2/54 (CG)⨁◯◯◯ Very lowCritical
      Patients on acute medication at 24 h
      1Randomised controlled trialSerious
      Funded by the manufacturer.
      Insufficient information about sequence generation process.
      Method of concealment is not described to allow a definite judgement.
      Not seriousSerious
      Inappropriate comparator.
      Serious
      Small sample size.
      None18/52 (IG) vs. 21/54 (CG)⨁◯◯◯ Very lowImportant
      Serious adverse device effects
      1Randomised controlled trialSerious
      Funded by the manufacturer.
      Insufficient information about sequence generation process.
      Method of concealment is not described to allow a definite judgement.
      Not seriousSerious
      Inappropriate comparator.
      Serious
      Small sample size.
      None0/52 (IG) vs. 0/54 (CG)⨁◯◯◯ Very lowCritical
      Adverse device effects
      1Randomised controlled trialSerious
      Funded by the manufacturer.
      Insufficient information about sequence generation process.
      Method of concealment is not described to allow a definite judgement.
      Not seriousSerious
      Inappropriate comparator.
      Serious
      Small sample size.
      NoneIntolerance to paresthesia in 3/52 (IG) vs. 1/54 (CG). Nausea after stimulation in 1/52 (IG) vs. 0/54 (CG)⨁◯◯◯

      Very low
      Important
      Abbreviations: CG = Control group, CI = Confidence interval, IG = Interventional group, n = number, pts = Patients, RCT = Randomised controlled trial, VAS = Visual Analogue Scale.
      GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect. Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect.
      a Funded by the manufacturer.
      b Insufficient information about sequence generation process.
      c Method of concealment is not described to allow a definite judgement.
      d Inappropriate comparator.
      e Small sample size.
      Challenges with interpreting the data arise when in the acute treatment use of e-TNS, one study outlines the baseline most bothersome migraine symptoms (nausea, vomiting, sensitivity to light and sound) [
      • Mann J.
      Abortive Treatment of Migraine With the Cefaly® Abortive Program Device.
      ], but these baseline symptoms are not outlined in the remaining two acute treatment studies, thus undermining their internal validity [
      • Chou E.D.
      Acute Treatment of Migraine With e-TNS (ACME).
      ,
      • Chou D.E.
      • Gross G.J.
      • Casadei C.H.
      • Yugrakh M.S.
      External trigeminal nerve stimulation for the acute treatment of migraine: open-label trial on safety and efficacy.
      ].
      In terms of external validity, the differences between inclusion and exclusion criteria undermine the generalizability of conclusions. The reason is that prevention studies represent the real clinical context, as the patients used the e-TNS in their homes, but the acute treatment studies were conducted in the hospital setting, yet the e-TNS should be used in the home setting for acute treatment as well.
      Furthermore, compliance is considered to be one of the key issues. In the preventive studies, compliance ranged in the case series from 81.8% to 83.3% [
      • Di Fiore P.
      • Bussone G.
      • Galli A.
      • Didier H.
      • Peccarisi C.
      • D’Amico D.
      • et al.
      Transcutaneous supraorbital neurostimulation for the prevention of chronic migraine: a prospective, open-label preliminary trial.
      ,
      • Russo A.
      • Tessitore A.
      • Conte F.
      • Marcuccio L.
      • Giordano A.
      • Tedeschi G.
      Transcutaneous supraorbital neurostimulation in “de novo” patients with migraine without aura: the first Italian experience.
      ,
      • Vikelis M.
      • Dermitzakis E.V.
      • Spingos K.C.
      • Vasiliadis G.G.
      • Vlachos G.S.
      • Kararizou E.
      Clinical experience with transcutaneous supraorbital nerve stimulation in patients with refractory migraine or with migraine and intolerance to topiramate: a prospective exploratory clinical study.
      ], and in the preventive RCT [
      • Schoenen J.
      • Vandersmissen B.
      • Jeangette S.
      • Herroelen L.
      • Vandenheede M.
      • Gerard P.
      • et al.
      Migraine prevention with a supraorbital transcutaneous stimulator: a randomized controlled trial.
      ], it was 61.7% (IG) vs. 54.4% (CG). It was not reported in the acute treatment RCT [
      • Chou D.E.
      • Gross G.J.
      • Casadei C.H.
      • Yugrakh M.S.
      External trigeminal nerve stimulation for the acute treatment of migraine: open-label trial on safety and efficacy.
      ], but in one acute treatment case series, it was 82.7% [
      • Mann J.
      Abortive Treatment of Migraine With the Cefaly® Abortive Program Device.
      ]. In a survey with 2313 patients, 46.6% of patients were unsatisfied who, in terms of compliance, used the device for the recommended period of time only in 48.6% of cases [
      • Magis D.
      • Sava S.
      • d’Elia T.S.
      • Baschi R.
      • Schoenen J.
      Safety and patients’ satisfaction of transcutaneous supraorbital neurostimulation (tSNS) with the Cefaly device in headache treatment: a survey of 2,313 headache sufferers in the general population.
      ]. The real-life compliance with e-TNS is put into question because the relatively high case series compliance data are contrasted with lower RCT and survey data.

      3.2 Socio-economic and ethical considerations

      When considering socio-economic and ethical aspects of e-TNS, the effects have to be reflected over against the principles of beneficence, non-maleficence, autonomy, distribute justice, and uncertainty. On the one hand, by being applied in the home setting (thus securing patient autonomy and easy access), e-TNS claims to reduce the burden on the in−/outpatient sector by reducing the pharmacological and other healthcare spending associated with the use of the current preventive and acute treatment options for EM and CM patients (freeing of resources and accordingly distribute justice). And, if proven to be more effective than the best practice comparators, it may also reduce the economic loss associated with decreased economic productivity of migraine patients at the expense of the out-of-pocket payment for the device made by the patients themselves [
      • Brown J.S.
      • Papadopoulos G.
      • Neumann P.J.
      • Price M.
      • Friedman M.
      • Menzin J.
      Cost-effectiveness of migraine prevention: the case of topiramate in the UK.
      ]. E-TNS also claims to be associated with less side effects than the current pharmacological therapy and thus, it may better protect the principles of medical beneficence and patient's autonomy.
      On the other hand, however, the lack of clarity behind the mechanism of action of e-TNS casts doubts over its positive safety profile [
      • Didier H.A.
      • Di Fiore P.
      • Marchetti C.
      • Tullo V.
      • Frediani F.
      • Arlotti M.
      • et al.
      Electromyography data in chronic migraine patients by using neurostimulation with the Cefaly device.
      ,
      • Magis D.
      • D’Ostilio K.
      • Thibaut A.
      • De Pasqua V.
      • Gerard P.
      • Hustinx R.
      • et al.
      Cerebral metabolism before and after external trigeminal nerve stimulation in episodic migraine.
      ,
      • Vecchio E.
      • Gentile E.
      • Franco G.
      • Ricci K.
      • de Tommaso M.
      Effects of external trigeminal nerve stimulation (eTNS) on laser evoked cortical potentials (LEP): a pilot study in migraine patients and controls.
      ,
      • Ekizoglu E.
      • Sozer-Topçular N.
      • Baykan B.
      • Oge A.E.
      Assessment of excitability at the brainstem and cortex in primary headaches with allodynia.
      ]. Moreover, the fact the e-TNS is placed in such proximity to the brain area (at the patient's forehead) casts questions about the possible long-term ADEs (and the respective principles of non-maleficence) that are not yet measured by the current evidence. It is not clear to what extent the electrical field applied in such close proximity to the brain for such an extended period of time influences the brain. It is also difficult to unify different levels of stimulus intensity, and so to prevent breaching the principle of non-maleficence, larger controlled trials are needed to match the size of the population that e-TNS targets. At the time of writing, there was only one ongoing RCT for the acute treatment use of e-TNS that aimed to recruit 600 patients with an estimated primary completion date of October 2018, however, it lacks the measurement of any longer-term outcomes (more than 24 h post intervention) (NCT03465904).

      3.3 Limitations of evidence

      Given the small sample size of the studies on e-TNS, the current conclusions about effectiveness and the positive safety profile are considered to be inflated. The evidence found was only partly relevant to answering the research question. Both RCTs identified were relevant for excluding placebo effects, but best practice interventions should be used as comparators for establishing the clinical benefit of e-TNS. The reason is that the target population of e-TNS are not only patients refractory to medication, but also drug responsive patients, which makes e-TNS aim to replace the use of medication. That is why a controlled trial comparing e-TNS to either of the above outlined (preventive or acute) treatments is necessary. At the time of writing of this systematic review, there was one preventive trial available, but it was placebo controlled and its broad inclusion criteria and low quality of reporting did not allow for data extraction and data analysis [
      • Przeklasa-Muszynska A.
      • Skrzypiec K.
      • Kocot-Kepska M.
      • Dobrogowski J.
      • Wiatr M.
      • Mika J.
      Non-invasive transcutaneous Supraorbital Neurostimulation (tSNS) using Cefaly((R)) device in prevention of primary headaches.
      ]. Also, an RCT on the combination of flunarizine and e-TNS compared to e-TNS or flunarizine alone was brought to our attention after the completion of our study [
      • Jiang L.
      • Yuan D.L.
      • Li M.
      • Liu C.
      • Liu Q.
      • Zhang Y.
      • et al.
      Combination of flunarizine and transcutaneous supraorbital neurostimulation improves migraine prophylaxis.
      ]. For the sake of not undermining the methodology of this systematic review however, this study was not retrospectively included in the analysis. The study included 154 patients and claimed that for episodic migraine patients at three months follow-up, 50% responder rate was significantly higher in the combination therapy group (78.43%) as opposed to monotherapy groups of flunarizine (46.15%) or transcutaneous supraorbital neurostimulation (39.22%) alone. Reduction of migraine intensity and rescue medication intake was observed to be higher in combination group as well with no significant difference in adverse events.
      Furthermore, outcomes measured in the studies were judged to be relevant to patients' experience, but a standardized evaluation of satisfaction was lacking (especially in the RCTs). It is important to note that the patient relevant endpoint of QoL was not measured or reported in any of the studies. Also, in acute treatment studies with medication, patients are followed for 48 h for the purpose of measuring headache recurrence. The follow-up of 24 h that is applied to all three acute treatment studies is considered to be too short [
      • Chou E.D.
      Acute Treatment of Migraine With e-TNS (ACME).
      ,
      • Mann J.
      Abortive Treatment of Migraine With the Cefaly® Abortive Program Device.
      ,
      • Chou D.E.
      • Gross G.J.
      • Casadei C.H.
      • Yugrakh M.S.
      External trigeminal nerve stimulation for the acute treatment of migraine: open-label trial on safety and efficacy.
      ]. Also, the standard outcome parameter recommended by the International Headache Society is the proportion of patients being pain free at two hours post treatment. This outcome was used as the primary outcome parameter only in Mann et al. [
      • Mann J.
      Abortive Treatment of Migraine With the Cefaly® Abortive Program Device.
      ]. It was not measured at all in the remaining studies. Moreover, consistency of the effect of e-TNS is undermined because, in medication studies, several attacks must be treated in one person to prove that the acute therapy works and that was not the case.

      4. Conclusion

      While e-TNS has the potential to improve migraine symptoms (in terms of migraine attacks, migraine days, and headache days), improve patients' autonomy, and reduce the total medication intake, its non-invasive nature needs to be put in the context of paucity of knowledge about its mechanism of action and the lacking long-term safety profile. With regards to short-term safety, no serious adverse events occurred in any of the studies. Furthermore, the potential cost-effectiveness of e-TNS needs to be contrasted with the small effects measured by the VAS, the question of the clinical importance of these effects, and the small sample size included in the studies. For its establishment in the standard practice, high quality comparative data, studies with larger sample sizes, studies with standard primary outcome parameters, patient relevant outcome parameters, and precise reporting is needed.

      Funding

      This work was supported by Ludwig Boltzmann Institute for Health Technology Assessment , Vienna, Austria. MS and SW are employees of the institute, while HJ and KZ were contracted for this study at the time of writing. The Ludwig Boltzmann Institute for Health Technology Assessment did not interfere with study design, collection, analysis and interpretation of data, writing of the report, and the decision to submit the article for publication.

      Declaration of Competing Interest

      None.

      Acknowledgement

      We are most thankful to Tarquin Mittermayr for his excellent help with the literature search.

      Appendix A. Supplementary data

      References

        • Headache Classification Committee of the International Headache Society
        The international classification of headache disorders 3rd edition.
        Cephalalgia. 2018; 38: 1-211
        • Steiner T.J.
        • Stovner L.J.
        • Vos T.
        GBD 2015: migraine is the third cause of disability in under 50s.
        J. Headache Pain. 2016; 17: 104
        • Lipton R.B.
        • Silberstein S.D.
        Episodic and chronic migraine headache: breaking down barriers to optimal treatment and prevention.
        Headache. 2015; 55: 103-122
        • Serrano D.
        • Lipton R.B.
        • Scher A.I.
        • Reed M.L.
        • Stewart W.B.F.
        • Adams A.M.
        • et al.
        Fluctuations in episodic and chronic migraine status over the course of 1 year: implications for diagnosis, treatment and clinical trial design.
        J. Headache Pain. 2017; 18: 101
        • Diener H.C.
        • Gaul C.
        • Kropp P.
        Leitlinie der Deutschen Gesellschaft für Neurologie (DGN) in Zusammenarbeit mit der Deutschen Migräne- und Kopfschmerzgesellschaft (DMKG).
        Deutsche Gesellschaft für Neurologie, 2018 (Available from:)
        • Riederer F.
        • Penning S.
        • Schoenen J.
        Transcutaneous supraorbital nerve stimulation (t-SNS) with the Cefaly((R)) device for migraine prevention: a review of the available data.
        Pain Ther. 2015; 14 (Epub 2015/10/16): 14
        • Anonymous
        A transcutaneous electrical nerve stimulation device (Cefaly) for migraine prevention.
        Med. Lett. Drugs Ther. 2014; 56: 78
        • Goadsby P.J.
        • Holland P.R.
        • Martins-Oliveira M.
        • Hoffmann J.
        • Schankin C.
        • Akerman S.
        Pathophysiology of migraine: a disorder of sensory processing.
        Physiol. Rev. 2017; 97: 553-622
        • Didier H.A.
        • Di Fiore P.
        • Marchetti C.
        • Tullo V.
        • Frediani F.
        • Arlotti M.
        • et al.
        Electromyography data in chronic migraine patients by using neurostimulation with the Cefaly device.
        Neurol. Sci. 2015; 36: 115-119
        • Magis D.
        • D’Ostilio K.
        • Thibaut A.
        • De Pasqua V.
        • Gerard P.
        • Hustinx R.
        • et al.
        Cerebral metabolism before and after external trigeminal nerve stimulation in episodic migraine.
        Cephalalgia. 2017; 37: 881-891
        • Vecchio E.
        • Gentile E.
        • Franco G.
        • Ricci K.
        • de Tommaso M.
        Effects of external trigeminal nerve stimulation (eTNS) on laser evoked cortical potentials (LEP): a pilot study in migraine patients and controls.
        Cephalalgia. 2018; 38 (333102417728748): 1245-1256
        • Institute of Health Economics (IHE)
        Quality Appraisal of Case Series Studies Checklist.
        Institute of Health Economics, Edmonton (AB)2014 ([2018/01/10]; Available from:)
        • Higgins J.
        • Altman D.G.
        • Gøtzsche P.C.
        • Jüni P.
        • Moher D.
        • Oxman A.D.
        • et al.
        The cochrane collaboration’s tool for assessing risk of bias in randomised trials.
        BMJ. 2011; 343 (d5928)
        • Guyatt G.
        • Oxman A.D.
        • Akl E.A.
        • Kunz R.
        • Vist G.
        • JEA B.
        GRADE guidelines: 1. Introduction-GRADE evidence profiles and summary of findings tables.
        J. Clin. Epidemiol. 2011; 64: 383-394
        • Schoenen J.
        • Vandersmissen B.
        • Jeangette S.
        • Herroelen L.
        • Vandenheede M.
        • Gerard P.
        • et al.
        Migraine prevention with a supraorbital transcutaneous stimulator: a randomized controlled trial.
        Neurology. 2013; 80: 697-704
        • Chou E.D.
        Acute Treatment of Migraine With e-TNS (ACME).
        • Chou D.E.
        • Shnayderman Y.M.
        • Winegarner D.
        • Rowe V.
        • Kuruvilla D.
        • Schoenen J.
        Acute migraine therapy with external trigeminal neurostimulation (ACME): a randomized controlled trial.
        Cephalalgia. 2018; 0: 1-12
        • Mann J.
        Abortive Treatment of Migraine With the Cefaly® Abortive Program Device.
        ([17.07.2018]; Available from:)
        • Di Fiore P.
        • Bussone G.
        • Galli A.
        • Didier H.
        • Peccarisi C.
        • D’Amico D.
        • et al.
        Transcutaneous supraorbital neurostimulation for the prevention of chronic migraine: a prospective, open-label preliminary trial.
        Neurol. Sci. 2017; 38: 201-206
        • Russo A.
        • Tessitore A.
        • Conte F.
        • Marcuccio L.
        • Giordano A.
        • Tedeschi G.
        Transcutaneous supraorbital neurostimulation in “de novo” patients with migraine without aura: the first Italian experience.
        J. Headache Pain. 2015; 16: 69
        • Vikelis M.
        • Dermitzakis E.V.
        • Spingos K.C.
        • Vasiliadis G.G.
        • Vlachos G.S.
        • Kararizou E.
        Clinical experience with transcutaneous supraorbital nerve stimulation in patients with refractory migraine or with migraine and intolerance to topiramate: a prospective exploratory clinical study.
        BMC Neurol. 2017; 17: 97
        • Chou D.E.
        • Gross G.J.
        • Casadei C.H.
        • Yugrakh M.S.
        External trigeminal nerve stimulation for the acute treatment of migraine: open-label trial on safety and efficacy.
        Neuromodulation. 2017; 20: 678-683
        • Olsen M.F.
        • Bjerre E.
        • Damkjaer Hanse M.
        • Hilden J.
        • Landler N.E.
        • Tendal B.
        • et al.
        Pain relief that matters to patients: systematic review of empirical studies assessing the minimum clinically important difference in acute pain.
        BMC Med. 2017; 15
        • Magis D.
        • Sava S.
        • d’Elia T.S.
        • Baschi R.
        • Schoenen J.
        Safety and patients’ satisfaction of transcutaneous supraorbital neurostimulation (tSNS) with the Cefaly device in headache treatment: a survey of 2,313 headache sufferers in the general population.
        J. Headache Pain. 2013; 14: 95
        • Brown J.S.
        • Papadopoulos G.
        • Neumann P.J.
        • Price M.
        • Friedman M.
        • Menzin J.
        Cost-effectiveness of migraine prevention: the case of topiramate in the UK.
        Cephalalgia. 2006; 26: 1473-1482
        • Ekizoglu E.
        • Sozer-Topçular N.
        • Baykan B.
        • Oge A.E.
        Assessment of excitability at the brainstem and cortex in primary headaches with allodynia.
        J. Clin. Neurophysiol. 2015; 32: 119-129
        • Przeklasa-Muszynska A.
        • Skrzypiec K.
        • Kocot-Kepska M.
        • Dobrogowski J.
        • Wiatr M.
        • Mika J.
        Non-invasive transcutaneous Supraorbital Neurostimulation (tSNS) using Cefaly((R)) device in prevention of primary headaches.
        Neurol. Neurochir. Pol. 2017; 51 (Epub 2017/02/06): 127-134
        • Jiang L.
        • Yuan D.L.
        • Li M.
        • Liu C.
        • Liu Q.
        • Zhang Y.
        • et al.
        Combination of flunarizine and transcutaneous supraorbital neurostimulation improves migraine prophylaxis.
        Acta Neurol. Scand. 2019; 139: 276-283