Abstract
Prompted by the results of gabaergic drugs, such as valproate and topiramate, we performed this pilot study to assess the effect of gabapentin in cluster headache. Eight patients suffering from episodic cluster headache and four suffering from chronic cluster headache were studied. All of them had failed to respond to traditional prophylactic drugs. The design of the study was an open trial. The main parameter for effectiveness was the number of daily attacks. Gabapentin was given at the daily dosage of 900 mg. All patients were pain free after a maximum of 8 days after starting therapy, with a bout duration thus reduced to 16–40% of the average previous bouts (only applies to episodic cluster patients). We hypothesize that the gabaergic action of gabapentin, perhaps combined with other mechanisms, such as calcium channel blockade, may be responsible for its remarkable effects on cluster headache.
Keywords
Introduction
Cluster headache is a painful syndrome of unknown origin, affecting one side of the face with extremely high intensity stabbing pain, in short lasting attacks, which recur once or more times a day, usually clustered in periods of 15–30 days, with long lasting remissions.
It has an estimated prevalence varying from 0.05% to 0.8%, but an arguably realistic extrapolation indicates that up to 0.24% of the population may suffer from cluster headache (for a review see 1).
The distressing characteristic of pain, its frequent recurrence within 24 h and its appearance at night time may be devastating for the patient; hence administration of an effective prophylactic treatment is mandatory. Unfortunately, there are no available therapies with a predictable and reliable outcome. Approximately 70% of patients respond either to ergotamine, methysergide, pizotifen, lithium carbonate, corticosteroids or verapamil (for a review see 2). More recently introduced drugs for cluster headache are valproate (3–5) and topiramate (6), tested in a limited number of patients with somewhat promising results.
In our experience patients do often respond satisfactorily when treated for the first time, but after several bouts they become refractory to the drugs used. The choice of alternatives then becomes increasingly difficult. Further and more pressing problems develop when the episodic form of cluster evolves into a chronic one, an occurrence happening in about 12.9% of cases (7). We thought that the anticonvulsant drug gabapentin, with a demonstrated analgesic effect in various pain syndromes and very few side-effects and interactions (for a review see 8), could safely be tried in a group of patients who had become resistant to most of the above-mentioned therapies. This paper reports the results obtained in an open trial.
Patients and methods
Twelve patients agreed to take part in this pilot trial. Eight of them had been suffering from the episodic form of cluster headache for a period ranging from 3 to 30 years. Four of them were affected by the chronic form, with an illness duration from 1 to 8 years. Diagnosis was made according to classification by the International Headache Society (9). All subjects underwent neurological examination and an MRI study of the head and neck, which revealed no abnormalities. The main characteristics of the headache and a summary of the patient history are given in Table 1.
Clinical characteristics of patients before starting gabapentin
BD=average bout duration in days; 4 patients had the chronic form. NB=number of previous bouts of cluster headache that the patient may remember. IDY=total illness duration in years. DA=number of attacks per day. FT=Do the attacks occur at fixed times during day or night? NA = Are night attacks a constant feature? IP=intensity of pain, graded from 0 (no pain) to 3 (maximum pain). AS = Are autonomic signs, like miosis, reddening of the eye, tearing, eyelid swelling, present?
All patients had been treated with various prophylactic drugs in the past: ergotamine, lithium carbonate, verapamil, methysergide and valproate. Unfortunately, effectiveness of these preparations had faded away long since, and only high doses of corticosteroids could check the attacks in some patients. For this reason, no patient had been taking any preventive drug in the previous month before the trial. Only attack therapy was used, either in the 1-month pre-trial period or during the trial itself. Attack therapy could be either sumatriptan, ergotamine tartrate or oxygen inhalation.
Patients suffering from the episodic form of cluster headache started treatment within 4–8 days from the beginning of a new bout.
The design of the study was a pilot, open trial. Gabapentin was prescribed at the starting dosage of 100 mg t.i.d., to be increased in 3 days up to 300 mg t.i.d.
Patients with episodic cluster discontinued gabapentin after 60 days, whereas those suffering from the chronic form were instructed to continue gabapentin at 300 mg t.i.d. for 6 months.
Results
Table 2 summarizes the attack pattern for each patient immediately before and after starting therapy. The duration of bouts before therapy is given in column 2 (‘Bout’) only for patients suffering from episodic cluster headache, because this cannot be determined in patients with chronic cluster. Number and intensity of attacks were almost immediately affected by treatment, with complete remission after a maximum of 8 days. Although not reported in the table, it is worth mentioning that the timing of the attacks changed in some instances, before they abated completely. The last column of the table reports the total number of pain days in the present bout, given by the sum of pain days before commencing therapy plus the number of pain days after therapy was started. This allows immediate comparison with the usual length of bout for the same patient as reported in Table 1. One can easily notice that, whereas in the previous bouts the usual duration was never below 30 days, in the treated bout the maximum duration (as given in the final column in Table 2) was 13 days. A follow-up visit was performed 3 months after discontinuation of therapy in the cases with episodic cluster and showed that no relapse had occurred. All chronic cluster patients were also seen after 4 months since the start of therapy, which they were still taking as scheduled, and none of them reported new attacks.
Pattern of attacks immediately before and after starting treatment. Columns marked day 1 to day 8 report the number of attacks for each day
Bout=days of pain of the present bout before starting therapy with gabapentin. Not applicable for patients with chronic cluster. Chr=patient suffering from the chronic cluster form. NA = not applicable. TDP=total days of pain in the present bout, given by the sum of days of pain before starting therapy plus number of days previous to complete remission of attacks. Not applicable for patients with chronic cluster.
The only side-effect detectable in our series was drowsiness, reported by two patients. It was, however, mild enough not to prompt discontinuation or change of the treatment schedule.
Discussion
Spontaneous remissions may hamper judgement on the efficacy of cluster headache therapy. While spontaneous remission is out of question in the four patients with chronic cluster, we cannot rule out the chance that at least some of the episodic cluster patients may have been suffering from an unusually short bout. This is particularly true for cases 1, 7, 9, 10 and 12, who had only been having attacks for 4–6 days before starting therapy. Such a period is shorter than the 7 days recommended by the International Headache Society guidelines (10). We had been prompted to make an early start to the therapy by explicit requests from the patients who, in any case, could not remember having short bouts in the past. Taking into due account this last consideration, we are prone to believe that the dramatic shortening of bout duration (from 16% to 40% of the usual average, as can be inferred by comparing the ‘TDP’ column in Table 2 with the ‘BD’ column in Table 1) may be due to an actual therapeutic effect.
As this was the first time that gabapentin was used in cluster headache, the study was intended as a pilot, open design. A further reason that prompted us to carry out an open trial was the fact that the patients had not responded to other drugs. We considered it unethical, in the circumstances, to further delay a therapeutic attempt by introducing placebo. The pain of cluster headache is of very high intensity, coming in very definite attacks affecting the same site; these characteristics are shared by trigeminal neuralgia, and one may reasonably consider that the two conditions are not susceptible to a prolonged placebo effect. Thus we may safely assume that the results obtained in this trial are due to the effect of gabapentin.
Although the mechanism of action of gabapentin as an anti-epileptic drug is not fully identified, we do know that it interacts with Ca channels, and that it increases the synthesis of GABA within the central nervous system (11). Wheeler et al. (6) found that topiramate was helpful in treating cluster headache, and they credited such a result to a GABA-mediated action of the drug. On the other hand, calcium entry blockers are also effective in some cases of cluster headache (12). It is just possible that both mechanisms are involved in the action of gabapentin in cluster headache.
Footnotes
Acknowledgements
This work is part of the project ′Pathophysiology of cluster headache', supported by a grant from the Italian Ministry for University, Science and Technology.