Abstract
Background
Non-invasive vagus nerve stimulation has initial evidence of efficacy in migraine and cluster headache. However, little is known about its role in the management of refractory chronic headaches.
Methods
We evaluated the preventive and abortive effects of non-invasive vagus nerve stimulation in 41 consecutive patients with refractory primary chronic headaches in an open-label prospective clinical audit. Headache diaries were used to collect clinical information. Those who obtained at least 30% reduction in headache days/episodes after three months of treatment were considered responders and were offered treatment continuation.
Results
Twenty-three patients with chronic migraine, 12 with chronic cluster headache, four with hemicrania continua and two with short-lasting unilateral neuralgiform headache attacks with autonomic symptoms (SUNA) were treated. Two of 23 chronic migraine patients, one of 12 chronic cluster headache patients, and two of four hemicrania continua patients were considered responders. None of the patients with SUNA benefited from the therapy. Two chronic migraine patients were able to reduce the pain severity of moderate migraines with non-invasive vagus nerve stimulation.
Conclusion
Non-invasive vagus nerve stimulation may not constitute an effective acute nor preventive treatment in refractory chronic primary headaches. The encouraging effect in hemicrania continua warrants further evaluation in larger studies.
Keywords
Introduction
In recent years, neurostimulation approaches have offered novel therapeutic avenues for the management of disabling headache conditions that are refractory to established pharmacological treatments (1). Invasive neurostimulation techniques targeting central (2) and peripheral (3,4) neuronal structures have yielded encouraging results in the management of chronic migraine (CM) and chronic trigeminal autonomic cephalalgias (TACs). However, the invasive nature of surgical procedures, along with the high rate of hardware-related adverse events and equipment costs, have limited their use to a few selected patients (5).
Meaningful headache improvement in patients with CM and chronic cluster headache (CCH) treated with invasive vagus nerve stimulation (VNS) primarily for refractory epilepsy (6) and refractory depression (7), led to experimentation using a non-invasive hand-held neurostimulation device targeting the cervical branches of the vagus nerve (gammaCore®) in the management of primary headaches. Clinical trials assessing the safety and efficacy of nVNS as an abortive and preventive option in migraine and cluster headache (CH) have shown encouraging results (8–12).
A position statement of the European Headache Federation (EHF) supports the need for further research on the role of non-invasive neurostimulation therapies in the treatment pathway of patients who may otherwise be candidates for invasive neurostimulation approaches (13). Furthermore, the National Institute of Health and Clinical Excellence (NICE) in the United Kingdom (UK) encourages the use of non-invasive neurostimulation therapies in the context of audit or research for the management of migraine and CH (14). Our tertiary referral Headache Centre assesses patients with difficult-to-treat chronic headache disorders. A significant proportion of them have failed numerous medical treatments and may potentially be candidates for invasive neurostimulation approaches. For this reason, we aimed to assess whether non-invasive neurostimulation approaches, such as nVNS, may have a role in the treatment pathway of these patients before invasive neurostimulation procedures are considered.
Methods
Participants and study design
This is a prospective audit assessing the effectiveness, compliance and tolerability of vagus nerve stimulation delivered via the portable, battery-driven, CE-marked medical device, gammaCore (electroCore LLC, Basking Ridge, NJ). Forty-one consecutive medically refractory patients meeting the International Headache Society (IHS) criteria for CM and TACs, seen at the Headache Centre at Guy’s and St Thomas’ Hospital, London, UK, between January 2014 and August 2016, were included in this audit. The definitions of refractory CM and CCH were based on the EHF recommendations (15,16). In view of the lack of guidance on the definition of refractoriness for the other TACs, we considered refractory the hemicrania continua (HC) patients with the unremitting subtype who failed to tolerate long-term exposure to indometacin and failed to respond to or tolerate trials of greater occipital nerve blockade, topiramate, verapamil, melatonin and gabapentin or pregabalin. Similarly, chronic short-lasting neuralgiform headache attacks with autonomic symptoms (SUNA) patients were considered refractory if they had failed to tolerate or respond to adequate trials of lamotrigine, topiramate, gabapentin, pregabalin, greater occipital nerve blockade and at least one of either carbamazepine or oxcarbazepine (17).
Patients were trained to use the device as a preventive and abortive treatment either in monotherapy or in conjunction with other treatments. Patients with active neurostimulation devices, cardiac pacemakers, or with a significant history of cardiac arrhythmia were not offered the treatment. Patients kept a headache diary for at least a month at baseline. Treatment with nVNS was offered for three months to assess its effectiveness. Subsequently, patients were seen in our outpatient headache clinic to discuss their progress. This project is an audit of outcome so, in line with national audit guidelines, ethics approval was not required (http://www.hra-decisiontools.org.uk/research/).
Stimulation paradigm
The nVNS device produces a proprietary low-voltage electrical signal comprising a 5 kHz sine wave burst lasting for 1 millisecond (five sine waves, each lasting 200 microseconds), with such bursts repeated once every 40 milliseconds (25 Hz), generating a 24 V peak voltage and 60 mA peak output current; users can adjust the stimulation amplitude (11). Patients were instructed by an electroCore representative to use two consecutive nVNS doses (90 seconds each) on one side of the neck or alternating right and left sides, three times a day, as a preventive stimulation paradigm. As an acute treatment for headache episodes, patients were advised to use up to three additional consecutive doses before resorting to their usual abortive treatment.
Data collection
Data on nVNS treatment outcomes were prospectively collected from headache diaries, specifically designed for migraine and CH. Headache-related disability was assessed using the Headache Impact Test (HIT-6) for migraine patients only, since no disability tools are currently validated for TACs. Data on device safety, tolerability and treatment compliance were collected during outpatient visits. Frequency, intensity (Numeric Rating Scale (NRS); 0 = no pain to 10 = very severe pain) and duration of headache episodes, days with acute headache medication consumption and change in headache-related disability (where appropriate) were evaluated after three months of treatment and compared to baseline in both the CM and TACs groups. Satisfaction with nVNS therapy was evaluated using patients’ subjective impression of change (as a percentage: 0% = no headache change, 100% = total headache relief).
Clinical outcomes
In line with the Initiative on Methods, Measurement, and Pain Assessment in Clinical Trials (IMMPACT) recommendations and the IHS clinical trials sub-committee (18,19), we considered ≥ 30% reduction of headache days after three months’ treatment as a clinically-relevant outcome to justify treatment continuation. These patients were referred to as responders. A moderate to severe headache day was defined as any headache lasting ≥ 4 hours on an NRS ≥ 4/10. A migraine day was defined as a day with migraine or probable migraine according to the ICHD-3β criteria (20). In view of the real-life exploratory nature of this report and the complex group of patients selected, other outcome measures, such as change in headache severity, and patients’ subjective impression of change, were taken into account when considering whether to continue the treatment beyond three months. Furthermore, compliance with the treatment protocol was assessed by looking at the time taken for a patient to deplete the preset 300 doses of each nVNS device, though this information was not available for every patient. Safety and tolerability were assessed by direct patient questioning in clinic at three-month follow-up and during future scheduled follow-ups for those who continued the treatment. Patients were instructed to contact us via email or telephone should they experience any adverse reactions in between scheduled follow-up appointments. The data were collated on a Microsoft Excel spreadsheet and kept in a password-protected file within the hospital’s computers.
Results
Demographic characteristics and failed medical treatments of chronic migraine patients.
F: female; M: male; CM: chronic migraine; MOH: medication overuse headache; SHM: sporadic hemiplegic migraine; nVNS: non-invasive vagus nerve stimulation. Prophylactic medication classes: anticonvulsants, beta-blockers, antidepressants, calcium antagonists, 5-HT2 receptor antagonists.
Demographic characteristics and failed medical treatments of chronic trigeminal autonomic cephalalgia patients.
F: female; M: male; CCH: chronic cluster headache; GONB: greater occipital nerve blockade; ONS: occipital nerve stimulation. Prophylactic medication classes: Calcium channel blockers, anticonvulsants, corticosteroids, lithium carbonate, melatonin.
Migraine
Clinical characteristics of refractory chronic migraine patients pre- and post- non-invasive vagus nerve stimulation treatment.
HIT-6: Headache Impact Test score; md: missing data; Pt: patient.
Abortive effect: None of the patients treated with nVNS reported being able to abort migraines entirely. However, two patients (patients 21 and 22) reported being able to reduce the pain severity of moderate migraines to a manageable level of severity with nVNS, though their abortive treatment intake remained the same as at baseline.
Headache related disability: According to their HIT-6 scores, all but one of the patients were considered severely disabled (Grade IV, score ≥60) at baseline. After three months, 11 patients reported a reduction of more than two points on the HIT-6, reflecting an improvement in headache-related disability that may be considered clinically significant (21). However, none of these moved down from grade IV level of impact. One patient (patient 19) had a substantial improvement on the HIT-6, resulting in a reduction from grade IV to grade I impact level. All the other HIT-6 scores were unchanged (Table 3). Data on HIT-6 was incomplete for three patients.
Treatment continuation: The two responders (patients 13 and 19) had been using the treatment respectively for 12 and eight months at the time of the last follow up and both are continuing.
Three other patients (patients 4, 5 and 23) continued the treatment beyond the initial three-month trial period. In one of these patients (patient 4) this decision was on the basis of a reduction in the headache intensity, which the patient considered to be a 40% improvement. Patient 5 demonstrated a reduction of headache and migraine days, which did not surpass the 30% threshold but was subjectively considered to be a 50% improvement. Of these three patients, one discontinued the treatment after 12 months due to lack of ongoing benefit (patient 4), one discontinued the treatment after five months because of no benefit and one patient (patient 5) had been using the treatment for 21 months and is still continuing. We observed no cumulative benefit beyond the three-month trial in any of those who continued the treatment.
Trigeminal autonomic cephalalgias
Cluster headache
Clinical characteristics of refractory chronic cluster headache patients pre- and post- non-invasive vagus nerve stimulation treatment.
md: missing data.
Abortive effect: None of the patients reported headache relief using nVNS to abort a CH episode.
Treatment continuation: Patient 4 continued nVNS for 10 months, at which stage the patient noticed a worsening of their condition for three consecutive months despite the therapy and hence discontinued the treatment.
Hemicrania continua
Clinical characteristics of refractory hemicrania continua patients pre- and post- non-invasive vagus nerve stimulation treatment.
HIT-6: Headache Impact Test score; md: missing data.
Short-lasting unilateral neuralgiform headaches
One patient had a 10-year history of chronic SUNA characterised by 60–300 right-sided headache attacks/day, lasting between two to 40 seconds, on an NRS of 8–10/10, accompanied by ipsilateral eyelid oedema and conjunctival injection. The second patient had an eight-year history of SUNA characterised by 10–25 right-sided attacks/day, lasting between 120 to 180 seconds on an NRS of 7/10, accompanied by ipsilateral lacrimation, ptosis, and nasal blockage. Neither of these patients reported any noticeable improvement after using nVNS for three months either as an abortive or as a preventive treatment.
Adverse events and adherence to treatment
There were no suspected unexpected serious adverse reactions. Mild or moderate anticipated adverse reactions were reported in five out of 40 patients (12.5%). These included temporary hoarseness/sore throat in two patients, swollen/red skin around the face and neck in one patient, nausea in one patient (which resolved by not using nVNS immediately after meals), increased frequency of bowel movements/flatus in one patient and facial twitching in one patient. No patients discontinued the treatment solely due to adverse reactions.
We were able to obtain data for 16 of the patients in our cohort (11 with CM, four with CH and one with HC), regarding the dates when they requested replacement nVNS devices. Since each device was pre-loaded with 300 doses, sufficient for 50 days’ preventive treatment, we were able to postulate that at least seven of these 16 patients were not compliant with the recommended preventive dosing regime (six patients with CM, 55%; and one with CH, 25%) on the basis of when they requested replacement devices. However, this does not take into account any additional doses that may have been used to try and treat headaches acutely, suggesting that an even greater number of patients may have been non-compliant with the preventive treatment regime.
Discussion
This audit aimed to assess the effectiveness of nVNS therapy in patients with primary chronic headache disorders refractory to established treatments. Our experience suggests that nVNS may not constitute an effective option in this group of patients, with the exception of HC patients, where the therapy might represent a useful treatment for those who do not tolerate indometacin.
Migraine
In the migraine group, a reduction of 30% or more in the number of headache days was observed in just two patients out of 23. Despite different study methodologies, our overall treatment outcomes were in keeping with the findings of the EVENT study (10). In this randomised-control pilot trial, nVNS used for two months showed no significant reduction in the number of headache days in 30 CM patients compared to the group treated with sham (n = 29 patients). The proportion of treatment responders using a threshold of 50% reduction of headache days in the EVENT study was higher (three out of 30 patients, 10%) than in our report, where no patients reported a 50% reduction in headache days after three months of nVNS treatment but two patients reported a reduction in headache days of at least 30% (8.7%), which we considered a realistic outcome in our refractory group of patients.
Our marginally less positive outcomes compared with those of the EVENT study could be explained by the severity of our patients’ conditions, which were refractory to medical treatment and hence more difficult-to-treat. However, no indication of the classes, nor number of failed preventive treatments was provided in the reporting of the EVENT study. It is also possible that the different definition of a headache day adopted in the EVENT study (“Any day on which a participant recorded a headache”), compared to the IHS definition of a headache day which we adopted (“a headache for at least 4 hours on a NRS of at least 4/10”), contributed to the slightly higher responder rate (19). The limitation of using “any headache” when establishing a primary outcome is that a reduction of short and mild headache episodes (<4 hours and <4 on NRS) or longer (>4 hours) but mild (<4 NRS) headache episodes may not be as clinically-relevant as a reduction in long-lasting moderate or severe headache episodes, hence a 50% reduction in headache days may not reflect a meaningful reduction in headache-related disability.
In our CM patients who experienced an initial improvement and continued nVNS therapy beyond the three months’ trial, no cumulative effect was noticed (range: 5–18 months). This is in contrast to the EVENT study, where a cumulative effect was postulated in those who continued nVNS treatment beyond the randomised controlled phase. Of the 59 patients included in the randomised phase, 44 patients completed the two-month phase; of these, 15 patients completed the open-label six months’ treatment phase. Those treated with nVNS reported a significantly higher reduction in headache days compared to those who started in the sham arm and subsequently continued with nVNS (10). However, it is possible that only the subgroup of patients, namely the self- identified responders willing to continue the treatment, reported this effect (n = 15).
The effectiveness of nVNS in refractory migraine patients was also explored in a small open label study (22). However, only 10 of the 20 patients studied had chronic migraine, whereas the others had episodic migraine. The authors of this study claimed that all patients enrolled were medically refractory, including episodic migraine patients and CM patients who had not tried Botulinum toxin type A (BoNTA). According to both the American and the European guidelines on the definition of refractory migraine, only those patients with CM who have failed to respond to BoNTA treatment can be labeled as refractory (15,23). For these reasons, a comparison with our study is not possible.
The encouraging migraine abortive effect of nVNS suggested in two open-label studies (8,9) was not noticed in our group of patients, although the audit was not powered to assess the efficacy of nVNS as an abortive migraine treatment. As suggested by Barbanti et al. (8), it may be possible that CM patients respond less to nVNS when used acutely than episodic migraine patients, which may explain the outcome differences amongst our patient group.
Cluster headache
In the CCH group, only one patient obtained >30% reduction in weekly CH frequency and was willing to continue the therapy beyond the three-month trial. Unfortunately, the benefit diminished after some months. These results are disappointing compared to the outcomes of two open-label studies that evaluated the preventive and abortive effect of nVNS in CCH (12,24). In the PREVA study, the use of nVNS in addition to the standard of care was compared to standard of care alone in CCH patients. The former group of patients benefited significantly in terms of reduction of weekly CH attacks from using nVNS to standard of care, compared with those who continued with standard of care only (12). It is possible that the refractory nature of the condition in the subjects included in our audit may have contributed to the low proportion of responders. It is also possible that the higher response rate reported in the PREVA study reflects the relatively short use of nVNS therapy (two months).
It is more difficult to explain the difference in outcomes of our audit compared to those of a previous audit conducted in 11 CCH patients treated with nVNS, seven of whom were considered refractory (24). The authors reported some benefit in eight of 11 CCH patients, though only five of the 11 patients had a one-year follow-up and no breakdown of the treatment outcomes in the refractory group was reported. It is possible that the lower proportion of responders in our audit resides in a different methodology for data collection. The majority of our patients provided prospective CH diaries. The clinical decision on treatment continuation beyond the three-month trial was made on the basis of patients’ headache diary data as well as their subjective estimation of benefit, whereas in the previous audit, the treatment outcome was based on patients’ estimation of benefit only (24).
The recent Food and Drug Administration (FDA) release on nVNS for the acute treatment of pain associated with episodic CH was based on the outcomes of two prospective, double-blind, placebo-controlled randomised studies: the already published ACT1 (25), and the ACT2, the outcome of which was recently presented at the American Academy of Neurology meeting (Boston 22–28 April 2017). The outcomes of these clinical trials suggested a poor abortive effect of nVNS in CCH patients, confirming the findings of our report. In the ACT1 study, the proportion of patients who achieved pain relief using nVNS compared to sham in the cohort of CCH was not statistically significant. However, the response rate was significantly higher in the nVNS treated cohort compared to the sham treated cohort, when only the subgroup of episodic CH patients was considered (24), suggesting a meaningful effect of nVNS in aborting CH attacks in episodic CH.
Hemicrania continua
An encouraging proportion of responders to nVNS was observed in our HC group. Our results are in line with initial evidence of efficacy of nVNS in difficult-to-treat HC patients, (26,27), suggesting a possible role of nVNS in HC patients who do not tolerate indometacin or for whom long term use of indometacin is contraindicated.
SUNA
The use of nVNS for three months in our two SUNA patients did not provide any meaningful benefit. Given our limited experience using nVNS in SUNA it is difficult to reach any definite conclusion on the potential effect in these conditions. However, in view of the disappointing initial results in these two patients, we have refrained from offering this treatment to other SUNA patients until further data is published.
The limitations of this audit include the presence of incomplete diaries and the lack of a sham arm. Furthermore, it would be difficult to achieve a reliable sham-controlled study, maintaining adequate blinding of participants to their treatment allocation, when using a paraesthesia-inducing device such as nVNS (5). However, whilst refractory headache patients may have particularly raised expectations of novel treatments, the low proportion of responders suggests a limited placebo effect in our group of patients. The strengths of our audit include the real-word nature of this report, which intended to provide the first evidence on the use of nVNS in a clinical setting; the homogenous selection of patients according to the recently published consensus on the definition of refractory headaches (15,16) and the use of NICE UK audit tools to assess treatment outcomes and headache-related disability.
Limited evidence-based effective therapeutic options are available for patients with refractory primary headaches. In view of the disappointing efficacy and tolerability outcomes of occipital nerve stimulation (ONS) in CM clinical trials (28,29) and the lack of RCTs testing ONS in CCH, there is currently an urgent need for novel treatments in this challenging-to-treat population. Non-invasive neurostimulation procedures could be a reasonable option before more invasive and expensive treatments are considered. Unfortunately, our results suggest that in a real world setting, nVNS may not constitute a meaningful preventive nor acute therapeutic option in this population. It remains to be investigated whether introducing the device at earlier stages of the headache management pathways may provide better outcomes in chronic primary headache sufferers. The results of the use of nVNS in HC patients seem more promising and warrant further studies in a larger group of patients.
Clinical implications
Clinical trials have shown the potential utility of non-invasive vagus nerve stimulation (nVNS) in migraine and cluster headache treatments. However, in a real-world setting, nVNS may not constitute a meaningful preventive nor acute therapeutic option in patients with primary refractory headache disorders, especially chronic migraine and chronic cluster headache. In patients with refractory hemicrania continua, nVNS may constitute an effective preventive option that warrants being explored in larger studies.
Footnotes
Declaration of conflicting interests
The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: Dr Trimboli: nothing to declare. Dr Al-Kaisy has received travel sponsorship and speaker fees from Medtronic and Nevro Corp, and he is the principal investigator in separate studies sponsored by Medtronic and Nevro Corp. Dr Andreou: reports a travel grant from eNeura. Mrs Murphy: received a travel grant from Electrocore. Dr Lambru has received speaker honoraria, funding for travel, and has received honoraria for participation in advisory boards sponsored by Allergan. He has received speaker honoraria, funding for travel from electroCore, Nevro Corp and Autonomic Technologies.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The audit received no specific grant. Electrocore provided the devices for a three-month trial for each patient, but had no influence in data collection or writing the manuscript.