Studies of drugs in epilepsy cited by author are not evidence based

EDITOR—We were surprised to read Brodie's letter claiming the existence of “hard empirical evidence” to support a mechanistic approach in the management of epilepsy.1 He refers to an unpublished study in which patients with partial seizures resistant to monotherapy with carbamazepine (a sodium channel blocker) were randomised to take additional valproate or vigabatrin (drugs with GABA-ergic mechanisms).2 In those patients who responded to dual therapy, withdrawal of carbamazepine was attempted, with a view to achieving monotherapy with either valproate or vigabatrin. Altogether 7% became seizure free with valproate or vigabatrin alone, while a further 14% became seizure free with dual therapy. Because this study did not contain a placebo group or a group not receiving treatment, the results cannot provide reliable evidence that patients not responding to a sodium channel drug may respond to a drug with a GABA-ergic mechanism. Similarly, because the study did not contain a group allocated to take a second sodium channel drug (for example, lamotrigine), it cannot provide evidence that patients failing to respond to a sodium channel blocker are more likely to respond to a GABA-ergic drug than to a second sodium channel blocker. For the same reasons, it cannot provide evidence to support the hypothesis of “rational polytherapy,” which deems it more rational to treat patients requiring polytherapy with drugs with differing mechanisms of action.

Brodie refers to a second open uncontrolled study, in which lamotrigine was added to carbamazepine, phenytoin, or valproate.3 Patients taking a combination of lamotrigine and valproate seemed to respond best. This study, however, is confounded by too many factors to allow any inference with respect to a synergistic action between these two drugs.

We agree with Brodie that the evidence underpinning the management of epilepsy must be clinically relevant and scientifically credible. Currently, however, no scientifically robust data are available to allow epilepsy specialists to adopt a mechanistic approach. At best, the evidence cited by Brodie generates hypotheses that need testing in large pragmatic studies.

If we are to adopt an evidence based approach to the management of epilepsy we need to have a clear understanding of what constitutes good and reliable evidence. Epilepsy specialists will also need ready access to that evidence, and systematic reviews produced by the Cochrane Epilepsy Group should provide a valuable up to date resource for those wishing to apply an evidence based approach to the management of epilepsy.