Neurostimulation in the treatmentof primary headaches

Sarah Miller,1 Alex J Sinclair,2 Brendan Davies,3 Manjit Matharu1

▸ Additional material ispublished online only. To viewplease visit the journal online(http://dx.doi.org/10.1136/practneurol-2015-001298).

1Headache Group, Institute ofNeurology and The NationalHospital for Neurology andNeurosurgery, London, UK2Neurometabolism, Institute ofMetabolism and SystemsResearch, College of Medicaland Dental Sciences, TheUniversity of Birmingham,Birmingham, UK3Department of Neurology, RoyalStoke University Hospital, Stoke-on-Trent, UK

Correspondence toDr Alex J Sinclair,Neurometabolism, Institute ofMetabolism and SystemsResearch, College of Medicaland Dental Sciences, Universityof Birmingham, Wolfson Drive,Edgbaston B15 2TT, UK;a.b.sinclair@bham.ac.uk

Accepted 11 April 2016Published Online First5 May 2016

To cite: Miller S, Sinclair AJ,Davies B, et al. Pract Neurol2016;16:362–375.

ABSTRACTThere is increasing interest in usingneurostimulation to treat headache disorders.There are now several non-invasive and invasivestimulation devices available with some open-label series and small controlled trial studies thatsupport their use. Non-invasive stimulationoptions include supraorbital stimulation (Cefaly),vagus nerve stimulation (gammaCore) andsingle-pulse transcranial magnetic stimulation(SpringTMS). Invasive procedures includeoccipital nerve stimulation, sphenopalatineganglion stimulation and ventral tegmental areadeep brain stimulation. These stimulation devicesmay find a place in the treatment pathway ofheadache disorders. Here, we explore the basicprinciples of neurostimulation for headache andoverview the available methods ofneurostimulation.

BACKGROUNDHeadache conditions can be classifiedinto episodic and chronic forms (box 1).Chronic headache is a major globalhealth issue affecting up to 4% of thepopulation.1 Most patients with chronicheadache attending neurology centreshave chronic migraine or cluster head-ache. The estimated prevalence ofchronic migraine is 2% and of chroniccluster is 0.02%.2 Although most patientsare helped with medical treatments, a sig-nificant minority cannot tolerate or proverefractory to pharmacological treatments.Neurostimulation therapies with periph-eral or central targets may help thesepatients. Here, we provide an overviewof this emerging field of headachemanagement.

NEUROSTIMULATION FOR PAINElectrical stimulation to treat pain is nota new concept. Scribonius Largus, ca. CE46, advocated fish ‘electrotherapy’ totreat headache.3 Modern neurostimula-tion for pain has its roots in functionalneurosurgery, which initially relied ondestructive lesions before focusing on

neurostimulation to manage intractablepain. Neurostimulation works bymanipulating central or peripheral painpathways using electrical or magneticimpulses. It aims is modify the painsystem in such a way to reduce painlevels.There are multiple targets for neurosti-

mulation in headache, including the pos-terior hypothalamus/ventral tegmentalarea, sphenopalatine ganglion (SPG),occipital nerve, vagus nerve, supraorbitalnerve and cortex. These have beenselected because of the recent recognitionof anatomically relevant pathophysio-logical pathways and hypothesisedmechanisms in both animal and humanmodels of migraine and cluster headache.Applying electrical or magnetic stimula-tion to pain pathways can modify centralneurotransmitters.4 For preventive treat-ment, these changes aim to cause a winddown of the central sensitisation thatoccurs in chronic headache. For acutetreatment, these changes probably blockthe processes responsible for attack gen-eration (cortical spreading depression orbrainstem activation). Figure 1 sum-marises the pathways most likely involvedin the mechanism of action of these neu-rostimulation treatments.Currently available non-invasive

devices target the supraorbital and vagusnerves (electrical stimulation) or thecortex (magnetic stimulation). Availableinvasive stimulation techniques involvethe occipital nerves and SPG (peripheraltargets) and the ventral tegmental area(central target). Such techniques mayhelp those patients who wish to avoid,are refractory to or intolerant of previousdrug therapies. Devices that allow acutetreatment of attacks may be help thosewho cannot use triptans or in whomacute medications are ineffective or over-used. Pregnancy—another situationwhere it may be difficult to use preventa-tive and acute headache treatments—may,

REVIEW

362 Miller S, et al. Pract Neurol 2016;16:362–375. doi:10.1136/practneurol-2015-001298

on June 8, 2020 by guest. Protected by copyright.

http://pn.bmj.com

/P

ract Neurol: first published as 10.1136/practneurol-2015-001298 on 5 M

ay 2016. Dow

nloaded from

in theory, be a specialised situation where neurostimu-lation devices could be used. However, the safety ofsuch devices in pregnancy is not established.Laboratory data suggest that exposure to electricalstimulation techniques is safe in animal studies and

limited open-label studies suggest safety in occipitalnerve stimulation and transcranial magneticstimulation.

NON-INVASIVE NEUROSTIMULATION DEVICESSupraorbital nerve stimulation and the Cefaly deviceThe trigeminovascular system has a crucial role inhead pain.5 The supraorbital nerve is a branch of thefirst division of the trigeminal nerve. Transcutaneoussupraorbital nerve stimulation has been developed as apotential treatment for headache. The Cefaly device isan external transcutaneous supraorbital nerve stimula-tor that is battery powered and applied to the fore-head using a headband-like device (figure 2).Currently, the device is not available on the National

Figure 1 Headache pain pathways targeted by neurostimulation. A simplified diagram of the various peripheral and central painpathways targeted by current neurostimulation devices. LC, locus coeruleus; NRM, nucleus raphe magnus; PAG, periaqueductal grey;SPG, sphenopalatine ganglion; SSN, superior salivary nucleus; TNC, trigeminal nucleus caudalis; V1, first division of the trigeminalnerve.

Box 1 Classification of episodic and chronicheadache

Episodic headacheHeadache occurs on <15 days a month for >3 months.Chronic headacheHeadache occurs on >15 days a month for >3 months.

REVIEW

Miller S, et al. Pract Neurol 2016;16:362–375. doi:10.1136/practneurol-2015-001298 363

on June 8, 2020 by guest. Protected by copyright.

http://pn.bmj.com

/P

ract Neurol: first published as 10.1136/practneurol-2015-001298 on 5 M

ay 2016. Dow

nloaded from

Health Service (NHS) but can be purchased directlyfrom the manufacturer. The cost is £249 for thedevice and around £50 for a 3-month supply ofreusable electrodes.

Possible role for the Cefaly deviceAcute treatment of episodic migraine

There is no well-controlled evidence for using theCefaly device in the acute treatment of episodic

migraine. A single pilot study of 10 patients using thedevice for acute migraine treatment showed no effecton 57% of treated attacks and was associated withpain freedom in only 13% of attacks.6

Preventative treatment of episodic migraine

Evidence for the use of the Cefaly device in prevent-ing episodic migraine comes from one small,manufacturer-sponsored sham-controlled trial, thePREvention of MIgraine using the STS Cefaly (thePREMICE study) and company postmarketing surveydata. The sham-controlled study involved only 67 sub-jects with episodic migraine, who after a one monthrun-in period of normal treatment used the Cefaly orsham device for 3 months and reported a significantreduction in migraine days by 29.7% from 6.94 to4.88 days (p=0.023) in the active group compared toa non-significant change of 4.9% from 6.54 to 6.22days (p=0.608) in the sham group.7 For comparison,topiramate 100 mg daily decreased migraine days by44% in a pooled analysis of controlled studies.8 Thetherapeutic gain for the 50% responder rate of supra-orbital nerve stimulation was 26.1% compared with23.5% for topiramate but the therapeutic gain inmigraine day reduction is much higher in pooledstudies of topiramate at 24.5% compared with thePREMICE trial at 12%.8 The responder rate shown inthe PREMICE study was within the range quoted forother migraine preventatives such as propranolol andantiepileptic medications.9 10 The postmarketingsurvey of 2313 subjects using the device to preventepisodic migraine reported 53% of users were ‘satis-fied’ with the treatment, as determined by the numbercontinuing treatment after a 40-day trial period.11 Inthe PREMICE trial, 70.6% declared themselves satis-fied after 3 months of treatment.7 In contrast, usingstopping of treatment as a measure of patient satisfac-tion, one American health insurance survey ofmigraine preventatives found that 26.6% of patientson antidepressants, 29.8% on antiepileptic medica-tions and 32.4% on beta-blockers were satisfied andremained on treatment after 6 months.

Using the Cefaly deviceBox 2 outlines a preventative regimen involving presetstimulation sessions to be used daily. As with allmigraine preventatives, treatment must continue for3 months before assessing efficacy, so encouragingtreatment adherence is important. The manufacturershave a preprogrammed 20 min session ‘acute setting’for terminating a migraine attack, despite there beingno evidence to support its use in acute migraine.The adverse effects of the Cefaly device appear

mild and transient.11 Some patients experienceintolerance to paraesthesia, drowsiness, worsening ofheadache and reversible forehead irritation. Table 2lists its contraindications.

Figure 2 Non-invasive supraorbital nerve stimulation. (A) and(B) show how a patient would use the Cefaly device to deliverpreventative treatment for migraine. Image published withpermission of individual.

Box 2 Using the Cefaly device to prevent episodicmigraine

Place sticky electrodes on the forehead and attach thedevice to the electrode (figure 2).For preventative treatment of episodic migraine:To deliver the preset preventative mode, press the deviceon-switch twice to deliver a 20 min session.Patients should receive daily preventative sessions for atleast 3 months before assessing efficacy.

REVIEW

364 Miller S, et al. Pract Neurol 2016;16:362–375. doi:10.1136/practneurol-2015-001298

on June 8, 2020 by guest. Protected by copyright.

http://pn.bmj.com

/P

ract Neurol: first published as 10.1136/practneurol-2015-001298 on 5 M

ay 2016. Dow

nloaded from

Proposed mechanism of actionWe do not yet know the mode of action of supraorbitalnerve stimulation in migraine. In healthy volunteers,stimulation with supraorbital nerve stimulation has anacute sedative effect.12 Although it is questionablewhether this influences the preventive treatment ofmigraine, it does suggest that supraorbital nerve stimu-lation can change brain activity. Current understandingof migraine pathophysiology suggests that there is neur-onal hyperexcitability of the pain neuromatrix: trigemi-novascular activation appears to be the basis for theheadache phase and central sensitisation the mechan-ism that may contribute to the transformation from epi-sodic to chronic forms. Supraorbital nerve stimulationprobably winds down the trigeminal pain pathwaysthrough altering activity within the trigeminovascularsystem both peripherally and centrally.

Vagus nerve simulation and the gammaCore deviceThe vagus nerve is a mixed motor and sensory nervethat is important in controlling autonomic responses;it projects to several higher centres that are importantin pain regulation. Following reports of migraineimprovement in patients receiving vagus nerve stimu-lation for epilepsy, the nerve became a target for head-ache treatment. The gammaCore device is a handheldtranscutaneous vagus nerve stimulator applied to theneck (figure 3). Devices are available charged witheither 150 or 300 treatment cycles, after which a newdevice must be purchased. If used for prevention only,the 300-treatment device lasts around 50 days; thus, apatient needs two devices to have the efficacy assessedat 3 months. Currently, the device is not available onthe NHS but individual funding requests may be con-sidered. If purchased direct from the manufacturer,the 300-treatment device costs around £550.

Evidence for the gammaCore devicePreventative treatment of cluster headache

The current evidence for its use in preventing clusterheadache attacks is limited to a manufacturer-sponsored

trial involving 97 subjects. This trial of standard of careplus vagus nerve stimulation versus standard of carealone was conducted on the preventative and acutetreatment of chronic cluster headache using thegammaCore device. Regular use of gammaCore for4 weeks was associated with a significantly reducedattack frequency in the active compared with standardof care group (7 vs 2 fewer attacks per week).13

Similarly, more subjects in the active group reported>50% fewer weekly attacks compared with those onstandard treatment (52% vs 9%).14 For comparison, theresponder rate for verapamil 360 mg daily was 80%versus 0% for placebo in the only available smallrandomised-controlled trial.14

Acute treatment of cluster headache

The above randomised study found no effect of acutegammaCore treatment on headache intensity.13

However, in a small open-label study of 19 patientsusing the device for acute treatment, 47% of attacksended within 11 min (average duration beforegammaCore use was 75 min).15 All 13 patients usingtriptans before starting using gammaCore treatmentcould reduce their intake, and four patients stoppedtriptans completely. The standard acute treatments forcluster headache attacks are subcutaneous sumatriptanand high-flow oxygen. In one randomised double-blind study of 39 patients, subcutaneous sumatriptanaborted 74% of attacks within 15 min compared with26% of those treated with placebo.16 In another ran-domised control study of 76 patients, high-flowoxygen aborted 78% of attacks within 15 min,whereas placebo aborted only 20% of attacks.17

Preventative treatment of chronic migraine

There are no data to support the use of gammaCoreto prevent chronic migraine. A single controlled trialof 59 patients comparing active gammaCore to shamtreatment showed that, after 2 months, active treat-ment was not associated with a reduction of headachedays (−1.9 days with gammaCore vs 0.20 sham(p=0.124)).18 Although open-label extension datasuggest that a longer duration of treatment might beeffective, we need more controlled studies to investi-gate this.

Acute treatment of migraine

There have been no controlled studies in acutemigraine. Two small open-label studies examinedgammaCore in treating acute migraine pain. The first,involving 27 patients treating 80 attacks, used twostimulation cycles administered 15 min apart duringan attack.19 Twenty-two per cent ofmoderate-to-severe attacks achieved pain freedomwithin 2 h of treatment, a benefit similar to that ofnaproxen 500 mg. Another study of 24 patients with79 attacks showed significantly reduced pain scoresfrom baseline at 2 h following treatment, with a 50%reduction in 46% at 2 h, increasing to 61% at 4 h

Figure 3 Non-invasive vagus nerve stimulation. How a patientwould use the gammaCore device to deliver acute orpreventative treatment for cluster headache. Image publishedwith permission of individual.

REVIEW

Miller S, et al. Pract Neurol 2016;16:362–375. doi:10.1136/practneurol-2015-001298 365

on June 8, 2020 by guest. Protected by copyright.

http://pn.bmj.com

/P

ract Neurol: first published as 10.1136/practneurol-2015-001298 on 5 M

ay 2016. Dow

nloaded from

(report available only in poster form). In arandomised-controlled study of acute migraine treat-ment in 92 patients, 67% achieved a pain-free rateusing 100 mg sumatriptan (placebo rate 28%).20 21

Possible role of the gammaCore deviceCurrent evidence suggests that the gammaCore deviceshould only be considered to prevent cluster head-ache. There is no robust evidence for using it as anacute treatment (table 1). Nonetheless, it might beconsidered as an acute treatment for patients withmore than two cluster attacks daily (ie, more attacksthan can safely be treated with subcutaneous suma-triptan), in those with contraindications to triptansand in those where all acute medications for clusterheadache are ineffective. There is no evidence as yetfor using gammaCore in migraine.

Using the gammaCore deviceBox 3 outlines our recommended regimens for usinggammaCore. The activated device delivers a singleprogramme cycle of 2 min. For preventative treatmentof cluster headache, we recommend daily use ofgammaCore for 3 months before assessing efficacy. Ifthe patient wishes to use it for acute treatment ofcluster headache, we recommend that they use it fortwo programme cycles at attack onset. The reportedadverse effects are mild and include transient hoarse-ness, voice change, skin irritation, muscle ache anduncomfortable paraesthesia. Table 2 lists itscontraindications.

Proposed mechanism of actionThe vagus nerve has several connections to higherbrain centres that are important in pain regulation,such as the nucleus tractus solitarius and spinal

Box 3 Using the gammaCore device to treatcluster headache

The patient locates their carotid pulse in the neck andplaces the device over this.The device can be used on either side of the neck.The stimulation is adjusted by turning a thumbwheeluntil a deep vibration is felt inside the neck. The patientshould aim to use the maximum tolerable intensity(stimulation setting three or more in most people).When correctly positioned, the subject should feel apulling at the corner of the mouth. This is a normalresponse and patients should be encouraged to look outfor it.For acute treatment of cluster headache attacks (evidencebase poor):Deliver two programme cycles (a total of 4 min) at theattack onset. Repeat after 15 min if needed.For preventative treatment of cluster headache attacks:Deliver one programme cycle (total of 2 min) three timesa day.The regimen can be increased to two cycles (total of4 min) three times daily if required.Preventative sessions should be delivered daily for3 months before assessing efficacy.

Table 1 Neurostimulation and devices and the conditions in which they can be used

Supraorbitalstimulation

Vagus nervestimulation

Transcranialmagneticstimulation

Occipital nervestimulation

Sphenopalatineganglion stimulation

Deep brainstimulation

Device name Cefaly gammaCore SpringTMS Pulsante

Acute migraineattacks

X ✓ ✓(with and withoutaura)

X X X

Prevention ofepisodic migraine

✓ X X X X X

Prevention ofchronic migraine

X X X ✓ X X

Acute clusterattacks

X ✓ X X ✓ X

Prevention ofepisodic clusterheadache

X ✓* X X X X

Prevention ofchronic clusterheadache

X ✓ X ✓ ✓(studies ongoing)

✓

Other TACs X X X ✓(prevention ofchronic intractableTACs)

X ✓(prevention ofchronic intractableTACs)

*Usefulness may be dictated by length of episodic cluster bout; if bout lasts <3 months, it may be difficult to assess as treatment may take this long tohave a clear effect.TAC, trigeminal autonomic cephalalgia.

REVIEW

366 Miller S, et al. Pract Neurol 2016;16:362–375. doi:10.1136/practneurol-2015-001298

on June 8, 2020 by guest. Protected by copyright.

http://pn.bmj.com

/P

ract Neurol: first published as 10.1136/practneurol-2015-001298 on 5 M

ay 2016. Dow

nloaded from

trigeminal nucleus. Early studies suggested that theacute effect of vagus nerve stimulation is mediated bydirect inhibition of afferents to the caudal trigeminalnucleus.22 More recently, neuroimaging studies haveshown that chronic vagus nerve stimulation inhibitsactivation of the thalamus, limbic system, dorsal pons,locus coeruleus and nucleus tractus solitarius, all ofwhich are structures identified in imaging studies aspart of the pain matrix of headache.23 There is alsoevidence that it may inhibit pain by reducing the con-centration of glutamate in the trigeminal nucleus cau-dalis; this in turn may reverse central sensitisation inchronic headache.24

Transcranial magnetic stimulation and the SpringTMSdeviceTranscranial magnetic stimulation applies a brief singlemagnetic pulse to the scalp and underlying cortex.This pulse induces electrical fields in the cortex, alter-ing neurotransmitter release and disrupting corticalspreading depression. The SpringTMS device is arechargeable handheld device that delivers a singlepulse of magnetic stimulation to the back of the head.As with other non-invasive devices, it is not widelyavailable through NHS sources but individual fundingrequests can be made (with special arrangements forclinical governance, audit and outcome data collec-tion, as per National Institute for Health and CareExcellence recommendations).25 In the private sector,it is available from the manufacturer for a free3-month trial after which the patient must pay £158every month for continuing use.

Evidence for SpringTMSAcute treatment of migraine with and without aura

The evidence for acute treatment of migraine comesfrom only one small manufacturer-sponsored studyand postmarketing surveys. A sham-controlled studyon SpringTMS in the acute treatment of migrainewith aura in 164 patients reported a pain-freeresponse rate higher in those with active comparedwith sham treatment at both 2 h (39% vs 22%) and

24 h (29% vs 16%) for a therapeutic gain of 17%(p=0.0179).26 As above, in a previous trial of acutemigraine treatment, the pain-free rate was 67% for100 mg sumatriptan with a placebo rate of 28%.20

In an open-label postmarketing survey of 462 sub-jects in the UK using SpringTMS as acute treatmentfor migraine with and without aura, authors obtained3-month follow-up data on only 190 patients.27 Thediscontinuation rate within the 3-month follow-upperiod was 55% (n=105) mainly due to inadequatebenefit (n=49), cost (n=−17) or inconvenience(n=15). Of the 190 subjects who completed 3 monthsfollow-up, 62% reported some reduction in migrainepain and 59% some reduction in attack duration. Thegroup did not provide further clarification withresponder rates and so we cannot make meaningfulcomparisons to other acute treatments for migraine.

Preventative treatment of migraine

There is no controlled evidence to support the use ofSpringTMS in the prevention of migraine.

Possible uses of SpringTMSFrom the limited evidence available, SpringTMSmight have a role in the acute treatment of migrainewith and without aura in those at risk of overusingacute medications or if acute pharmacological treat-ments are ineffective (table 1). As with the otherdevices, there are no clear safety data in children or inpregnancy; however, three women among the post-marketing data completed normal pregnancies.27

Current evidence does not support the use ofSpringTMS in the preventative treatment of migraineor for the treatment of cluster headache.

Using SpringTMSBox 4 gives our recommended regimens forSpringTMS in acute migraine attacks. The device isheld against the occiput and the pulse delivered withthe press of a button (figure 4). The device should beused as early as possible into a migraine attack.Following review, the clinician can adapt the numberof pulses delivered if needed. Patients should use the

Table 2 Contraindications for non-invasive neurostimulation

Supraorbital stimulation Vagus nerve stimulation Transcranial magnetic stimulation

Device name Cefaly gammaCore SpringTMS

Contraindicatedfor use

Recent brain or facial trauma(<3 months)

Skin abrasion on the forehead inthe area of application of theelectrodeAllergy to acrylate

Active implantable medical device, such as apacemaker, defibrillator, cochlear implant andother implanted electronic deviceMetallic implants near the treatment siteHistory of significant carotid atherosclerosisCervical vagotomy

Active implantable medical device, such as apacemaker, defibrillator, cochlear implant andother implanted electronic deviceImplants affected by a magnetic field†*

Epilepsy

Cautious use Electrohypersensitivity Vasovagal syncopeSkin irritation near the treatment site

Cardiac conditions

Devices should not be used while driving or operating machinery.*Dental implant and fillings are not a contraindication to using transcranial magnetic stimulation.†Including aneurysm clips or coils, cerebrospinal fluid shunts, bullets or pellets lodged in the head or upper body, metal plates, screws, staples or suturesin skull, neck, shoulders, arms or hands, electrodes, radioactive seeds, stents, filters, metallic heart valves, facial tattoos with metallic ink.

REVIEW

Miller S, et al. Pract Neurol 2016;16:362–375. doi:10.1136/practneurol-2015-001298 367

on June 8, 2020 by guest. Protected by copyright.

http://pn.bmj.com

/P

ract Neurol: first published as 10.1136/practneurol-2015-001298 on 5 M

ay 2016. Dow

nloaded from

device for at least 3 months to assess efficacy as thereis some evidence of an increased benefit at 12 weekscompared with 6 weeks.27 Reported side effects aretransient and mild; these include dizziness, light-headedness, tingling and worsening of migraine pain.Table 2 lists its contraindications.

Proposed mechanism of actionPatients with migraine are thought to have a state ofbrain hyperexcitability, and this has been shown intranscranial magnetic stimulation studies.28 Thishyperexcitable cortex leads to a lowered threshold forcortical spreading depression, a wave of depolarisationof neural membranes, which has been linked to thegeneration of migraine aura and activation of menin-geal nociceptors directly or via descending facilitationpathways.29 In animal studies, single-pulse transcranialmagnetic stimulation inhibits cortical spreadingdepression. Thus, in acute migraine, single-pulse tran-scranial magnetic stimulation has the potential to ter-minate aura and reduce the severity of headache inthose with aura. Repetitive transcranial magneticstimulation reduces cortical hyperexcitability bymodulating concentrations of neurotransmitters suchas dopamine and glutamate in the caudate and hippo-campus.28 Thus, repetitive transcranial magneticstimulation may produce long-term changes in neur-onal excitability, reversing central sensitisation andreducing headache frequency. However, much of thework with repetitive transcranial magnetic stimulationhas been based on laboratory studies and it has notyet translated into a useful clinical treatment effect.

INVASIVE NEUROSTIMULATION DEVICESOccipital nerve stimulationThe occipital nerves are a target for stimulation dueto the anatomical overlap between the trigeminal andcervical afferents in the trigeminocervical complex.This allows stimulation of the occipital region tomodulate pain in the trigeminal distribution. Occipitalnerve stimulation is a non-destructive surgical proced-ure where electrodes are placed subcutaneously in theoccipital region and then wired to a battery pack inthe chest or abdomen (figure 5). Stimulation para-meters can be adjusted to control paraesthesia.Patients can adjust their device and recharge thebattery using a handheld remote control. Occipitalnerve stimulation is not widely available and shouldonly be performed in highly specialised services withestablished surgical and headache teams.

Evidence for occipital nerve stimulationOpen-label studies have shown possible efficacy inpreventing chronic migraine, chronic cluster headache,hemicrania continua and short-lasting unilateral neur-algiform headache attacks.30 The average responserate is around 70% for chronic cluster headache and56% for chronic migraine.

Preventative treatment of refractory chronic migraine

Controlled trials using occipital nerve stimulation inchronic migraine have given mixed results. TheOccipital Nerve Stimulation for the Treatment ofChronic Migraine study was a randomised-controlledstudy of 61 subjects comparing adjustable (28

Figure 4 Transcranial magnetic stimulator. (A) and (B) showhow a patient would use the SpringTMS device to deliver apulse to the occipital region in order to treat acute migrainewith aura attacks. Image published with permission ofindividual.

Box 4 Using SpringTMS for the acute treatment ofmigraine with and without aura

Use the device at onset of migraine symptoms of painand/or aura.The device is held against the occiput and a buttonpressed to initiate the pulse.For acute treatment of migraine:Deliver two sequential pulses as early as possible intothe attack.Continue delivering two pulses every 15 min for 2 h oruntil attack resolves.After first month, reassess and, if needed, increase tothree pulses with each treatment.After second month, reassess and, if needed, increase tofour pulses delivered with each treatment.

REVIEW

368 Miller S, et al. Pract Neurol 2016;16:362–375. doi:10.1136/practneurol-2015-001298

on June 8, 2020 by guest. Protected by copyright.

http://pn.bmj.com

/P

ract Neurol: first published as 10.1136/practneurol-2015-001298 on 5 M

ay 2016. Dow

nloaded from

patients) and preset stimulation (16 patients) andmedical management (17 patients).31 A positiveresponse (a 50% reduction in monthly headache daysor a >3-point reduction in pain scores) was seen in39% of the adjustable group, 6% in the preset groupand 0% in the medical group. Another randomisedstudy of 125 subjects comparing active with shamstimulation showed no significant difference inmigraine day reduction between the groups after 3months.32 The third randomised trial of occipital nervestimulation for chronic migraine looked at 157patients. The proportion of subjects achieving a >50%reduction in daily pain scores was not significantly dif-ferent in the active and sham groups (17% vs 14%).However, the reduction of headache days was signifi-cantly higher in the active group (27% vs 15%) as wasthe proportion of patients achieving at least a 30%reduction in pain scores (38% vs 19%). Pain specialistsnow widely accept that a 30% improvement in chronicpain represents a clinically meaningful improvement;thus, this study outcome favoured the use of occipitalnerve stimulation.33 A 2015 systemic review andmeta-analysis of the effectiveness of occipital nervestimulation found that in a pooled analysis of the threeabove controlled trials, occipital nerve stimulation wasassociated with a mean reduction of 2.59 migraine daysper month after 3 months compared with sham con-trols.34 However, analysis on other outcome measureswas hampered by incomplete publications and poorreporting of data. Open-label follow-up data gavelimited evidence of long-term efficacy.

Preventative treatment of chronic cluster headache

There are no controlled trials of occipital nerve stimu-lation in chronic cluster headache. However, positive

open-label studies support its use (table 3). Onerecent review paper conducted a pooled analysis ofthe nine published studies involving 91 patients andfound that patients reported an average 67% reduc-tion of attack frequency.30

Possible uses of occipital nerve stimulationOccipital nerve stimulation, as with all invasive neuro-stimulation devices, must be reserved for those withhighly medically refractory headaches who have notresponded to all other treatments (table 1). TheEuropean Headache Society has guidelines on the useof invasive neurostimulation and patient selection (seeonline supplementary data table).41

Occipital nerve stimulation can be considered forpreventative treatment of refractory chronic migraineor cluster headaches. Patients must be assessed by aspecialist team before consideration of surgery andmust be deemed significantly disabled by their head-aches. Following the implant, patients cannot haveMRI scans and so occipital nerve stimulation shouldbe avoided in those with coexisting conditions thatmay require future MRI scanning (ie, multiple scler-osis). Occipital nerve stimulation has no role in theacute treatment of migraine or cluster headacheattacks.

Using occipital nerve stimulationBilateral leads should be implanted even with unilat-eral headaches as pain often swaps sides with unilat-eral stimulation. Following implantation, the device isprogrammed to give a comfortable level of paraesthe-sia in the distribution of the greater occipital nerve.The device is left on at all times, and generallypatients are advised not to alter the settings unless the

Figure 5 Skull X-rays showing occipital nerve stimulation electrode placement. Lateral (A) and anterior posterior (B) skull viewsshowing the occipital nerve stimulation electrodes (dotted lines) and leads (solid lines).

REVIEW

Miller S, et al. Pract Neurol 2016;16:362–375. doi:10.1136/practneurol-2015-001298 369

on June 8, 2020 by guest. Protected by copyright.

http://pn.bmj.com

/P

ract Neurol: first published as 10.1136/practneurol-2015-001298 on 5 M

ay 2016. Dow

nloaded from

paraesthesia becomes painful or unnoticeable.Stimulation is not to be altered during acute attacks.Current rechargeable batteries should last around 5–7 years, and patients are advised to charge them atleast once weekly to prevent system failure. It can takeup to 3 months post-programming to detect anychange in headache severity or frequency. It is import-ant that patients are followed up frequently in theearly postoperative period to ensure optimum stimula-tion settings are used. If after 1 year of continuouscomfortable stimulation there has been no change,then it is highly unlikely that the patient will gain anybenefit. In these patients, we switch off the device forat least 3 months before offering removal to ensurethey have not failed to recognise the extent of anyimprovement.Adverse event rates vary between centres but are

reduced if the procedure is carried out in specialistcentres.42 Hardware-related adverse events that mayrequire additional surgical input include lead migra-tion (13%), lead fracture (4%) and erosion of an elec-trode through the skin (4%). Other commonlyreported adverse events include pain over the batterysite (18%), infection (10%) and painful stimulation(17%).

Proposed mechanism of actionThe trigeminal cervical complex is a group of brain-stem and cervical spinal regions at which there is ananatomical convergence between the cervical and tri-geminal systems. The trigeminal cervical complex is akey relay system for pain from the head and facialregions to higher pain processing centres in the hypo-thalamus, thalamus and brainstem. The exact mechan-ism of action of occipital nerve stimulation is unclear,and it is likely to act via a non-specific modulatoryeffect on pain-control systems. Fluorodeoxyglucose

positron emission tomography imaging of patientswith occipital nerve stimulation for chronic clusterheadache found that metabolism normalised in severalareas of the pain matrix following treatment.36

However, the ipsilateral posterior hypothalamus,which becomes activated during acute cluster head-aches, continued to show hyperactivation. Similarfindings with persistent hyperactivity of the dorsalrostral pons, which becomes activated in migraine,occurred with H2

15O-positron emission tomography inpatients undergoing occipital nerve stimulation forchronic migraine.43 These findings support thehypothesis that occipital nerve stimulation activatesdescending pain-control systems and restores equilib-rium in antinociceptive pathways.

SPG stimulation and the Pulsante SPG microstimulatorThe SPG is an extracranial structure lying in the ptery-gopalatine fossa containing sympathetic and parasym-pathetic neurones. It has connections to thetrigeminovascular system, superior salivatory nucleusand posterior hypothalamus—all areas that have animportant role in the generation of cluster headacheattacks. The Pulsante device is a miniaturised implan-table neurostimulator with an integral lead and abattery. The lead is placed within the pterygopalatinefossa using a minimally invasive transoral approach.The patient controls the device using a handheldremote control. SPG stimulation is available on aresearch basis in two centres in the UK. The device isavailable in a limited capacity in some European sites,principally in Denmark and Germany.

Evidence base for SPG stimulationAcute treatment of chronic cluster headache

A multicentre trial of 28 patients using the Pulsantedevice reported a significant difference in the number

Table 3 Summary of main study outcomes for occipital nerve stimulation in chronic cluster headache

StudyNumber ofpatients

Meanfollow-up[range](months)

Patientsimproved>50%

Change attackfrequency(%)

Change attackseverity(%)

Preventative treatmentreduction*

Magis et al35 36 14 36.62[11–64]

12/14(86%)

−94.6 +2.3 29%

Burns et al37 38 14 17.5[4–35]

10/14(71%)

−33 +8 43% (triptans)

de Quintana-Schmidtet al39

4 6+ 4/4(100%)

−56 −48 21%

Fontaine et al40 13 14.6[3–34]

10/13(76%)

−68 −49 62%

Mueller 201356 24 21.5[4–47]

21/24(88%)

−40 −3840% reduction daily triptandose

NHNN (authors) cohort(submitted forpublication)

63 46[2–129]

36/63(57%)

−49 −25 83%(41% triptans)

*Figures show the percentage of patients on preventative drugs at baseline who have reduced or stopped medication at follow-up unless otherwise stated.NHNN, National Hospital for Neurology and Neurosurgery.

REVIEW

370 Miller S, et al. Pract Neurol 2016;16:362–375. doi:10.1136/practneurol-2015-001298

on June 8, 2020 by guest. Protected by copyright.

http://pn.bmj.com

/P

ract Neurol: first published as 10.1136/practneurol-2015-001298 on 5 M

ay 2016. Dow

nloaded from

of resolved cluster attacks within the active comparedwith sham group at 15 min (67% vs 7%).44 After2 months of treatment, 31% of the active group werestill using acute medications compared with 77% ofthe sham group.

Preventative treatment of chronic cluster headache

Although the above study was designed to examinethe acute use of SPG stimulation, 43% of subjectsusing the device to treat attacks reported a >50%reduction in the weekly cluster attack frequency after2 months.

Possible uses of SPG stimulationAn expert consensus on the use of SPG stimulation inchronic cluster headache was published in 2014 (table1; see online supplementary table).45 The procedurecould be considered as an acute, and possibly a pre-ventative, treatment for those with medically refractorychronic cluster headache who have failed all availablemedical treatments. At present, recommendations arethat only those with strictly unilateral cluster attacksshould be implanted. The device may be particularlyuseful in those patients who have not responded to orwho have contraindications to triptans and those whohave a high number of daily attacks.

Using SPG stimulationAs with the other invasive neurostimulation techni-ques, SPG stimulation should only be carried out inspecialist centres and patients must be assessed forsuitability by a multidisciplinary team prior to surgery.Immediately post implant, the patients need to bereviewed on a 1–2 weekly basis to achieve stimulationsettings, resulting in comfortable paraesthesia in thesoft palate. At the start of an attack, the patient placesthe handset to the cheek above the implant and acti-vates the device (figure 6). Stimulation should be con-tinued for at least 15 min. After this time, if the attackcontinues, the patient should switch off the deviceand use their normal rescue medication. Even ifpatients stop having regular attacks, then the devicecould be used as a preventative giving 15 min ofstimulation once or twice a day. There are ongoingstudies to determine the most effective preventativeregimens. Adverse events include misplacement ormigration of the leads (15%), infection (6%) or mildtransient sensory deficit in the maxillary division ofthe trigeminal nerve (81%).44

Proposed mechanism of actionCluster headache pathology probably involves inter-action between trigeminal inputs and the cranial para-sympathetic outflow from the superior salivatorynucleus via the SPG. Postganglionic fibres from the SPGinnervate facial structures and meningeal blood vessels.These fibres release neurotransmitters that activate tri-geminal nociceptors resulting in activation of the tri-geminal system, which, in turn, has a positive feedback

on the parasympathetic outflow. This pathway isreferred to as the trigemino-autonomic reflex.46 SPGstimulation probably works by interrupting this system,resulting in the termination of acute attacks via a directeffect on the trigeminal inputs and parasympatheticoutflow; it may prevent attacks by inducing long-termchanges in neurotransmitter release.

Deep brain stimulation (ventral tegmentum/posteriorhypothalamic region)Functional neuroimaging techniques show that theposterior hypothalamic region is activated duringcluster headache attacks.47 Stimulation of this regionincreases blood flow through areas of the pain matrix.Deep brain electrodes were first implanted in thisregion to treat a patient with cluster headache in2001.48 Further work localised the site of implant-ation to the ventral tegmental area rather than theposterior hypothalamus.49

Using stereotactic imaging guidance, an electrode isplaced into the ventral tegmental area ipsilateral tothe site of pain (figure 7). The device is programmedand kept active at all times. The method is not avail-able for treating headache on the NHS in England.

Evidence for deep brain stimulationPreventative treatment of chronic cluster headache

There are now >60 open-label cases of deep brainstimulation in the treatment of chronic cluster head-ache with an overall response rate of 66%.30 A recentsingle randomised placebo-controlled trial of 11patients treated with deep brain stimulation in chroniccluster headache used a 1-month period of activeversus sham stimulation.50 Although there were nodifferences in response rate between groups after1 month, this may have been because of too short afollow-up period: we now know that it takes 3–6 months to see a response to deep brain stimulation.

Acute treatment of cluster headache

Deep brain stimulation does not help in the acutetreatment of cluster attacks.51

Possible uses of deep brain stimulationEuropean Headache Guidelines apply to deep brainstimulation patient selection (table 1; see onlinesupplementary data table).41 Patients should haveproven refractory to all other treatments, includingother neurostimulation techniques, and need to havenormal brain anatomy. Implantation must only occur inhighly specialist centres and all candidates must havebeen approved by a multidisciplinary team, including apsychologist. There is no evidence to support the use ofdeep brain stimulation in chronic migraine.

Using deep brain stimulationDeep brain stimulation is reserved for patients withend-of-line refractory chronic cluster headache.Following implantation, a specialist programmes the

REVIEW

Miller S, et al. Pract Neurol 2016;16:362–375. doi:10.1136/practneurol-2015-001298 371

on June 8, 2020 by guest. Protected by copyright.

http://pn.bmj.com

/P

ract Neurol: first published as 10.1136/practneurol-2015-001298 on 5 M

ay 2016. Dow

nloaded from

device, which is then left switched on at all times.Unlike the other invasive neurostimulation devices,the patient is advised not to change their stimulationsettings. Initially patients should be followed up3-monthly to ensure the stimulation settings areadequate. Current implants involve a non-rechargeable battery with a lifetime of 5–7 years. As

with occipital nerve stimulation, if there is noimprovement after 1–2 years of stable, adequatestimulation, then the treatment is unlikely to help.Deep brain stimulation carries the potential for

serious risk. There is a single reported case of apatient with cluster headache dying from a post-operative intracerebral haemorrhage. The overall inci-dence of symptomatic haemorrhage in any deep brainstimulation surgery is around 2%, although our ownhighly experienced functional neurosurgery unit hasan incidence of 0.5%.52 Other adverse events fromour patients include transient diplopia or vertigo withchanges in programming (almost universal), neck stiff-ness (12%) and pain over the implant (19%).

Proposed mechanism of actionVentral tegmental area deep brain stimulation prob-ably acts upon pain circuits involved maintainingchronic cluster headache.53 Stimulation of this regionactivates the hypothalamus, thalamus, somatosensorycortex, anterior cingulate, and the ipsilateral trigem-inal nucleus and ganglion. These structures are alsoactive during acute cluster attacks.54 On the basis thatthe therapeutic effect takes several weeks, it has beenhypothesised that ventral tegmentum deep brainstimulation induces a functional modulation of thepain processing network in cluster headache ratherthan pure inhibition of hypothalamic activity.

CONCLUSIONHeadache syndromes are the most common disordersof the nervous system.1 In Europe, 1-year prevalenceof adults reporting headache was 51%, migraine 14%and chronic daily headache 4%.55 Traditionalmigraine preventatives have problems with low toler-ability and efficacy and so there is a demand for newtreatment options. However, all the studies to date

Figure 6 Sphenopalatine ganglion (SPG) stimulation. The SPGmicrostimulator (Pulsante device) in use. During an attack ofcluster headache, the patient holds the handset to the cheek toactivate the stimulator. Kindly reproduced with permission fromATI Technologies.

Figure 7 Deep brain stimulator for chronic cluster headache. (A) Skull X-ray shows placement of occipital nerve stimulationelectrodes and leads. (B) Postoperative MRI brain showing occipital nerve stimulation lead in right ventral tegmental area (arrow).

REVIEW

372 Miller S, et al. Pract Neurol 2016;16:362–375. doi:10.1136/practneurol-2015-001298

on June 8, 2020 by guest. Protected by copyright.

http://pn.bmj.com

/P

ract Neurol: first published as 10.1136/practneurol-2015-001298 on 5 M

ay 2016. Dow

nloaded from

addressing neuromodulation for headache have beeninadequate either due to lack of suitable placebo,small study populations, or both. Therefore, atpresent the conclusions that can be drawn from themare limited. Given the impact of chronic headaches onquality of life and the potential benefit that neurosti-mulation devices has for them, we clearly need robustrandomised-controlled trials on these treatments.There are issues involved in conducting such trials,the main one being what constitutes an adequateplacebo. With many of these devices, there arestimulation-related adverse effects that may unblind atrial subject—such as paraesthesia or muscle twitching—but which would be difficult to imitate with shamstimulation. Another issue with previous studies hasbeen the use of sham stimulation where the electricalcurrents are below the level believed to be clinicallyeffective. However, this level has never been studiedand it is possible that studies using this method haveused an active placebo, a complication in interpretingthe data. From the available efficacy data, neurostimu-lation treatments appear similar to that of standarddrugs but the major potential benefits of the treat-ments lie in their favourable adverse event profiles.Although current costs are high, this may be offset insome cases by reduced acute medication use (namelytriptans) with successful treatment. At present, thesetherapies could be considered for patients with drugtolerance issues or those who have proven medicallyrefractory. However, clinicians must discuss the limita-tions of these devices and the limited evidence fortheir use. In the future, if there is robust evidencethen neurostimulation devices may take a prominentplace alongside the arsenal of currently availablepharmacological treatment options for headache.

Contributors All authors: article drafting, key concepts, revisionand final approval of the manuscript.

Competing interests SM has received educational and travelgrants from St Jude Medical and Medtronic. She has receivedpayment for the development of educational presentations fromAllergan. AJS is funded by an NIHR Clinician ScientistFellowship (NIHR-CS-011-028) and the Medical ResearchCouncil, UK (MR/K015184/1). BD was a NICE CG 150Headache guidelines group member 2010-2012, NICEEvidence Update Work member 2014 and provided headacherelated advisory Consultancy work for NICE in 2015. He hasreceived advisory consultancy fees from AMGEN, USA, in2013 and advisory consultancy fees and lecture honoraria fromAllergan, USA, in 2013 and 2015 and lecture honoraria fromBMJ Masterclass 2014. He is a Migraine Trust Medical Trustee.MM serves on the advisory board for Allergan and St JudeMedical, and has received payment for the development ofeducational presentations from Allergan, Merck Sharpe andDohme and Medtronic.

Patient consent Obtained.

Provenance and peer review Commissioned; externally peerreviewed. Reviewed by Nicola Giffin, Bath, UK.

Open Access This is an Open Access article distributed inaccordance with the terms of the Creative CommonsAttribution (CC BY 4.0) license, which permits others todistribute, remix, adapt and build upon this work, forcommercial use, provided the original work is properly cited.See: http://creativecommons.org/licenses/by/4.0/

REFERENCES1 World Health Organisation. Atlas of headache disorders and

resources in the world 2011. A collaberation project of theWorld Health Organisation and Lifting the Burden. WHO,2011.

2 Scher AI, Stewart WF, Liberman J, et al. Prevalence of frequentheadache in a population sample. Headache 1998;38:497–506.

3 Scribonius L. Compositiones medicae. Padua: Frambottus, 1655.4 Oshinsky ML, Murphy AL, Hekierski H Jr, et al. Noninvasive

vagus nerve stimulation as treatment for trigeminal allodynia.Pain 2014;155:1037–42.

5 Sinclair AJ, Sturrock A, Davies B, et al. Headachemanagement: pharmacological approaches. Pract Neurol2015;15:411–23.

6 Gerardy PY, Fabry D, Fumal A, et al. A pilot study onsupra-orbital surface electrotherapy in migraine. Cephalalgia2009;29:101–78.

7 Schoenen J, Vandersmissen B, Jeangette S, et al. Migraineprevention with a supraorbital transcutaneous stimulator:a randomized controlled trial. Neurology 2013;80:697–704.

8 Bussone G, Diener HC, Pfeil J, et al. Topiramate 100 mg/dayin migraine prevention: a pooled analysis of double-blindrandomised controlled trials. Int J Clin Pract 2005;59:961–8.

9 Mulleners WM, Chronicle EP. Anticonvulsants in migraineprophylaxis: a Cochrane review. Cephalalgia 2008;28:585–97.

10 Linde K, Rossnagel K. Propranolol for migraine prophylaxis.Cochrane Database Syst Rev 2004;(2):CD003225.

11 Magis D, Sava S, d’Elia TS, et al. Safety and patients’satisfaction of transcutaneous supraorbital neurostimulation(tSNS) with the Cefaly® device in headache treatment: asurvey of 2,313 headache sufferers in the general population.J Headache Pain 2013;14:95.

12 Piquet M, Balestra C, Sava SL, et al. Supraorbitaltranscutaneous neurostimulation has sedative effects in healthysubjects. BMC Neurol 2011;11:135.

13 Gaul C, Diener H, Solbach K, et al. EHMTI-0364.Non-invasive vagus nerve stimulation using gammacore® for

Key points

▸ Controlled evidence for the use of neuro-stimulationdevices is relatively sparse, particularly for the non-invasive options; however, open-label series appearto be positive, especially for invasive techniques.

▸ Non-invasive neurostimulation targets include thevagus nerve, supraorbital nerves and the cortex (viatranscranial magnetic stimulation).

▸ Non-invasive devises are not routinely available onthe NHS without special permissions. However theirrole is expanding and potentially useful.

▸ Invasive neurostimulation techniques included occipi-tal nerve stimulation, sphenopalatine ganglion stimu-lation and ventral tegmental area deep brainstimulation.

▸ These techniques are reserved for those with refrac-tory primary headache conditions only, due to costand potential risks, and should only be conducted inhighly specialist centres.

REVIEW

Miller S, et al. Pract Neurol 2016;16:362–375. doi:10.1136/practneurol-2015-001298 373

on June 8, 2020 by guest. Protected by copyright.

http://pn.bmj.com

/P

ract Neurol: first published as 10.1136/practneurol-2015-001298 on 5 M

ay 2016. Dow

nloaded from

prevention and acute treatment of chronic cluster headache:report from the randomized phase of the preva study.J Headache Pain 2014;15(Suppl 1):I7.

14 Leone M, D’Amico D, Frediani F, et al. Verapamil in theprophylaxis of episodic cluster headache: a double-blind studyversus placebo. Neurology 2000;54:1382–5.

15 Nesbitt AD, Marin JC, Tompkins E, et al. Initial use of a novelnoninvasive vagus nerve stimulator for cluster headachetreatment. Neurology 2015;84:1249–53.

16 The Sumatriptan Cluster Headache Study Group. Treatment ofacute cluster headache with sumatriptan. N Engl J Med1991;325:322–6.

17 Cohen AS, Burns B, Goadsby PJ. High-flow oxygen fortreatment of cluster headache: a randomized trial. JAMA2009;302:2451–7.

18 Silberstein S, Neves Da Silva A, Calhoun A, et al. Non-InvasiveVagus Nerve Stimulation for Chronic Migraine Prevention in aPropective, Randomized, Sham-Controlled Pilot Study (theEVENT Study): report from the Double-blind Phase. Headache2014;54:1–74.

19 Goadsby PJ, Grosberg BM, Mauskop A, et al. Effect ofnoninvasive vagus nerve stimulation on acute migraine: anopen-label pilot study. Cephalalgia 2014;34:986–93.

20 Cady RK, Sheftell F, Lipton RB, et al. Effect of earlyintervention with sumatriptan on migraine pain: retrospectiveanalyses of data from three clinical trials. Clin Ther2000;22:1035–48.

21 Gallagher RM, Dennish G, Spierings EL, et al. A comparativetrial of zolmitriptan and sumatriptan for the acute oraltreatment of migraine. Headache 2000;40:119–28.

22 Lambru G, Matharu MS. Peripheral neurostimulation inprimary headaches. Neurol Sci 2014;35(Suppl 1):77–81.

23 Kraus T, Kiess O, Hösl K, et al. CNS BOLD fMRI effects ofsham-controlled transcutaneous electrical nerve stimulation inthe left outer auditory canal—a pilot study. Brain Stimul2013;6:798–804.

24 Beekwilder JP, Beems T. Overview of the clinical applicationsof vagus nerve stimulation. J Clin Neurophysiol2010;27:130–8.

25 NICE. Transcranial magnetic stimulation for treating andpreventing migraine, 2014. http://guidance.nice.org.uk/ipg477

26 Lipton RB, Dodick DW, Silberstein SD, et al. Single-pulsetranscranial magnetic stimulation for acute treatment ofmigraine with aura: a randomised, double-blind, parallel-group,sham-controlled trial. Lancet Neurol 2010;9:373–80.

27 Bhola R, Kinsella E, Giffin N, et al. Single-pulse transcranialmagnetic stimulation (sTMS) for the acute treatment ofmigraine: evaluation of outcome data for the UK post marketpilot program. J Headache Pain 2015;16:535.

28 Lipton RB, Pearlman SH. Transcranial magnetic simulationin the treatment of migraine. Neurotherapeutics2010;7:204–12.

29 Charles AC, Baca SM. Cortical spreading depression andmigraine. Nat Rev Neurol 2013;9:637–44.

30 Magis D, Schoenen J. Advances and challenges inneurostimulation for headaches. Lancet Neurol2012;11:708–19.

31 Saper JR, Dodick DW, Silberstein SD, et al. Occipital nervestimulation for the treatment of intractable chronic migraineheadache: ONSTIM feasibility study. Cephalalgia 2011;31:271–85.

32 Lipton R, Goadsby PJ, Cady R, et al. PRISM study: Occipitalnerve stimulation for treatment-refractory migraine.Cephalalgia 2009;29(Suppl 1):30.

33 Dworkin RH, Turk DC, McDermott MP, et al. Interpretingthe clinical importance of group differences in chronic painclinical trials: IMMPACTrecommendations. Pain2009;146:238–44.

34 Chen YF, Bramley G, Unwin G, et al. Occipital nervestimulation for chronic migraine–a systematic review andmeta-analysis. PLoS ONE 2015;10:e0116786.

35 Magis D, Allena M, Bolla M, et al. Occipital nerve stimulationfor drug-resistant chronic cluster headache: a prospective pilotstudy. Lancet Neurol 2007;6:314–21.

36 Magis D, Bruno MA, Fumal A, et al. Central modulation incluster headache patients treated with occipital nervestimulation: an FDG-PET study. BMC Neurol 2011;11:25.

37 Burns B, Watkins L, Goadsby PJ. Treatment of medicallyintractable cluster headache by occipital nerve stimulation:long-term follow-up of eight patients. Lancet2007;369:1099–106.

38 Burns B, Watkins L, Goadsby PJ. Treatment of intractablechronic cluster headache by occipital nerve stimulation in 14patients. Neurology 2009;72:341–5.

39 de Quintana-Schmidt C, Casajuana-Garreta E, Molet-Teixidó J,et al. [Stimulation of the occipital nerve in the treatmentof drug-resistant cluster headache]. Rev Neurol2010;51:19–26.

40 Fontaine D, Christophe Sol J, Raoul S, et al. Treatment ofrefractory chronic cluster headache by chronic occipital nervestimulation. Cephalalgia 2011;31:1101–5.

41 Martelletti P, Jensen RH, Antal A, et al. Neuromodulation ofchronic headaches: position statement from the EuropeanHeadache Federation. J Headache Pain 2013;14:86.

42 Sharan A, Huh B, Narouze S, et al. Analysis of Adverse Eventsin the Management of Chronic Migraine by Peripheral NerveStimulation. Neuromodulation 2015;18:305–12.

43 Matharu MS, Bartsch T, Ward N, et al. Centralneuromodulation in chronic migraine patients withsuboccipital stimulators: a PET study. Brain 2004;127(Pt 1):220–30.

44 Schoenen J, Jensen RH, Lanteri-Minet M, et al. Stimulation ofthe sphenopalatine ganglion (SPG) for cluster headachetreatment. Pathway CH-1: a randomized, sham-controlledstudy. Cephalalgia 2013;33:816–30.

45 Jürgens TP, Schoenen J, Rostgaard J, et al. Stimulation of thesphenopalatine ganglion in intractable cluster headache: expertconsensus on patient selection and standards of care.Cephalalgia 2014;34:1100–10.

46 May A, Goadsby PJ. The trigeminovascular system in humans:pathophysiologic implications for primary headache syndromesof the neural influences on the cerebral circulation. J CerebBlood Flow Metab 1999;19:115–27.

47 May A, Bahra A, Büchel C, et al. Hypothalamic activation incluster headache attacks. Lancet 1998;352:275–8.

48 Leone M, Franzini A, Bussone G. Stereotactic stimulationof posterior hypothalamic gray matter in a patient withintractable cluster headache. N Engl J Med2001;345:1428–9.

49 Matharu MS, Zrinzo L. Deep brain stimulation in clusterheadache: hypothalamus or midbrain tegmentum?. Curr PainHeadache Rep 2010;14:151–9.

50 Fontaine D, Lazorthes Y, Mertens P, et al. Safety and efficacyof deep brain stimulation in refractory cluster headache:a randomized placebo-controlled double-blind trialfollowed by a 1-year open extension. J Headache Pain2010;11:23–31.

REVIEW

374 Miller S, et al. Pract Neurol 2016;16:362–375. doi:10.1136/practneurol-2015-001298

on June 8, 2020 by guest. Protected by copyright.

http://pn.bmj.com

/P

ract Neurol: first published as 10.1136/practneurol-2015-001298 on 5 M

ay 2016. Dow

nloaded from

51 Leone M, Franzini A, Broggi G, et al. Acute hypothalamicstimulation and ongoing cluster headache attacks. Neurology2006;67:1844–5.

52 Zrinzo L, Foltynie T, Limousin P, et al. Reducing hemorrhagiccomplications in functional neurosurgery: a large case seriesand systematic literature review. J Neurosurg 2012;116:84–94.

53 Leone M, Franzini A, Proietti Cecchini A, et al. Success,failure, and putative mechanisms in hypothalamic stimulationfor drug-resistant chronic cluster headache. Pain2013;154:89–94.

54 May A, Leone M, Boecker H, et al. Hypothalamic deep brainstimulation in positron emission tomography. J Neurosci2006;26:3589–93.

55 Stovner LJ, Zwart JA, Hagen K, et al. Epidemiologyof headache in Europe. Eur J Neurol 2006;13:333–45.

56 Mueller O, Diener HC, Dammann P, et al. Occipital nervestimulation for intractable chronic cluster headache ormigraine: a critical analysis of direct treatment costs andcomplications. Cephalalgia 2013;33:1283–91.

REVIEW

Miller S, et al. Pract Neurol 2016;16:362–375. doi:10.1136/practneurol-2015-001298 375

on June 8, 2020 by guest. Protected by copyright.

http://pn.bmj.com

/P

ract Neurol: first published as 10.1136/practneurol-2015-001298 on 5 M

ay 2016. Dow

nloaded from