Peripheral Nerve Stimulation for Treatment of Headaches: An Evidence-Based Review

Headaches are one of the most common medical complaints worldwide, and treatment is often made difficult because of misclassification. Peripheral nerve stimulation has emerged as a novel treatment for the treatment of intractable headaches in recent years. While high-quality evidence does exist regarding its use, efficacy is generally limited to specific nerves and headache types. While much research remains to bring this technology to the mainstream, clinicians are increasingly able to provide safe yet efficacious pain control.


Introduction
Headaches are one of the most common medical complaints worldwide, with a prevalence rate of 46% in the general population [1]. A majority of headaches are of the migraine, tension, and cluster subtype, but many other forms exist. The healthcare burden of this is profound, as headaches consistently rank as the fourth or fifth most common reason for emergency-department visits in the United States, and they make it into the top 10 most disabling conditions according to the World Health Organization (WHO) [2]. There is also a significant economic impact from the burden of disease and the often-high cost of treatment. Unfortunately, diagnosis can be difficult, due in part to the complexity of its classification, contributions of referred pain from the spine, and presence of craniofacial pain syndromes. As our knowledge of the nervous system and pathophysiology of headaches and craniofacial pain has advanced, therapeutics have steadily moved towards directly targeting nerves of interest. After conservative measures have been exhausted, peripheral nerve stimulation (PNS) and field stimulation (PNFS) have emerged as a novel therapy for intractable headaches. Given the relatively recent advent of neuromodulation, this review seeks to help clinicians better understands the efficacy of PNS for headaches. We hope to summarize the high-yield evidence for the physician caring for these patients and provide an evidence-based overview of treatment strategies.

Conventional Management
Due to the complexity in etiology and presentation of headaches and craniofacial pain, treatment can be challenging. Typically, the first step in management is to explore and avoid known triggers, often in the form of caffeine, stress, poor sleep hygiene, and aspartame [5]. If pain continues despite this, the next step is typically pharmacological treatment in the form of oral medications, due to their relatively safe side-effect profile and simplicity of use.
Typically trialed in a stepwise fashion, escalations in management are made if no improvement in symptoms is seen. These are often measured by pain scoring (visual or otherwise) and by measures of disability, including days off work and days with symptoms. Oral pain medications are typically classified as either preventative (prior to an acute episode) or abortive (during an attack) [6]. While specific pharmaceuticals are used for the different headache subtypes, nonsteroidal anti-inflammatory drugs, anticonvulsants, selective serotonin reuptake inhibitors, tricyclic antidepressants, triptans, and corticosteroids are typically used as abortive medications [6,7]. Conversely, preventative therapies often include beta blockers, calcium channel blocks, and other anticonvulsants.
Despite multimodal approaches in management, many refractory cases resistant to conventional treatments subsist. Interestingly, medication overuse headaches have become increasingly prevalent, leading to a need for non-pharmacological interventions. Novel treatments pioneered and trialed in recent years include botulinum injections, nerve blocks, acupuncture, physical therapy, trigger-point injections, deep brain stimulation, and facet joint radiofrequency ablation for cervicogenic headaches [4,7]. While a thorough discussion of the evidence behind the techniques is outside the scope of this review, it is clear that there is no one-size-fits-all approach. Increasingly, PNS has been utilized as a minimally invasive intervention for these complex disorders. Evidence for its use is the primary focus of this review.

Indications
While specific guidelines are sparse, the use of PNS has generally been indicated for chronic neuropathic pain disorders originating from specific peripheral nerves. The source of pain should logically follow that of a specific nerve distribution. While the exact mechanism of pain relief is unclear, it is thought to involve activation of central endogenous pain-modulation pathways [8]. As with any other pain modality, the success of PNS relies largely on patient selection. A typical patient for PNS is one who has failed conventional medical treatments and one for whom surgery has been deemed inappropriate [9]. After establishing a diagnosis and exhausting conservative treatment measures, patients are typically referred to a specialist for further discussion of symptoms, physician examination, and review of relevant imaging. Comorbid psychological illnesses should also be explored, as these can significantly reduce intervention success rates [10][11][12]. PNFS indications are even more nebulous, as these involve targeted stimulation of smaller mostly unnamed nerves at the site of pain [11]. Various targets and techniques are outlined below, in addition to the high-yield evidence for each.

Contraindications and Complications
There are generally few absolute contraindications for PNS. These typically include medical allergies, local infections, coagulopathies, immunocompromised status, comorbidities preventing fluoroscopic needle guidance, and patient refusal [11,13]. As these are peripheral procedures, the risk of permanent damage is relatively low. Complications are also similarly mild, but they can include headaches, muscle cramping, subcutaneous hematomas, seromas, and local skin infections [8]. Other common complications, along with their incidence seen in literature, are highlighted in Table 2 below. One of the most common themes in preventing success of treatment is improper lead placement or postoperative migration, causing displacement of electrodes from their target nerves [14]. While correctable, this is often frustrating for the patient and can worsen outcomes. There are numerous possible targets for PNS, with varying levels of evidence for each. Outlined is a high-yield summary of the common targets for headaches/craniofacial pain and broad evidence for their use (Table 3). These were further classified as highquality evidence (primary from randomized controlled trials), moderate-quality (a mix of both randomized controlled trials and observational studies), or low-quality evidence (observational studies primarily from case reports or case series). For the physician making treatment decisions on these techniques, this quick guide can serve as a resource for decision-making on interventional approaches. Additional details regarding each highquality trial are also presented for each stimulation technique. Overall, research tends to skew towards weaker observational studies versus higher-quality RCTs. Figure 1 breaks down the study types of location of stimulation, whereas Figure 2 breaks down the available literation by indication.

Occipital Nerve Stimulation (ONS)
One of the most well-studied targets for neurostimulation is the occipital nerve ( Table 3). The occipital nerves are a group of nerves that originate from the C2/C3 spinal nerves and innervate a significant portion of the scalp, ear, and other nearby structures. The three major occipital nerves include the greater occipital nerve (GON), lesser (or small) occipital nerve (LON), and third (least) occipital nerve (TON) [75,76]. Common pathologies treated by occipital nerve stimulation include occipital neuralgia, chronic migraines, and cluster headaches (Table 4).  A review of the literature yielded a total of six randomized controlled trials, all of which involved the use of ONS for chronic migraines. Of these, three were industrysponsored multicenter studies (PRISM study, ONSTIM, and Dodick et al.) [8,23,30]. The PRISM study was one of the first, but the results were published only as an abstract presentation with no formal manuscript submitted [23]. A total of 140 patients were randomized to either bilateral ONS or sham, with 125 completing the 12-week follow-up. No statistical difference was found between the treatment group and sham for reduction in migraines day per month. Later, Saper et al. completed ONSTIM, a multicenter feasibility trial where respondents were randomized to adjustable stimulation, preset stimulation, or medical management groups [30]. For the 66 patients who completed their headache diaries, 39% of patients receiving adjustable stimulation reported a >50% reduction in headache days or three-point greater decrease in average overall pain scores; symptom reduction on average was only 6% for preset stimulation, and 0% for medical management. Furthermore, Silberstein and Dodick et al. later completed a RCT with 157 patients at both the 12-and 52-week follow-ups [8]. While there was no significant reduction in VAS (i.e., by more than 50%) at 12 weeks, patients did report a significant reduction in headache days, headache reduction, and migraine-related disability. At 52 weeks, 59.5% of patients reported a 30% reduction in headache frequency or intensity, and 48% had an improvement of >50%. Overall, this study helped to cement the long-term viability of ONS for chronic migraines. In 2017, Mekhail et al. completed a single-center RCT, with 20 patients randomized to either an active or control group for 12 weeks, after which they received open-label treatment for an additional 40 weeks [17]. Overall, they found an overall average reduction in headache days per month by 8.51, with 60% of patients receiving >30% relief in headache days or pain intensity and 35% reporting >50%. Furthermore, MIDAS and Zung PAD scoring (migraine related disability) were reduced for all patients in the study.
Two single-site crossover RCTs have found similar results [32,33]. Serra and Marchioretto's study randomized 30 patients with chronic migraines or medication-overuse headaches to either stimulation or sham, with a crossover after one month had passed or if headache had worsened [32]. Overall, they found significantly lower headache intensity and frequency in the stimulation arm, decreased MIDAS scores at the 1-year follow-up, significantly increased quality of life, and decreased triptan and NSAID use at each subsequent follow-up. Later, Slotty et al. evaluated various stimulation thresholds (effective stimulation, subthreshold, and no stimulation) in migraine patients already being treated with ONS and the importance of paresthesia [33]. They found a significant difference in pain (VAS) with suprathreshold stimulation (where patients reported paresthesia during stimulation) and subthreshold stimulation (1.98 ± 1.56 vs. 5.65 ± 2.11). Similar results were seen when comparing subthreshold stimulation and no stimulation (5.65 ± 2.11 vs. 8.45 ± 0.99). Overall, this study found that, while paresthesia was not required to achieve pain reduction, suprathreshold stimulation yielded better pain control, highlighting the importance of stimulator customization.
In 2017, Liu et al. published an RCT evaluating transcutaneous occipital nerve stimulation to prevent migraines [24]. This study randomized 110 patients into five groups who received transcutaneous ONS (tONS) at various frequencies (2, 100, and 2/100 Hz), underwent sham tONS, or received oral topiramate. Overall, they found that tONS at all frequencies, and topiramate, all yielded a significant reduction in headache days compared to baseline at the 3-month follow-up. However, only the tONS at 100 Hz and topiramate groups showed a significant difference when compared with sham. Interestingly, tONS, topiramate, and sham all led to decreases in VAS, potentially speaking to the placebo-effect of tONS. Overall, this study suggests that tONS may also be effective in the prevention of chronic migraines.

Trigeminal Nerve Stimulation
The trigeminal nerve carries sensory components for much of the head and innervates muscles in the lower jaw. It then divides into the ophthalmic (V1), maxillary (V2), and mandibular (V3) branches [78]. The ophthalmic further divides into the supraorbital nerve and the supratrochlear nerve, while the auriculotemporal nerve is a branch of the mandibular nerve.
Jiang et al. completed a RCT evaluating the combination of flunarizine and transcutaneous supraorbital neurostimulation (tSNS) on episodic migraines on patients randomized to flunarizine only, tSNS only, or both flunarizine and tSNS [55]. They found that while monthly migraine days was decreased in all three groups, pain reduction by >50% was greatest (78%) in the combination therapy group than either flunarizine (46%) or tSNS (39%) alone, showing the efficacy of pharmacological and interventional therapies in treating migraines (Table 5).
In 2019, Chou et al. conducted a multicenter RCT with 109 patients evaluating the efficacy of tSNS for acute migraine attacks. They found that one hour of tSNS treatment resulted in a 59% decrease in VAS versus a 30% decrease with sham [54]. Overall, these studies help support the efficacy of external neuromodulation as viable treatment for migraine headaches.
Several studies have also been conducted on the efficacy of neuromodulation on multiple concurrent nerve targets ( Table 6). The first RCT to show the effectiveness of noninvasive supraorbital and supratrochlear PNS for migraines was completed by Schoenen et al. in 2013 [57]. In this study, 67 patients with at least two migraine attacks per month were randomized to either sham or stimulation with daily sessions of tSNS with Cefaly device. After 3 months of treatment, the stimulation group experienced a significant reduction in the average number of migraine days, with 38% achieving a >50% response. This study overall demonstrated a 26% therapeutic gain, which is within the range of those reported for other commonly used migraine treatments.

Vagal Nerve Stimulation
The vagus nerve plays a major role in the autonomic nervous system, regulating metabolic homeostasis, control of various organs/glands/muscles in the body, and mediating the transfer of sensory information throughout the body [80]. Historically, neuromodulation of this nerve has played a role in treating epilepsy and depression. In more recent years, there has been a movement towards both invasive and noninvasive methods of vagus nerve stimulation for the treatment of migraine and cluster headaches (Table 7).
Gaul et al. conducted a large multicenter PREVA RCT in the UK, evaluating the use of noninvasive VNS as an adjunct to standard of care (oxygen and triptans) for chronic cluster headaches [69]. As one of the largest RCTs with a total of 97 patients, this study found a significantly greater reduction in the mean number of attacks per month (5.9 vs. 2.1) compared to standard of care. In addition, 40% of patients with the adjunct nVNS reported a greater than 50% reduction in pain as compared to just 8.3% for the control. Overall, this study suggests that nVNS has a beneficial role as an adjunct to standard of care for cluster headaches. Silberstein et al. conducted the multicenter double-blinded RCT called ACT1 to evaluate the use of nVNS on acute treatment of episodic and chronic cluster headaches [73]. Interestingly, the study found a significantly higher response rate (pain relief based on scoring) for episodic cluster headaches versus the sham, but not for the chronic cohort. Goadsby et al. conducted a similar study on the European population and found comparable results [70]. This suggest that nVNS is more effective at treating episodic cluster headache than chronic cluster headache.
Subsequently, Silberstein completed the EVENT RCT to assess the use of nVNS for chronic-migraine prevention [67]. Patients were randomized to nVNS or sham, with treatments administered personally at a specific time of the day. Overall, they found no statistically significant difference at the 2-month follow-up. Following this, a small subset of patients completed a further 6 months of open-label treatment, where statistical significance (decreased headache days) was achieved. While limited, this suggests that continued nVNS use may have a benefit for chronic-migraine prevention.
Finally, Tassorelli et al. completed the PRESTO RCT to evaluate nVNS as an acute treatment for migraines [74]. Overall, this found that nVNS was superior to the sham in aborting the first treated attack of acute migraine at 30 and 60 min, with repeat-measures testing supporting superiority in pain freedom from 30 to 120 min. The pain-free response rate at 120 min was also similar to those achieved by oral triptans and NSAIDs, suggesting their usefulness for acute migraines.

Peripheral Nerve Field Stimulation
Peripheral nerve field stimulation involves the placement of electrodes near the area of the pain without direct contact to a specific peripheral neve [84]. While research in this area is relatively new, several observational studies have been completed to evaluate this treatment's efficacy. Verrills et al. evaluated 83 patients who had undergone PNFS targeting the occipital, supraorbital, and infraorbital nerves for chronic daily headache, chronic migraines, and occipital neuralgia [20]. Overall, they found a mean NRS decrease of 4.8, with 68% of the patients experiencing >50% reduction in pain. Furthermore, 23/35 of the patients at follow-up reported a moderate-to-extreme decrease in analgesic use at follow-up, highlighting its possibility for headache treatment.
In 2018, Ishiyama et al. applied C2 PNFS, using electroacupuncture for primary headache [85]. In this observational study of 54 patients, significant decreases in NRS pain was found with C2 PNFS use, alongside decreases in HIT-6 and SDS (self-rating depressing scale), and decreases in monthly headache days. While more research needs to be done for PNFS, this remains a safe and exciting field for further study.

Remote Electrical Neuromodulation
Remote electrical neuromodulation (REN) is a novel technique for the treatment of acute migraine. The device is placed on the upper arm and stimulates peripheral nerves in the region to induce conditioned pain modulation. Yarnitsky et al. published a randomized, double-blinded, sham-controlled multicenter study across 12 sites with 252 adults to evaluate pain relief at 2 h post-stimulation [86]. They found that active stimulation was more effective in pain relief at 2 h versus the sham, sustained up to 48 h after treatment. While further research remains, this suggests REN can be an effective treatment for acute migraines.

Conclusions
Peripheral nerve stimulation offers promising treatments for intractable headaches. Their success relies on accurate diagnosis and appropriate patient selection. While highquality evidence does exist regarding its use, additional research is needed in the field. For the physician wishing to utilize PNS in practice, the strongest up-to-date evidence for its use is ONS for treatment of chronic migraines, transcutaneous supraorbital nerve stimulation for migraines, and nVNS for cluster headaches. Preliminary evidence also suggests that interventional therapies may be beneficial as adjuncts to standard-of-care therapies (oxygen, triptans, etc.), thus forming a potential area of study. Looking towards the future, much work remains to bring PNS into the mainstream for headache intervention.