RESUMO
OBJECTIVE: To examine serologic markers of vascular risk under treatment with commonly used antiepileptic drugs (AEDs) in the elderly in a randomized setting, and to determine whether the reduced exposure to hydroxymethylglutaryl-CoA reductase inhibitors ("statins") caused by carbamazepine reduces the effectiveness of the drugs as lipid-lowering agents. METHODS: Standard lipid fractions, lipoprotein(a), and C-reactive protein (CRP) were examined in a subset of those participating in the STEP-ONE trial, in which elderly patients with new epilepsy were randomized to treatment with carbamazepine, lamotrigine, or levetiracetam. Separate comparisons were made by individual AED, among those treated with statins, and, for CRP, among those treated with anti-inflammatory drugs. RESULTS: One hundred ninety-four patients had the aforementioned serologic measurements. In patients not taking statins, those treated with carbamazepine had higher total cholesterol than those treated with levetiracetam (+16.6 mg/dL, P = 0.053), with values from patients on lamotrigine intermediate, whereas cholesterol fractions were subject to drug-gender interactions which did not show a consistent pattern. Lipoprotein(a) was significantly lower in lamotrigine patients than in the carbamazepine and levetiracetam groups. After accounting for the effects of steroids, CRP was higher in carbamazepine patients than in other patients. Patients taking a statin had lower lipid levels than those not taking a statin regardless of AED, but the differences between statin-treated and non-statin-treated patients were much larger (50%-100% or more) in the lamotrigine and levetiracetam groups than in the carbamazepine group (P = 0.035 for interaction effect of statin use and AED on total cholesterol). SIGNIFICANCE: Here, we demonstrate that carbamazepine significantly interferes with the ability of statins to lower total cholesterol, thus making it a poor choice for hyperlipidemic patients or those with cardiovascular disease. Native AED effects on lipids were inconsistent and subject to drug-gender interaction, in contrast with other studies; further investigation is necessary to determine if these latter findings are genuine or spurious.