Abstract
Background
Occipital nerve stimulation (ONS) is a treatment with evidence in refractory chronic cluster headache (CCH). However, the variable response rate and cost make it necessary to investigate predictors of response.
Methods
This is a cross-sectional study conducted through the review of medical records of CCH patients from six hospitals in Madrid. Epidemiological and clinical variables were compared between patients with ONS failure and the rest. ONS failure was defined as the need for device withdrawal or switch off because of lack of response or adverse events.
Results
From a series of 88 CCH, 26 (29.6%) underwent ONS surgery, of whom 13/26 (50.0%) failed because lack of response. ONS failure group had an earlier headache onset (mean ± SD) of 27.7 ± 6.9 vs. 36.7 ± 11.8 years, p = 0.026) and a higher smoking rate (100% vs. 42.9%, p = 0.006). Stational fluctuations (58.3% vs. 7.7%, p = 0.007) and nocturnal exacerbations (91.7% vs. 53.9%, p = 0.035) were more frequent in the ONS failure group as well. There was no difference between groups in diagnostic delay, years of evolution prior to surgery, mental illness, comorbidity with other headache disorders or chronic pain conditions or prior response to occipital nerves anesthetic blocks.
Conclusions
Some clinical features such as an early debut, smoking and seasonal or circadian fluctuations could be related to failure of ONS in refractory CCH.
Keywords
Introduction
Occipital nerve stimulation (ONS) is an electric neuromodulation technique intended to treat headache disorders and cranial neuralgias in which the occipital nerves are implied in the pathophysiology (1). Chronic cluster headache (CCH) is a relatively uncommon primary headache disorder (2) for which management is often challenging in headache clinics. In 2014, The European Headache Federation (EHF) defined the diagnostic criteria for refractory CCH (rCCH), to reach a consensus on the classification of patients who do not respond to conventional therapies. These criteria include CCH or probable CCH diagnosis as defined by the International Classification of Headache Disorders, 3rd edition (ICHD-3) with at least three attacks per week that impact the patient's quality of life despite preventive or symptomatic treatment, in which at least three preventive drugs have failed when used at the maximum tolerated dose over a sufficient time. Secondary headaches mimicking CCH needs to be ruled out by negative investigation with brain magnetic resonance imaging. The preventive trials need to be drugs with good clinical experience for the prevention of CCH and/or have shown efficacy over placebo in randomized trials. The same EHF consensus document states that these include verapamil, lithium, oral or intravenous steroids, greater occipital nerve anesthetic block (GON block) with local corticoids, topiramate, methysergide, ergots, civamide and long-acting triptans (3). ONS has been postulated as a preventive therapy for rCCH. A systematic review and meta-analysis on the preventive treatment of rCCH concludes that, with data from more than 400 CCH patients who underwent therapy, ONS is the most studied therapy in CCH to date, with an aggregated response rate of 57.3% (95% confidence interval = 0.481–0.665, p < 0.001, I2 = 68.5%) (4). The largest ONS study in rCCH is a randomized, electrical dose-controlled clinical trial of 150 rCCH patients (ICON trial) (5). All patients received active neuromodulation with different protocols. A significant improvement in attack frequency was reported in all groups. More recently, long-term efficacy and safety data of ICON have been published, showing sustained benefit with good tolerability at up to five years of follow-up (6). Placebo effect cannot be ruled out in this study because a sham-controlled trial was not possible given the device technology available at the time. Nevertheless, all of the benefit could not be attributed to placebo effect because a sustained long-term response was demonstrated (5,6). On the other hand, this treatment is not without adverse events (AE), which make reintervention surgery necessary in case of local complications such as infection or lead migration, with the serious AE rate having been described as frequent as 23% in the largest studies (5). Furthermore, the cost associated with the therapy are as high as €28,186 per case treated (7) and new pharmacological therapies are being investigated in clinical trials, such as erenumab (NCT04970355), eptinezumab (NCT05064397) and rimegepant (NCT05264714), which may change the state of the art in the treatment of CCH in the near future. Galcanezumab, a monoclonal antibody targeting calcitonin-gene related peptide, was investigated for CCH in a randomized clinical trial, without achieving statistically significant improvement against placebo (8). However, some case series of rCCH patients report improvement with galcanezumab (4,9), and so this treatment is sometimes used on a compassionate basis. Despite the evidence supporting ONS, for the above reasons, the new European guidelines only recommend it as a third option (after non-invasive vagus nerve stimulation and sphenopalatine ganglion stimulation), before considering deep brain stimulation (10).
For all these reasons, it is of interest to headache specialist and functional neurosurgeons to investigate clinical predictors of ONS response or failure. Our objective in the present study is to analyze clinical factors that may be related to ONS failure in a sample of CCH patients treated with ONS.
Methods
Study design
We conducted a cross-sectional study through a review of the medical records of CCH patients in six public hospitals in Madrid (including five third-level hospitals), Spain (CLUSTER-MAD Registry). All the headache clinic physicians in the Headache Study Group of the Madrid Neurology Association were invited to participate to obtain the most accurate sample possible of patients diagnosed with CCH in this territory. A subanalysis on patients who underwent ONS surgery was later performed. This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.
Participants
All patients included were diagnosed as CCH by a headache specialist, according to the ICHD-3 and fulfilled the European Headache Federation diagnostic definition of refractory CCH. If they were diagnosed under the criteria of a previous version of the ICHD, they could only be included in the registry if they also met the diagnostic criteria according to the most recent classification. All patients had brain magnetic resonance imaging performed, with results within normal limits. Of these patients, we selected those who have had a definitive ONS implanted, following the recommendations of the Spanish Society of Neurology, with a follow-up of at least six months.
ONS selection and follow-up
Patients were considered for ONS if they met the rCCH EHF criteria (i.e. all patients selected for ONS met the rCCH criteria before this therapy). Surgery could be performed before or after trying other therapeutic alternatives following the usual clinical practice of the different centers that participated in the registry. Unresponsiveness GON blocks including local corticoids was not considered as an exclusion criterion for surgery, as some evidence suggest that it may not be predict ONS success (11). ONS was only indicated for CCH management, and not for other possible comorbid headache disorder such as migraine, although the presence of these was not an exclusion criterion. The objective of ONS is to generate relief in the patient by creating an electrical field. The paresthesia produced by stimulating the occipital nerves can be maintained and adjusted by altering the stimulation. All of the CCH patients received tonic stimulation and none burst. The stimulation parameters were adjusted individually, looking for active paresthesia in the cutaneous territory of the greater and lesser occipital nerves (notorious but under the threshold of dysesthesia or painful sensation). Amplitude was set according to the tolerance and sensation of the patients, pulse width could vary from 180 to 450 μs and frequency from 40 to 130 Hz.
Data collection
Data were collected through a survey in REDCap (Research Electronic Data Capture) (https://www.project-redcap.org) from January 2022 to May 2022. The collaborators had to provide data on all patients diagnosed with CCH or probable CCH (according to the ICHD-3 criteria) attending their headache clinic and with active follow up at the moment. The data were collected by a retrospective analysis of the medical records of these patients. The survey included data on family history, comorbidities, clinical presentation, refractoriness criteria, therapies used in the past, currently active treatments and clinical outcome at the time of data collection. ONS failure was defined as the need for device withdrawal or permanent switch-off because of lack of response or AE, at the time of data collection. Characteristics of ONS failure group and active ONS group were compared.
Statistical analysis
No statistical power calculation was performed before the study. Nominal variables were reported as percentages and compared using a chi-squared or Fisher's exact test, when applicable. A two-tailed Kolmogorov–Smirnov test was applied to examine whether the ratio variables followed a Gaussian distribution. Ratio variables were reported as the mean ± SD if they followed a Gaussian distribution; otherwise, they were represented as the median ± interquartile range. A two-tailed Mann–Whitney U-test was used to compare the differences between ordinal variables in two independent groups. p < 0.05 was considered statistically significant. Data analysis was performed using SPSS, version 23.00 (IBM Corp, Armonk, NY, USA).
Ethical approval
This project has been evaluated and approved by the local Ethics Committee for Clinical Research of La Paz University Hospital, Madrid (project code: PI-5575).
Results
In total, 88 CCH patients were included in the CLUSTER-MAD registry. ONS was performed in 29.6% (26/88). All of them met rCCH criteria. At the time of data collection, 50.0% (13/26) still receive active stimulation because of improvement attributed to this therapy, the remaining 50.0% were classified as ONS failure. In both groups, 3/13 (23.1%) were implanted with a unilateral device ipsilateral to the headache, the rest were implanted with bilateral ONS. Mean ± SD follow-up was 7.5 ± 4.6 years. In all cases, ONS failure happened because of lack of significative improvement. Three serious AE were reported, with the need of a second surgery (one local infection and two lead migrations), all of them in the active ONS group.
Demographical data and comorbidities
Onset of headache in ONS failure group was earlier than active ONS group, both cluster headache (CH) onset (27.7 ± 6.9 years vs. 36.6 ± 11.8 years, p = 0.026) and chronification or CCH onset (33.2 ± 6.5 years vs. 43.2 ± 9.7 years, p = 0.005). There were no significative differences in diagnostic delay (6.3 ± 7.7 years in ONS failure group vs. 6.0 ± 9.0 years in active ONS group, p = 0.927) or time to surgery since headache onset (11.1 ± 7.3 years vs. 12.5 ± 9.1 years, p = 0.668). Active smoking was seen in 76.9% (10/13) of ONS failure group, whereas 23.1% (3/13) active ONS group smoked (p = 0.006). There was no difference in the use of other substances. Finally, no differences were found in the rate of psychiatric disease, comorbid headache disorders or other chronic pain conditions, but a tendency to occur more commonly in the ONS failure group was observed. Chronic migraine was a relatively common comorbidity (1/13 in active ONS group, 4/13 in ONS failure group, p = 0.395), these patients received onabotulinum toxin A or topiramate and improved the frequency of migraine attacks significantly, but CH attacks remained active and thus the indication of ONS was made. Further information is provided in Table 1.
Demographical and comorbidity data in non-refractory and refractory chronic cluster headache.
ONS = occipital nerve stimulation; CH = cluster headache; CCH = chronic cluster headache.
Clinical presentation
ONS failure group presented more frequently as CCH since the beginning but the difference did not achieve statistical signification (53.9% (7/13) vs. 30.8% (4/13), p = 0.234), in contrast to active ONS group more frequently developing CCH from episodic CH. Both groups presented a clinical course with absence of remission periods in 76.9% (10/13) in contrast with a course in the form of active periods alternating with remissions (of less than 3 months). By contrast, there was differences in the presence of fluctuations or exacerbations (seasonal and nocturnal) both being more frequent in ONS failure group: seasonal exacerbations appeared in 58.3% (7/12) of ONS failure group vs. 7.7% (1/12) of active ONS group (p = 0.007) and nocturnal exacerbations in 91.7% (11/12) of ONS failure group vs. 53.9% (7/13) of active ONS group (p = 0.035). Both nocturnal and stational fluctuations in attack frequency relate to the clinical course since the onset of CCH, before ONS and most other preventive treatments were implemented. The response to GON block prior to ONS surgery was also analyzed, with a higher response to GON block in ONS failure group (40.0% (4/10) vs. 18.2% (2/11)) but without reaching statistical signification (p = 0.251). Further information is provided in Table 2.
Clinical presentation of non-refractory and refractory chronic cluster headache.
ONS = occipital nerve stimulation; ICHD-3 = International Classification of Headache Disorders, 3rd edition; CCH = cluster headache; ECH = episodic cluster headache; GON = greater occipital nerve.
All patients tried subcutaneous sumatriptan 6 mg.
Concomitant treatments
All patients met rCCH criteria prior to ONS, meaning they remained with at least three debilitant CH attacks per week of life despite at least three preventive drug trials with sufficient evidence for CCH (mainly verapamil, lithium, topiramate and/or systemic or local corticoids administered by GON block). Many therapies were used as off-label pharmacological therapies, such as gabapentin, pregabalin, carbamazepine, onabotulinum toxin A and endovenous ketamine. In the 13/26 patients improving after ONS, the response was attributed to this therapy because no other treatment was initiated or modified until ONS response could be evaluated. In these 13/26 patients, all pharmacological therapies could be decreased and 6/13 of them could retire preventive drugs completely. In patients with comorbid chronic migraine, topiramate and onabotulinum toxin A were maintained for control of migraine attacks. More information is provided in the Supplementary Material.
Discusion
In this series of rCCH patients who underwent ONS surgery, half of the cases classified as ONS failure. Clinical factors related to the ONS failure state were a younger age at debut, smoking and fluctuations in attack frequency with seasonal and nocturnal exacerbations.
CH is considered to have a genetic basis (12) in which environmental factors play a role as precipitants in subjects at risk (13). Alcohol and other vasodilators such as histamine and nitroglycerine are the most well-known substances that can trigger CH attacks. Tobacco has also been linked to CH, first when higher smoking rates in CH patients than controls were observed (14). A previous study has described that CH smokers have an earlier onset and with greater seasonal fluctuations than CH who do not smoke (15), so that it is reminiscent of our results.
CH attacks are considered to have their origin in a hypothalamic disfunction, as the circadian (nocturnal exacerbation) and supradian course (seasonal fluctuations) appear to indicate (13). Information from light travels by in the retinohypothalamic pathway, modulating hypothalamic activity via the suprachiasmatic nucleus. In the day and in the sunny seasons of the year, suprachiasmatic firing increases. CH attacks are typically more frequent at night and in the seasons with less daylight (16). Nevertheless, not all patients report seasonal fluctuations (from 37% to 80.5%) (15,17–23). Series from the same region also differ (19–22), and so this situation should be caused by factors other than geographical conditions.
With regard to ONS, the neuroanatomical basis for its therapeutic effect lies in the trigeminocervical complex (TCC), the functional unit formed by the spinal nucleus of the trigeminal nerve and the neurons in the dorsal horn of the spinal segments C1–C3 (24). The spinal nucleus of the trigeminal nerve receives afferences from the ipsilateral posterior hypothalamus, which is the reason for CH attacks having a trigeminal distribution. ONS active sensory non-nociceptive Aβ-fibers of the greater and lesser occipital nerves, which are tributary of C2–C3. In the dorsal horn of those spinal segments, these fibers modulate excitatory glutamatergic interneurons that receive information from nociceptive Aδ and C-fibers of the same territory as well. Because of these neural relations, the stimulation of Aβ-fibers via ONS (or other therapies such as GON block) are considered to be able to modulate TCC activity (25). This has been confirmed in vivo in humans in a double-blind placebo-controlled study of individuals undergoing GON block and then addressing the TCC activity using functional neuroimaging, revealing that GON block but not placebo significantly reduced nociceptive trigeminal activation (24). The neuromodulation of TCC can have benefits may be related to subsequential modulation of brain areas relevant in CH pathophysiology. Another functional imaging study reveals hypermetabolic brain regions of the central pain matrix in CCH that normalize after improving with ONS (26).
In the present study, nocturnal and seasonal exacerbations were more frequent in patients who failed ONS therapy. Not all patients of CH report fluctuations, and those who did were the group that responded better to the therapy. The cases of CH fluctuating independently of the season of the year or the day–night cycle or that have a high attack load without fluctuations may correspond to a permanent hypothalamic disfunction and the participation or other cerebral regions with implication in pain processing (central sensitization). because the ONS mechanism is related to modulation of TCC and other brain areas of the central pain matrix, these may be the cases that benefit most from ONS. By contrast, ONS may be less effective in CCH patients who are more likely to fluctuate because of environmental factors such as light (i.e. nocturnal and seasonal fluctuations), who are exposed to nicotine or who have more genetic risk (that could condition onset time). One of the main limitations of this therapy is the lack of improvement in hypothalamic regions (as revealed in the previously cited study) (26), which probably plays an important role in the correct selection of patients according to our results.
Previous studies have not found clinical predictors of responsiveness or failure to ONS in CCH. The ICON randomized controlled trial (5) and its extension phase (6) include an analysis for this matter, without significant results. The variables included in the analysis (that can be consulted in the articles and their appendices of the trial) are age, sex, number of years with CH, attack frequency at baseline, and the presence of autonomic symptoms and restlessness. Regarding the classic predominance of men, we found a slightly higher proportion of women in the group of non-responders to ONS, but no significant differences. Several studies have shown that there are phenotypic, clinical and demographic differences with respect to sex in these patients and, if we had a larger sample, we might have found differences by sex (27–29). Previous investigations did not find predictive value of other items such as previous response to GON block (11). In the present study, we analyzed the same variables with consisted results; however, to the best of our knowledge, this is the first study to include a wide range of epidemiological data, comorbidities and clinical variables in the intended analysis of predictors. An interesting recent review found discrepancies in the anatomical locations and corresponding landmarks of the greater and lesser occipital nerve. The surgical approaches differed in patient positioning, electrode placement, imaging techniques and overall efficacy (30) The lack of standardized implantation common to all CH patients is also likely to be a key factor in the different response rates between patients and centers.
The strengths of the present study include the multicentric design, having as collaborators neurologists of different assistance level hospitals (up to headache clinics of third level hospitals with functional neurosurgery teams) and the focus on clinical aspects of clinical CH presentation that are not always reported in clinical series. As a main limitation, treatment data were collected using the failure or discontinuation of the therapies rather than the response rate. We decided this was the most favorable decision taking into account that the registry did not only collect information about ONS, but on all the therapies tried in the patient's history, and so a specific response rate definition (i.e. decrease >50% in CH attacks) would have been difficult to collect in a retrospective review of medical records. In the specific case of ONS, the clinical decision of device removal or switch-off (“ONS failure”) is a non-ambiguous and objective way to indicate which patients did not benefit from the treatment, making it useful for the intended analysis. From a clinical standpoint, both device removal and switch-off correspond to the need to consider other treatments, and so both were considered as ONS failure. As previously mentioned, there are important anatomical variations in ONS implantation among neurosurgeons, making it difficult to evaluate a homogeneous response. However, it should be noted that, of the six hospitals participating in this study, all devices were implanted in two hospitals with extensive experience in neuromodulation. Finally, a prospective design could have been more adequate in the investigation of response predictors, and the small sample size of the implanted patients can make some of the analysis underpowered.
Conclusions
Some clinical features such as an early debut, smoking, and seasonal or circadian fluctuations could be related to failure of occipital nerve stimulation in refractory chronic cluster headache. These data could be relevant for a correct selection of patients prior to the implementation of these therapies but research is needed to confirm these findings.
In this series of refractory chronic cluster headache patients who underwent occipital nerve stimulation surgery, half of the cases classified as occipital nerve stimulation failure. Occipital nerve stimulation may be less effective in chronic cluster headache patients who are more likely to fluctuate with circadian or seasonal patterns, who are exposed to nicotine or who have a younger headache onset. By contrast, patients with a headache course without fluctuations benefit most from occipital nerve stimulation. This fact may be related to dysfunction of the central pain matrix. This treatment induce a modulatory effect over this network.
Supplemental Material
sj-docx-1-cep-10.1177_03331024241254078 - Supplemental material for Clinical predictors of therapeutic failure of occipital nerve stimulation in refractory chronic cluster headache
Supplemental material, sj-docx-1-cep-10.1177_03331024241254078 for Clinical predictors of therapeutic failure of occipital nerve stimulation in refractory chronic cluster headache by Javier A Membrilla, María-Luz Cuadrado, Nuria González-García, Jesús Porta-Etessam, Antonio Sánchez-Soblechero, Alberto Lozano Ros, Alicia Gonzalez-Martinez, Ana Beatriz Gago-Veiga, Sonia Quintas, Jaime S Rodríguez Vico, Alex Jaimes, Lucía Llorente Ayuso, Javier Roa, Carlos Estebas and Javier Díaz-de-Terán in Cephalalgia
Footnotes
Acknowledgements
We thank Íñigo de Lorenzo for his unconditional support.
Declaration of conflicting interests
JAM has received honoraria as a speaker from TEVA and Novartis and for advisory boards from Pfizer. AGM has received speaker honoraria from TEVA. ABGV has received honoraria for advisory boards and/or speaker panels from Novartis, Lilly, TEVA, Abbvie-Allergan, Lundbeck and Pfizer. JDT has received honoraria for advisory boards from speaker panels from Novartis, Lilly, TEVA, Abbvie-Allergan, Lundbeck Pfizer and Boston Scientific.
Ethical statement
This project has been evaluated and approved by the local Ethics Committee for Clinical Research of La Paz University Hospital, Madrid (project code: PI-5575).
Funding
Funding for article processing charges were provided by Boston Scientific. Boston Scientific has not been involved in the design, conduction or writing this nor any related manuscripts.
ORCID iDs
Supplemental material
Supplemental material for this article is available online.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
