Abstracts

RATIONALE: To compare the rates of infections among patients participating in double-blind efficacy and safety clinical trials of new antiepileptic drugs. METHODS: Data from placebo-controlled clinical trials listed in the drug label were evaluated to determine the rates of various types of infections. Adverse events reported by patients were converted to standard dictionary terms (e.g., COSTART, WHO-ART) by the sponsor s coding team. Each team used different standards for selection of the reporting terminology. Data for the placebo group were compared with data from the active drug group. RESULTS: In most populations, the incidence of various types of infections was similar in the active drug and placebo groups. Significantly higher rates of common colds and upper respiratory infections (URI) were listed for levetiracetam (13% vs 7% placebo); rhinitis for lamotrigine (14% vs 9%) and oxcarbazepine (11% vs 4%); vaginitis (lamotrigine 4.1% vs. 0.5%), pharyngitis (topiramate 7.1% vs 2.9%; tiagabine 7% vs 4%), and sinusitis (felbamate 3.5% vs 0). In some instances, placebo groups reported more problems than drug groups (tiagabine infection 13% vs 15% placebo). Problems reported by LEV patients as common colds or URI were coded as infection whereas the terms flu-like syndrome, respiratory disorder were used for other drugs. Levetiracetam responders who reported common colds or URI had higher QOLIE-31 scores for social functioning, suggesting that reduced seizure frequency allowed greater social activity and exposure to common colds and URI in the community. CONCLUSIONS: Reports of various types of infections were highly variable in clinical trials. Differences among drugs might be attributable to the coding styles of adverse event reporting using standardized dictionaries. The lack of uniform reporting makes it difficult to attribute infections to a specific drug or to compare infection rates among drugs.