Abstracts

RATIONALE: New antiepilepsy drugs (AEDs) are commonly studied in adults with refractory partial epilepsy. Patients may exhibit a temporary response, colloquially termed a "honeymoon effect", but become resistant to the same AED with prolonged use. Such a pattern of response could bias data from short-term pharmaceutical trials. Few previous studies have reviewed the comparative characteristics of patients and drugs involved in this phenomenon. METHODS: We prospectively studied 80 consecutive adult partial epilepsy patients of the adult epilepsy clinic, with known baseline seizure frequency, who started adjunctive AED therapy. Patients were categorized as responding to the AED if there was an overall >50% reduction in seizure frequency by the end of the first 4 months of therapy, including titration time. Temporary response was defined as increase to at least baseline seizure frequency after initial 4 months. We compared temporary to consistent responders by age, sex, localization and etiology of seizure, as well as number and type of AEDs used. RESULTS: The population consisted of 43 males and 37 females, with a mean age of 39.2 (range 19-77). Temporary response was seen in 22 (27.5%). 54 had interictal or ictal EEGs with temporal abnormalities, (28 left, 19 right and 7 bilateral). The etiologies represented were: cortical dysplasia (12), post-traumatic (14),cerebrovascular (11) tumor (7), and idiopathic (36). There was no association of any of the above clinical characteristics with temporary response to the AED. The added AEDs associated with temporary response, in descending order, were: formulations of carbamazepine (12), phenytoin(5), lamotrigine(3), gabapentin(2). Topiramate, tiagabine and divalproex sodium were unassociated with temporary response. Fifteen (77%) temporary responders were on 4 AEDs during the reponse period. CONCLUSIONS: It is not uncommon to observe a "honeymoon effect" to short-term adjunctive AED therapy. Our study suggests that the type and number of AEDs may be more associated with this phenomenon than patient characteristics. Further studies are needed to determine if specific patient profiles can indicate susceptibility to a "honeymoon effect".