3 Evidence

Clinical evidence

The clinical evidence comprises 8 published studies

3.1 The published studies comprise:

  • 3 randomised trials: 2 sham-controlled (ACT1, Silberstein et al. 2016, and ACT2, Goadsby et al. 2018) and 1 open label (PREVA, Gaul et al. 2016)

  • 1 post-hoc analysis of a randomised trial (Gaul et al. 2017)

  • a pooled analysis of 2 randomised trials (de Coo, 2019)

  • 3 non-comparative cohort studies (Nesbitt et al. 2015, Marin et al. 2018, and Trimboli et al. 2018).

    The evidence includes 410 patients with cluster headache. For full details of the clinical evidence, see section 3 of the assessment report.

The evidence includes the preventative and acute use of gammaCore and treatment refractory cluster headache

3.2 gammaCore was used in different ways in the studies. In the ACT1 and ACT2 studies, gammaCore was used as an acute treatment only. In all the other studies, gammaCore was used as a preventative and an acute treatment. Two cohort studies (Marin et al. 2018 and Trimboli et al. 2018) reported on gammaCore used for people classified as refractory to 1 or more standard treatments for cluster headache. In all other studies gammaCore was used in addition to standard of care treatments.

The studies show that gammaCore can reduce the frequency of cluster headache attacks and the intensity of pain during an attack

3.3 The evidence for gammaCore comprises a small number of studies which include comparative, non-comparative and observational studies. The external assessment centre (EAC) noted, however, that large randomised trials are not likely to be possible given the very low prevalence of cluster headaches. All but one of the studies was sponsored by the company. The studies had short follow-up times so there is no evidence for the long-term benefits of using gammaCore. The trials showed that some but not all people benefit from using gammaCore. When gammaCore was used preventatively, it reduced the frequency of cluster headache attacks (Gaul et al. 2016). When gammaCore was used acutely it reduced the intensity of attacks (ACT1 and ACT2). Using gammaCore improved measures of quality of life (Gaul et al. 2016).

The degree of clinical benefit is uncertain

3.4 In the 2 sham-controlled trials (ACT1 and ACT2), gammaCore worked better as an acute treatment for people with episodic cluster headache than for people with chronic cluster headache. The EAC noted, however, that these studies were not powered to allow such sub-group analysis. The EAC concluded that the published evidence suggests that people with cluster headache may benefit from using gammaCore (either preventatively or acutely) although the degree of benefit is uncertain.

Adverse events are mild to moderate

3.5 The reported adverse events were mild to moderate in all studies. No-one stopped using gammaCore because of adverse events. The clinical experts told the EAC that gammaCore is safe and easy to use and that there was no need for safety monitoring. The EAC concluded that there is enough evidence that gammaCore does not cause any serious device-related adverse events in people with cluster headache. For full details of the adverse events, see section 4.8 about adverse effects.

Cost evidence

The cost evidence comprises 3 published studies

3.6 There were 3 published studies: Morris et al. 2016 used data from the PREVA trial, Mwamburi et al. 2017 used data from ACT 1 and ACT 2 and Pietzsch et al. 2015 used data from a randomised, sham-controlled study (Schoenen et al. 2013).

The company says using gammaCore saves £450 per person in the first year

3.7 The company created a de novo cost analysis using a Markov model with a 1‑month cycle and 1‑year time horizon. The model only considered people with chronic cluster headache and did not include anyone with episodic cluster headache. The model considered gammaCore in addition to standard care acute treatments (oxygen, zolmitriptan and sumatriptan). Preventative medication (for example, verapamil) was not included in the model because it was assumed to be the same in both treatment arms. Data from the PREVA trial was used in the model to estimate the proportion of people whose cluster headaches responded to treatment with gammaCore; this study defined a treatment response as cluster headache frequency reduced by 50% or more. For full details of the cost evidence, see section 4 of the assessment report.

3.8 The company's model, which was unchanged by the EAC, reported that using gammaCore in addition to standard care would lead to cost savings of £450 per patient in the first year.

Cost savings depend on a free 3‑month initial period and reduced sumatriptan use

3.9 In the model gammaCore is free for the first 3 months, and it assumes that only people whose condition responds to treatment with gammaCore continue to use it. This is called a trial; however, this refers to the user trialling the technology for 3 months and does not mean that people using gammaCore will be automatically enrolled in a clinical trial. The model included less acute medication use by people in the gammaCore responder arm, based on data from the PREVA trial. The EAC's sensitivity analyses showed that the cost savings associated with gammaCore depend on the initial 3‑month period being free of charge and less sumatriptan use in the gammaCore responder arm.

The cost analysis has limitations because of lack of evidence

3.10 The model does not include costs for:

  • inpatient, outpatient or primary care

  • psychological support

  • invasive surgical procedures such as implanted sphenopalatine nerve stimulators

  • unlicensed medications used for treatment refractory cluster headaches, which might be avoided if gammaCore treatment is successful.

    These costs could not be included in the model for gammaCore because there is no evidence to base assumptions on. The EAC noted that gammaCore is likely to lead to a reduced need for care resources, and their associated costs, because people who it is effective for will have fewer, less severe cluster headaches. The EAC also noted that the data underpinning the economic model is part of a single small data set and post-hoc analysis that was only partially based in the UK. However it did not identify any other data that could be used to improve the analysis.

  • National Institute for Health and Care Excellence (NICE)