Migraine prophylaxis

A number of new options have appeared in recent years. In brief, topiramate is now a thoroughly established agent, and is listed on the Pharmaceutical Benefits Scheme (PBS) as a second-line option, after propranolol and pizotifen. Candesartan is very well tolerated by most patients and is attractive when previous agents have produced unacceptable side effects. Lisinopril seems to have few advantages over candesartan, and gabapentin is expensive; neither drug is widely used for migraine in Australia. The depth of data supporting the effectiveness of candesartan, lisinopril and gabapentin is limited; replication of the results quoted below would enhance confidence in these drugs. Botulinum toxin is also expensive and doubts remain about its efficacy. However, its adverse event profile is excellent and, anecdotally (and from published case series), it seems that some patients respond dramatically.

This is a relatively new anticonvulsant medication, and there have been three large randomised controlled trials which have shown efficacy. One randomised study of 483 patients showed that mean number of migraine days per month was significantly reduced with topiramate at daily doses of 100 mg (by 2.1 days) and 200 mg (by 2.4 days) compared with placebo (1.1 days).11 This effect was noted within the first month of treatment and persisted for the 26-week trial period. There was a also a trend towards improvement in the 50 mg group in this study. Adverse events leading to cessation of topiramate therapy included paraesthesia, fatigue, nausea and weight loss.11 Almost identical results were found in another large study12 and a smaller one.13 These data were analysed by a Cochrane review and found to be convincing.14 A further large study found similar efficacy to propranolol.15 Continuing benefit was demonstrated when patients who had been treated with topiramate for 6 months were randomly assigned to continue topiramate therapy or placebo over the next 6 months; headache days per 4 weeks increased by 1.19 with placebo and 0.10 with topiramate.16 Of note is the consistently reported side effect of weight loss with topiramate, in contrast to the weight gain often seen with such established medications as sodium valproate and pizotifen.

In practice, in the past couple of years, before its PBS listing in Australia, topiramate has been an attractive choice when other older agents have failed, and especially when the patient is overweight. The PBS listing provides subsidised treatment for migraine prevention only when both pizotifen and β-blockers are contraindicated or not tolerated. This results in the bizarre situation that inefficacy of these previous treatments is not sufficient to qualify for PBS subsidy. The PBS decision is based on lack of scientific evidence that topiramate is effective when other agents have failed, despite widespread anecdotal experience that this is often the case. Consequently, topiramate may be used in Australia according to the PBS guidelines or prescribed for suitable patients outside PBS guidelines and without PBS subsidy.

Side effects are mostly dose-related, and are less troublesome when topiramate is used for migraine than for epilepsy, as the doses are typically lower (100 mg per day compared with 200–400 mg per day). Side effects include drowsiness, dizziness, somnolence and an unusual problem with word-finding, paraesthesiae in hands and feet, and reduced appetite and weight loss.

A randomised, double-blind, crossover study of candesartan (16 mg daily) versus placebo has been performed.17 The mean number of days with headache in 12 weeks (the primary endpoint) was 18.5 with placebo versus 13.6 with candesartan (P = 0.001; n = 57). Several secondary endpoints also favoured candesartan, and 32% of patients had a greater than 50% reduction in migraines while taking candesartan compared with the period taking placebo. The drug was well tolerated — adverse events were similar in the two periods. Candesartan is starting to be used more widely by Australian neurologists, some of whom are enthusiastic proponents.

An earlier study had shown a clear benefit from the angiotensin-converting enzyme (ACE) inhibitor lisinopril (at a dose of 20 mg).18 The number of days with headache was reduced by 17% compared with the placebo arm; this benefit was somewhat less than in the later trial of candesartan,17 and adverse events, including cough, were more often an issue. ACE inhibitors are not widely used in Australia for migraine prophylaxis.

The clinical efficacy of ACE and angiotensin II inhibitors supports a role for the renin–angiotensin system in migraine. The benefits of these drugs are thought not to derive from the effect on blood pressure, but perhaps modulation of receptors in the central nervous system.19

One group studied gabapentin, titrated to a dose of 2400 mg per day.20 An unusual statistical approach with analysis of “modified intention to treat” populations detracts from this study. After 12 weeks of treatment, the median 4-week migraine rate was 2.7 for the patients treated with gabapentin and 3.5 for those taking placebo (P = 0.006), down from 4.2 and 4.1, respectively, during the baseline period. Additionally, 26 of 56 patients (46.4%) receiving a stable dose of 2400 mg of gabapentin per day and five of 31 patients (16.1%) receiving placebo showed at least a 50% reduction in the 4-week migraine rate (P = 0.008). Adverse events considered to be drug-related (especially somnolence and dizziness) resulted in 13 of 98 gabapentin-treated patients (13.3%) and three of 45 placebo-treated patients (6.7%) withdrawing from the study.

The use of botulinum toxin for migraine prophylaxis is particularly controversial. Large case series have been reported, with several experienced and respected experts convinced that this treatment is effective, at least in selected patients. Indeed, botulinum toxin has become a standard treatment option in many centres.21 This is despite the fact that several high-quality trials have failed to show effectiveness of botulinum toxin. Discussion of the data and of the paradox offers one possible explanation that a subgroup of non-responders is so large as to render the published trials under-powered.22 Recent studies suggest that headaches described as “imploding” or “ocular” may respond while “exploding” headaches may not.23 Independent prospective verification of this hypothesis is required.

Personal experience (R J S) suggests that, in a cohort of difficult and otherwise unresponsive patients, about 10% have a dramatic response and another 30% have a worthwhile response. Some case series claim an even higher response rate.21

Chronic daily headache, defined as headache on more than 15 days per month with more than 4 hours of headache on each of these days, is common, occurring in 4% of the population.24 Most cases troublesome enough to be referred to headache clinics arise from migraine,25 with transformation over time from an episodic to a very frequent or daily pattern. This situation has been termed “transformed migraine” or “chronic migraine”; there are recently published criteria for the diagnosis of “chronic migraine”.26 There is controversy as to the cause of this transformation, with some experts believing it very often occurs because of medication overuse, resulting in “rebound” headaches, while others feel that the pattern generally evolves spontaneously with more frequent medication use as a consequence.27 The truth probably lies between these positions; headaches in some patients certainly reduce dramatically once overused medications are withdrawn, while in others, they do not. All experts would agree that an escalating pattern of migraine frequency requires active intervention, and this generally involves minimising the regular use of acute treatments, particularly certain agents, and aggressively pursuing prophylaxis. This approach has led to guidelines on or recommendations for the maximum desirable frequency of use of various acute agents (Box 4).28,29

Once a pattern of chronic daily headache with medication overuse is established, it is necessary to withdraw the offending agent. Numerous protocols to aid successful withdrawal by controlling the inevitable exacerbation of headache have been proposed, supported by case series, but high-quality trials are lacking. Outpatient withdrawal may be effective on occasions, but for patients overusing codeine or substantial amounts of ergotamine or triptans, inpatient management is preferred.28

Agents used to aid medication withdrawal have included analgesics, tranquilisers, neuroleptics, amitriptyline, naproxen and valproate. Naproxen was shown to be better than symptomatic treatment with antiemetics and analgesics.30

The usual approach to inpatient management has included the use of fluid replacement and intravenous dihydroergotamine.

Many women with migraine have menstrually associated migraine (MAM). A subgroup of patients have attacks that occur exclusively with or just before menses: this is called “true menstrual migraine” and occurs in about 15% of women who have migraines.

Patients in whom most of the disabling attacks occur in relation to menstruation may benefit from the strategy of “miniprophylaxis”, in which a preventive medication is used for about a week at the time of vulnerability. Drugs used in practice in this way have included oestrogens, non-steroidal anti-inflammatory drugs, ergotamine, dihydroergotamine, methysergide and magnesium,37 but there have been few well designed studies. Triptans have also been used: there is an open-label study of 20 women with MAM given oral sumatriptan (25 mg thrice daily) and a small double-blind, placebo-controlled study of naratriptan (1 mg and 2.5 mg daily).38 In both cases the findings were positive, but curiously only for the lower dose of naratriptan.

Migraine aura is now recognised as a primarily neural event, involving spreading depression of activity, with secondary reduction in cerebral blood flow rather than a vascular or vasospastic process. In some patients, the aura symptoms predominate, with mild or no headache, or may be unusually prolonged. Prevention of such episodes may require a different approach from that used for more typical migraine. The information on this point has, however, been anecdotal.

A 2001 survey of North American neurologists found that verapamil and valproate were the preferred treatments, with verapamil especially popular for prolonged aura.39 β-Blockers were rarely used, perhaps because of concerns about limited compensatory vasodilator capacitance.39