Safety and efficacy of sphenopalatine ganglion stimulation for chronic cluster headache: a double-blind, randomised controlled trial

Background Chronic cluster headache is the most disabling form of cluster headache. The mainstay of treatment is attack prevention, but the available management options have little efficacy and are associated with substantial side-effects. In this study, we aimed to assess the safety and efficacy of sphenopalatine ganglion stimulation for treatment of chronic cluster headache. Methods We did a randomised, sham-controlled, parallel group, double-blind, safety and efficacy study at 21 headache centres in the USA. We recruited patients aged 22 years or older with chronic cluster headache, who reported a minimum of four cluster headache attacks per week that were unsuccessfully controlled by preventive treatments. Participants were randomly assigned (1:1) via an online adaptive randomisation procedure to either stimulation of the sphenopalatine ganglion or a sham control that delivered a cutaneous electrical stimulation. Patients and the clinical evaluator and surgeon were masked to group assignment. The primary efficacy endpoint, which was analysed with weighted generalised estimated equation logistic regression models, was the difference between groups in the proportion of stimulation-treated ipsilateral cluster attacks for which relief from pain was achieved 15 min after the start of stimulation without the use of acute drugs before that timepoint. Efficacy analyses were done in all patients who were implanted with a device and provided data for at least one treated attack during the 4-week experimental phase. Safety was assessed in all patients undergoing an implantation procedure up to the end of the open-label phase of the study, which followed the experimental phase. This trial is registered with ClinicalTrials.gov, number NCT02168764. Findings Between July 9, 2014, and Feb 14, 2017, 93 patients were enrolled and randomly assigned, 45 to the sphenopalatine ganglion stimulation group and 48 to the control group. 36 patients in the sphenopalatine ganglion stimulation group and 40 in the control group had at least one attack during the experimental phase and were included in efficacy analyses. The proportion of attacks for which pain relief was experienced at 15 min was 62·46% (95% CI 49·15–74·12) in the sphenopalatine ganglion stimulation group versus 38·87% (28·60–50·25) in the control group (odds ratio 2·62 [95% CI 1·28–5·34]; p=0·008). Nine serious adverse events were reported by the end of the open-label phase. Three of these serious adverse events were related to the implantation procedure (aspiration during intubation, nausea and vomiting, and venous injury or compromise). A fourth serious adverse event was an infection that was attributed to both the stimulation device and the implantation procedure. The other five serious adverse events were unrelated. There were no unanticipated serious adverse events. Interpretation Sphenopalatine ganglion stimulation seems efficacious and is well tolerated, and potentially offers an alternative approach to the treatment of chronic cluster headache. Further research is need to clarify its place in clinical practice. Funding Autonomic Technologies.