OnabotulinumtoxinA for Hemicrania Continua: open label experience in 9 patients

Patients receiving BoNT-A with the Headache Group at the National Hospital for Neurology and Neurosurgery were analyzed. Patients were diagnosed with HC in accordance to International Classification of Headache Disorder criteria (ICHD-3beta) [1]. All had unilateral headaches that had responded fully to an indomethacin trial (oral or intramuscular trials, detailed in Table 1). All patients were injected with BoNT-A as per the Phase 3 REsearch Evaluating Migraine Prophylaxis Therapy (PREEMPT) regime for chronic migraine, with patients having a modified regime (exclusion of occipital and cervical paraspinal sites) if they had an occipital nerve stimulator (ONS) in situ [6].

All data was collected prospectively with the use of headache diaries. Average monthly scores were calculated from a month pre-treatment and a three-month post-final treatment diary. Headache days were recorded as any day on which the subject recorded HC pain. Subjects were asked to score pain intensity on two scales: 1) pain free, mild, moderate and severe; and 2) verbal rating scale (VRS). Headache load (HAL) was calculated from diaries using the formula: ∑ (severity (VRS) × pain duration (hours). Disability scores consisting of Headache Impact Test (HIT-6), Migraine Disability Assessment (MIDAS) and Hospital Anxiety and Depression scale (HAD) were collected before and after treatment.

Responders to treatment were classified as those achieving a 50% or greater improvement in headache days rated as moderate to severe. Other outcomes included those achieving a 30% and 50% or greater improvement in HAL and 30% or more improvement in headache days rated as moderate to severe.

Median values pre- and post-treatment were compared using Wilcoxon Signed Ranks tests and a statistically significant result set at the 95% level (p = 0.05). Data was processed using IBM SPSS Version 22 for Windows.

The study was approved by Northwick Park Hospital Research Ethics Committee, Hampstead, London, and written consent obtained from all patients.

A total of nine patients with HC received treatment with BoNT-A, of whom six were females and three males (see Table 1). Median age at time of treatment was 48 years (19–61 years) and median duration of HC was 8 years (1–34 years). During exacerbations, migrainous features were present in six and autonomic features in all subjects. Three patients reported visual auras during exacerbations. Four patients had concomitant episodic migraine (EM) and one co-existent idiopathic stabbing headache. All patients were able to differentiate their co-existent headaches from HC and none of the co-existent headaches had responded to indomethacin trials.

Subjects had failed to respond to a median of seven previous treatments for HC. Two subjects had failed to respond to ONS and one was awaiting ONS implantation.

Reasons for treatment with BoNT-A are summarized in Table 1. Eight patients could not tolerate therapeutic doses of indomethacin due to gastro-intestinal (GI) side effects. Two patients complained of worsening of their EM with indomethacin doses required to suppress HC. Patients had a median of two treatments (range 2–6) with a median BoNT-A dose of 167 units (range 110–185 units) injected at each treatment.

The results of BoNT-A treatment are summarized in Table 2. Five subjects demonstrated a response of 50% or more in reduction of moderate or severe headache days to mild headache days or pain free and were classified as responders to treatment. Six subjects reported a 30% or greater response in reduction of moderate or severe headache days to mild headache days or pain free. The median reduction in total headache days was 90% (range 0–100) (p = 0.026) and in moderate to severe headache days 80% (range 0–100) (p = 0.012). Headache load showed a median improvement of 62% (range 0–100) with six patients demonstrating a 30% and 50% or more improvement. Significant improvements were also seen in average headache hours and average VRS (Table 2). The median subjective duration of response in the five responders post treatment was 11 weeks (range 6–20 weeks, mean 12 weeks). Five subjects reported a subjective benefit of 80% or more in their HC.

Four subjects were taking indomethacin prior to BoNT-A and all were able to stop regular use after treatment with two using indomethacin as required at a frequency of less than three times a month.

Headache disability scores showed a trend to improvement after BoNT-A (Table 3). HIT-6 showed a median change of 12 points (p = 0.069). This is above the three-point change suggestive of minimal clinical difference. MIDAS improved by a median of 51 points (p = 0.063).

Adverse events were reported in three subjects: one eyebrow ptosis, one frontalis over-activity and one transient worsening in headache before improvement was noted. All adverse events were rated as mild by patients and transient in nature.

This series is the largest so far of BoNT-A treatment for HC. Five out of nine patients showed a greater than 50% reduction in moderate or severe headache days to mild headache days or pain free with a median reduction in moderate and severe days of 80%. All four subjects taking daily indomethacin prior to treatment were able to stop regular use. Five patients reporting an 80% or more improvement in their HC and clinically significant improvements were seen in both HIT-6 and MIDAS.

The patient group had tried a median of seven previous preventatives and had suffered from HC for a median of 8 years at the time of BoNT-A treatment. The refractory nature of the group means that it is doubtful that our observations are due to spontaneous remission. Despite four patients reporting co-existing episodic migraine, all were clearly able to differentiate this from their HC. The phenotype of HC was secure in all subjects and all meet ICHD-3beta criteria including a complete response of their side-locked headache to an adequate indometacin trial. Although a number of subjects report migrainous symptoms associated with HC, this is an accepted feature commented on in epidemiological studies and the ICHD-3beta criteria [1,3]. Given that all patients were carefully phenotyped and could clearly differentiate their episodic migraine attacks from hemicrania continua taken together with the sparse evidence for the efficacy of botulinum toxin in episodic migraine, our data are consistent with a change in HC and not the co-existent episodic migraine.

This series is still small, and this must be considered when interpreting the results. Previous studies of BoNT-A have reported a significant placebo response and we cannot eliminate this as a potential confounding factor in our outcomes. However, the relatively high response rates taken together with the consistent efficacy of repeated BoNT-A injections and a mean duration of effect similar to that seen in other reports in TAC as well as chronic migraine suggest that the response to BoNT-A in this series cannot be attributed entirely to the placebo response [10,12,13].

The exact mechanisms by which BoNT-A produces therapeutic benefit remains unclear, but the neurotoxin is likely to function by multiple mechanisms, suppressing events associated with peripheral and central sensitization. Both migraine and TACs are believed to share a common pathophysiology comprising of the activation of the trigeminovascular system and involvement of neuroactive peptides such as calcitonin gene-related peptide (CGRP), vasoactive peptide (VIP) and glutamate [14]. Animal studies have provided evidence of BoNT-A suppressing nociception in peripheral trigeminovascular neurons and also suppressing CGRP and VIP release from these neurons [5] There is also data to support the hypothesis that the toxin may act via central mechanisms with studies showing retrograde axonal transport of active BoNT-A [15,16]. The potential target of BoNT-A in chronic migraine is the direct blockage of trigeminal neurons providing nociception to the head and face. Suppression of neuro-inflammatory mediator release leads to decreased activation of second-order neurons within the trigemino-cervical complex and brainstem. BoNT-A may therefore be assumed to exert its benefit by repressing the neuro-inflammatory mediators responsible for the maintenance of peripheral and central sensitization [17,18]. It is therefore possible that BoNT-A has a wider therapeutic potential than chronic migraine. It is interesting to speculate that the clinical and functional imaging similarities between migraine and HC may mean that BoNT-A has more of an impact in HC than the other TACs which are much more clinically distinct to migraine [19].

OnabotulinumtoxinA may be a promising alternative to Indomethacin in patients with HC who do not tolerate the drug. Treatment appears to be associated with a significant improvement in moderate to severe headache days and related disability. It may add another potential therapeutic agent for HC to the limited number available. However, further controlled studies are necessary to clarify the efficacy of BoNT-A in HC.