Abstracts

Rationale: Published quality guidelines call for minimizing risk of pharmacokinetic (PK) interactions when prescribing antiepilepsy drugs (AEDs), especially for new-onset, incident epilepsy. This usually means avoidance of enzyme inducing drugs (EIAEDs). Older persons may be at more risk for interactions because of vulnerability to side effects and changes in metabolic function. We sought to quantify this problem and to identify undesirable drug combinations most often used in a racially-balanced older US Medicare population.Methods: We conducted a retrospective analysis of claims data for a 5% random sample of Medicare beneficiaries at least 66 years old in 2009 augmented for adequate representation of minorities. We identifed 8,789 with at least one diagnostic code for epilepsy (345.xx) or two or more for seizures (780.xx) during 2009 with none during 2008 plus prescriptions for at least 60 days of an AED. The index date was the date of the first diagnostic code. We identified the 60 most common maintenance prescription non-antiepilepsy drugs (NAEDs) filled within 60 days of the index date and reviewed literature to determine which combinations with EIAEDs were subject to interactions. We calculated the proportion of users of each NAED who had a prescription for these AEDs: phenytoin (PHT), phenobarbital or primidone, and carbamazepine. We did not assess PK interactions between AEDs or with nonprescription drugs, nor pharmacodynamic interactions.Results: The study population was 21% White, 58% Black, 12% Hispanic, and 2.2% Native American. Two-thirds were women and 24% were 85 years or older. From the literature, we identifed 87 drug pairs with potentially important interactions between NAEDs and EIAEDs. Most interactions lower the serum level of the NAED (69), but a few (18) can raise serum levels of either or both drugs. 25% of patients were taking at least one NAED around the index date which can interact with EIAEDs . Persons taking interacting drug pairs, for example, included the 19% of persons taking metoprolol and 18% taking simvastatin who were also taking PHT. Some of these prescription combinations may have had little clinical effect, but there were some potentially life-threatening pairs: 11% of this population was taking warfarin, and of these, 24% took an EIAED; among the 5.6% on prednisone, 28%, and among the 5.4% on digoxin, 30%%.Conclusions: Conclusions Among a racially diverse population of older Medicare recipients with incident epilepsy, 25% were prescribed at least one drug which can produce PK interactions if combined with an enzyme-inducing AED. The proportion of patients at risk is likely higher because we assessed only the 60 most commonly- prescribed NAEDs, excluding nonprescription drugs and interactions among AEDs. Interacting combinations may be unavoidable for medical or financial reasons, but increased knowledge and computerized warning systems may reduce frequency of use. We plan further analyses to examine factors influencing this practice, such as demographics of patients and providers.