Background
Treatment of chronic cluster headache (cCH) is guided by the dual objectives of ending acute attacks and decreasing attack frequency. Current medical treatment for acute attacks includes the use of oxygen inhalation, subcutaneous or intranasal application of triptans, or intranasal lidocaine [
1‐
3]. Preventative medications usually consist of verapamil or lithium as first choice; alternatively or concurrently, steroids, topiramate, melatonin, long-acting triptans, and occipital nerve blocks are also commonly used as prophylactic treatments [
4‐
6]. The costs of such pharmaceutical treatments for cCH are substantial, and have previously been estimated to average €15,700 per year, with severely afflicted patients incurring even higher annual medication costs [
7].
Electrical stimulation of the sphenopalatine ganglion (SPG) has recently been proposed as an alternative treatment approach for cluster headache, relying on an implantable on-demand stimulator that is activated by the patients themselves (PULSANTE SPG Microstimulation Therapy; Autonomic Technologies Inc. (ATI); Redwood City, California, USA). The safety and effectiveness of the PULSANTE system was first investigated in the Pathway CH-1 study, a multicenter, randomized trial [
8]. This study found the therapy to have dual clinical benefits of acute pain relief and attack prevention, while demonstrating an acceptable safety profile comparable to similar surgical procedures. In recent years, further investigational and commercial experience with the system has been gained in the European Union and other markets [
9,
10]. In the EU, the PULSANTE system is available for the treatment of episodic and chronic cluster headache since obtaining CE mark approval in 2012. In the United States, a pivotal study is currently recruiting patients (NCT02168764).
Our objective in this study was to evaluate the long-term cost-effectiveness of SPG stimulation, as compared to medical management, based on the findings of the Pathway CH-1 study and a model-based extrapolation through 5 years. We chose the German healthcare system as the setting for this analysis because of the experience already gained with SPG stimulation in that market, and the general relevance of Germany as the largest European healthcare system.
Discussion
While cluster headache patients have benefitted from some meaningful advances in drug therapy since the 1990s, when injectable and intranasal triptans were introduced, no substantial improvements in pharmaceutical therapy have been made since. As a result, a sizable number of cCH attacks remain insufficiently treated, and drug therapy alone has been unable to achieve pain relief sufficiently. At the same time, currently used medications, both abortive and prophylactic, are associated with a number of potential severe side effects and tolerability concerns, including increased risk of cardiac events [
16,
17] and hypertension [
18].
Furthermore, [
16‐
20] injectable triptans are frequently associated with “triptan sensations,” characterized by distressing chest discomfort, palpitations, and flushing. Patients with more frequent cluster attacks per day are unable to use triptans because of the dose limit they can use in a 24 h period. In addition, triptan treatment is contraindicated for patients with cardiovascular disease, and around 10 % to 20 % of patients are not effectively treated by—or are resistant to—these therapies. High doses of preventive medications like verapamil have the potential to cause serious cardiac abnormalities, including bradycardia, heart blocks and severe hypotension needing aggressive and expensive interventions. Also, lithium is frequently not well tolerated and requires careful medical management of the patients. Topiramate is increasingly used but due to central side effects is not well tolerated by a substantial number of patients. Some patients respond only to corticosteroids, or need them frequently because other preventative medications are not effective throughout the therapy; in these patients severe side effects are unavoidable. Methysergide and injectable dihydroergotramine are rather effective in cluster headache therapy but currently unavailiable in most countries.
As a result, there is a continued need for improved or complementary treatment approaches for cCH patients. This pressing clinical need is further underscored by a recent study reporting the burden of cCH to patients and society: more than 50 % of cCH patients reported suffering symptoms of depression; 25 % were found to have suicidal tendencies; and about 25 % received some form of disability allowance [
21].
SPG stimulation using a permanent implantable neurostimulation system has recently been demonstrated to be a safe and effective treatment alternative that offers acute pain relief at effectiveness levels comparable to the latest drug therapies, while also having the beneficial effect of reducing the frequency of attacks [
8,
10]. To-date, it is the only therapy that has the potential of dual benefits of acute pain relief and a reduction in attack frequency, combined with a reduction in overall medication use [
10].
The costs of current drug treatments for cCH are staggering, averaging more than €12,000 per patient per year for acute medications alone. SPG stimulation has the potential to lower these costs, by reducing or eliminating the need for attack-ending medications, and by achieving preventive effects [
8‐
10].
At the same time, implantation of an SPG stimulation system is a significant one-time investment for payers upfront, averaging more than €30,000 for implantation, when the costs of the device and the implantation procedure are considered. Understanding the health-economic profiles of a drug-based treatment strategy versus an implantable neurostimulation strategy is therefore essential to inform decision-making by payers and providers.
Despite the upfront costs for the device and implantation, our analysis showed that SPG stimulation has the potential to be cost-effective in time frames as short as 3 years, and to reduce overall costs to payers in timeframes greater than 5 years, primarily through the reduced need for rescue medications, but also by means of its observed preventive effects.
The results of this study compare favorably to the cost-effectiveness findings for a number of well-accepted and currently reimbursed implantable neurostimulation systems used to treat other clinical conditions with high disease and cost burden, including deep brain stimulation for movement disorders [
22], spinal cord stimulation for failed back syndrome [
23], and cochlear implants [
24]. These comparable therapies have been found to be cost-effective, in that their increased upfront costs have been shown to be associated with improved outcomes that healthcare systems considered a worthy investment.
But none of those technologies were found to bring about cost savings at time horizons as short as those projected in our analysis of SPG stimulation. While the assessment methods used in these published studies resemble our own—in that the results were derived from model-based projections informed by shorter-term clinical trial results—the timeframes adopted for those analyses were a minimum of 5 years, and sometimes a lifetime horizon. The timeframe of 5 years used in our analysis is therefore conservative. Longer follow-up would have led to an even more favorable health-economic profile. This fact is worth noting in light of the relatively young age of cCH patients, as evidenced by the average age of 45 years for the subjects included in Pathway CH-1.
The results of our sensitivity analyses show effects on the cost-effectiveness profile of SPG stimulation that—directionally—would be expected. In the control cohort, higher baseline attack frequency leads to additional savings because more attacks need to be treated with medications. In the SPG stimulation cohort, higher SPG therapy effectiveness leads to overall cost savings at follow-up longer than 5 years, while lower therapy effectiveness reduces the cost-effectiveness, as evidenced by an increased ICER. However, it is noteworthy that SPG stimulation remained cost-effective or cost saving across all trial-informed scenarios, even when considering the lower bound of the 95 % confidence interval for therapy effectiveness reported in the Pathway CH-1 study.
Our study is subject to a number of limitations. First, the Pathway CH-1 study, while randomized and sham-controlled, included only 28 patients. As a result, our trial-informed base case assumption of 67.1 % pain relief will need to be confirmed by future studies with a larger sample size. Further, the CH-1-observed distribution suggests significant patient-to-patient variation in response to stimulation. In order to test the effect of this uncertainty, we ran scenario-analyses using the lower and upper bound of the 95 % confidence interval of this parameter (50.2 %–80.5 %). Even at the lower efficacy assumption, SPG stimulation was found cost-effective, with an ICER of €18,846 per QALY gained—well below commonly acknowledged willingness-to-pay thresholds. At the higher assumption, SPG stimulation was the dominating strategy, and was associated with overall savings of €3,262. Similar findings hold for the secondary effectiveness parameter, frequency response.
Second, our analysis assumes the primary treatment effect of pain relief observed in the Pathway CH-1 study is maintained over the full time horizon of the analysis, while the secondary treatment effect of frequency reduction is gradually declining. Even though follow-up data through 24 months are available and there is no evidence or clinical rationale that would support a contrary assumption, further follow-up data will be desirable to confirm our assumptions. The same holds for the assumed longevity of the device. Our scenarios of lower and higher treatment effectiveness—albeit constant—provide insight into the potential effects of gradual changes in primary pain relief over time, and of more or less pronounced reductions in frequency response.
Third, our assumptions about medication use in the control cohort were based on a prior publication reporting on drug usage in a cohort of cCH patients treated in a tertiary headache center. While these patients share mostly comparable characteristics with the Pathway CH-1 cohort, and more broadly with patients eligible for SPG stimulation, some variation might exist in actual drug usage. To account for this uncertainty, we considered two scenarios in sensitivity analyses that assumed drug utilization increased or reduced by 25 % from baseline, respectively. Further, our model assumed that attacks successfully treated with SPG stimulation would not require any attack-aborting medication, and that patients not successfully treated would revert to the standard regimen of attack-aborting drugs. While this assumption is in line with the Pathway CH-1 study protocol, individual patient utilization might differ.
Fourth, the control cohort used in our model was defined as sharing the same characteristics as the SPG stimulation cohort at baseline, both in terms of attack frequency and health-related quality of life, and maintaining these parameters constantly over the horizon of the analysis. While this assumption seems reasonable, some fluctuations in the patients’ severity of cCH might occur, leading to higher or lower costs and quality of life. However, these potential fluctuations could be expected to occur in both the simulated control cohort and the SPG stimulation cohort, reducing the impact of any such fluctuations on the results of this analysis.
Fifth, our analysis does not consider the potential beneficial health impact of reductions in drug utilization for patients receiving SPG stimulation. Reducing the use of triptans and other attack-ending medications by more than 50 % might lead to fewer drug-related side effects. Including these beneficial effects would have made the health-economic profile of SPG stimulation more favorable.
Sixth, our economic analysis is limited to direct medical cost, that is, costs directly related to medical treatments. The preventative effects, reduction in necessary medication use, and improvement in health-related quality of life demonstrated in Pathway CH-1, however, might lead to reduced absenteeism, increased productivity, and other societal benefits not accounted for in our analysis. Including these potential benefits – again – would have increased the overall savings associated with SPG stimulation and would have further improved the favorable health-economic profile of this therapy choice.
Finally, intangible benefits such as potential therapy-related reductions in suicidal ideations and improved family/social interactions are not considered in this analysis.
Competing interests
Jan B. Pietzsch, Ph.D.: Dr. Pietzsch is president, CEO, and shareholder of Wing Tech Inc., a technology consulting firm focusing on early-stage assessment of medical technologies. Wing Tech Inc. received consulting fees from Autonomic Technologies Inc. to develop the health-economic model used in this analysis.
Abigail Garner, M.S.: Ms. Garner worked as a consultant for Wing Tech Inc. on this project.
Charly Gaul, M.D.: Dr. Gaul received honoraria for participation in clinical trials, contribution to advisory boards or oral presentations from MSD, Berlin Chemie AG, St. Jude Medical, ATI, Electrocore, Complen Health GmbH, Allergan, Boehringer Ingelheim, Astellas, and Hormosan. He has no ownership interest and does not own stocks of any pharmaceutical company.
Arne May, M.D.: Dr. May is funded by the University Clinic of Hamburg and received unrestricted research support from LindeGas (RealFund) and scientific funds by the Deutsche Forschungsgemeinschaft (DFG) and European Framework (FP7), is or has been consultant or speaker for Pfizer, Bayer Vital, GSK, Allergan, ATI, MSD, Electrocore and Desitin and is an editorial board member of Cephalalgia, J Headache and Pain, European Neurology and Der Schmerz.
This work was initiated by the authors and Autonomic Technologies provided support to Wing Tech Inc. (JBP and AMG) for model development and necessary analyses. Autonomic Technologies had no influence on the results or manuscript development. The authors maintained the right to publish without approval of the funding source.
Authors’ contributions
JBP conceived of the study, participated in the study design and model construction, and drafted the manuscript. AM participated in the study design, critically reviewed the model and assumptions, and helped with drafting the manuscript. CG participated in the study design, critically reviewed the model and assumptions, and helped with drafting the manuscript. AMG conducted literature review and data searches and helped with drafting the manuscript. All authors read and approved the final manuscript.