Efficacy of hypothalamic stimulation for chronic drug-resistant cluster headache

The study of Seijo et al. recently published in this journal (1) provides further convincing evidence that posterior hypothalamic stimulation can be an effective treatment for carefully selected patients with drug-resistant chronic cluster headache. All five patients had a severe condition that had not responded to either drugs or previous surgical treatments. A number of aspects of the study deserve closer examination.

Patient selection

All patients had chronic cluster headache as defined by the International Headache Society (IHS) (2). According to the IHS, cluster headache is chronic if, in the preceding year, the pain-free period does not exceed 1 month. Thus, the condition can be chronic if there are just two or three headaches per month. For this reason the IHS definition of chronic cluster headache appears inadequate for selecting patients for surgery. This should not surprise us because the IHS classification was not designed for this purpose. In fact, all the patients in Seijo et al.’s series had cluster headache lasting for at least 2 years, with daily attacks (mean five) over the preceding year (1,3). In addition, before undergoing hypothalamic implant, the patients were tested with all the prophylactics usually mentioned in cluster headache treatment guidelines, including steroids (4). The selection of completely drug-resistant patients (5) seems correct both from the ethical and cost points of view: public health services should not be expected to support the high costs of this experimental and invasive procedure unless all other therapeutic avenues have been explored.

Bilateral implants from the outset?

Another characteristic of Seijo et al.’s series is that patients received implants only on the pain side (1). However, in one case (20%) controlateral headache attacks developed after implantation (1). We have had similar experience, leading us to suggest implanting electrodes bilaterally in a single operation (6). This would avoid the need for a second craniotomy should headaches develop on the opposite side to the original attacks. A similar approach is used for the much less invasive peripheral stimulation of the greater occipital nerve (7). Further studies will help us understand the best option here.

Microrecordings and targeting

Seijo et al. noted that recordings at their target site showed no specific pattern and always had low bioelectrical activity (1). This is probably because the area targeted (slightly different from that in previous series) is poor in cell bodies and mainly composed of nerve fibres. Previous studies report more pronounced electrical activity (8–10). The efficacy of stimulation in a low activity brain area suggests an effect by current diffusion to other structures that may be near, or not so near (11). A lack of correlation between precise target position and efficacy has been elegantly demonstrated by Fontaine’s group (12,13).

Follow-up and efficacy

After almost 3 years, two of Seijo et al.’s patients remain totally pain-free, in two others the attack frequency and need for subcutaneous sumatriptan have decreased by over 90%, and in patient 5 the number of attacks has halved (1). Steroids have been discontinued in all cases. Headache relief was achieved after about 2 months of stimulation with repeated adjustments of stimulation settings (1).

The authors also found that reappearance of attacks after 18 months in patient 1 was resolved after replacing the broken electrode. Similarly, in patient 2 the attacks transiently worsened after the stimulator had been transiently switched off (1). The authors suggested that these findings confirm the efficacy of hypothalamic stimulation in chronic drug-resistant cluster headache, consistent with the findings of other studies (for a review see (14)). The high proportion of responders led them propose changing the target coordinates compared with previous studies (1). The modified coordinates may also reduce the risk of ventricular wall perforation, because the electrode is positioned slightly more distant from the wall of the third ventricle.

Further studies are required to confirm the validity of Seijo et al’s coordinates (1). In the mean time, we encourage the authors to keep us updated on this interesting series of patients.