Topiramate for migraine prophylaxis: addressing the blood-brain barrier-related pharmacokinetic-pathophysiological disconnect

To the Editor: Professor Diener and colleagues present a pooled analysis of three large pivotal placebo-controlled trials and conclude that topiramate 100 mg/day is an effective clinically useful drug for migraine prevention (1). Neuromodula-tory actions on excitatory neurotransmitter receptors and voltage-gated neuronal ion channels are believed to be relevant to migraine pathogenesis and therapy. Although the results of this analysis are statistically significant in favour of topiramate (1), the biological significance of this analysis to migraine pathophysiology remains abstruse. Approximately 50% of the patients did not respond to topiramate prophylaxis (1), a limitation that must be underscored. Besides, there is no practical method to predict which patients within a migraine cohort are likely to respond to topiramate. Third, while topiramate was found superior to placebo (1), it was not found superior to propranolol, a negative facet that certainly merits mention in the report. That even the use of propranolol itself in migraine therapeutics is purely empirical is another issue that, despite its momentous implications, is hardly ever countenanced. Fourth, topiramate is not effective in preventing aura in migraine patients (2). Fifth, topiramate can prevent recurrences of episodic and chronic cluster headache (3), a primary headache variant in which brain cortical neuronal dysfunction has not been envisaged. Six, a double-blind, placebo-controlled, parallel study of lamotrigine, an inhibitor of neuronal release of the excitatory transmitter glutamate, in 53 migraine with aura and migraine without aura patients found the drug ineffective for prophylaxis (4). Seven, acetazolamide, another drug believed to exert a critical action on neuronal ion channels in migraine patients, has no direct action on the P/Q-type calcium channel (5) nor does it alter the penetration of sodium ions into the parietal cortex (6). The controlled clinical trial is designed to compare therapies using mathematical logic; clinical trials do not supplant the need to evolve the conceptual background (7). Above all, clinical trials must not be allowed to suspend biological disbelief. A remarkable conceptual disconnect prevails between drugs used empirically for migraine therapy and the pharmacokinetic absolute related to their brain penetrability (8,9). Migraine prophylactic agents, in contrast to migraine headache abortive agents, modulate primary pathophysiological processes; also, migraine aura abortive agents must be ameliorating the primary dysfunction in migraine. If the origin of migraine truly involves brain neuronal dysfunction, prophylactic and aura-aborting agents must necessarily readily cross the intact blood–brain barrier (BBB). Atenolol, a first-line migraine prophylactic agent, does not cross the intact BBB or critically influence any brain neuronal function; several other antimigraine agents with variable prophylactic effects also do not freely cross the intact BBB, including verapamil, nadolol or magnesium (7–9). Some therapists prefer to use verapamil prophylactically in patients with migraine aura, with or without headache, particularly when the auras are frequent or associated with hemiparesis (10). Verapamil also appears to be useful in the management of hemiplegic migraine, sporadic of familial (11), although the drug has no action on brain neuronal function (8,9). Nifedipine and verapamil do not freely cross the intact BBB but can instantaneously abort migraine aura (12). Further, epileptiform activity, paradoxically, occurs more commonly in patients with tension-type headache than in migraine patients; as reviewed by Bazil, both migraine and seizures in such patients can be controlled by nimodipine, verapamil or nifedipine, which do not readily penetrate the BBB (13). Next, the epileptogenic potential of tricyclic agents is well established; nevertheless, amitriptyline is a first-line migraine prophylactic agent. If both anticonvulsant and potentially convulsant (or pro-convulsant) agents that can decrease or increase brain cortical excitability, respectively, can prevent migraine attacks, a reorganisation of prevalent concepts is essential (9,13). Finally, peripheral frontotemporal application of nitroglycerine precipitates ipsilateral headache without involving brain neuronal dysfunction (14). If a low-efficiency migraine headache (not aura) preventive effect might indeed be attributed to topiramate, it must be qualified as being unpredictable, not superior to propranolol, and probably unrelated to alteration in brain neuronal glutamate metabolism or ion channel modulation. It is also premature to conclude that the limited migraine prophylactic action of topiramate is related to modulation of central or brain neuronal function. Serendipity rather than design continues to propel migraine therapeutics. At this stage of comprehension of migraine mechanisms, with empirical prophylactic management strategies varying widely from closure of patent foramen ovale (15) to pericranial injection of botulinum toxin (16), it is important to try to refine the theoretical basis of the diverse therapies currently available or under investigation.