Abstracts

Nonadherence is an important determinant of outcome with drug therapy. Drug concentrations are influenced by the extent of nonadherence; some drugs are expected to be more sensitive to missed doses than others. The primary factor affecting this sensitivity is the dosing schedule of the drug in relation to its pharmacokinetics. The objective of this simulation study was to quantitate changes in drug exposure in response to nonadherence to antiepileptic drug therapy. Pharmacokinetic parameters and standard doses of zonisamide (ZNS), oxcarbazepine (OXC), levetiracetam (LEV), topiramate (TPM) and carbamazepine (CBZ) were obtained from the literature. Induced and uninduced clearance (CL) values were used for ZNS, OXC and TPM simulations. In addition, CL values of +/- 1 standard deviation were identified to correspond to above and below average elimination. Simulations were based on a one-compartment model with first-order absorption and elimination using the pharmacokinetic program NONMEM. Three levels of nonadherence were built into simulations following the attainment of steady state. Concentrations were simulated following the omission of 1, 2 and 3 consecutive days of dosing to mimic nonadherence over a single day, a weekend, and a long weekend, respectively. Two additional days of full adherence followed the period of nonadherence in the simulation. In silico subjects were created for each drug, induced and uninduced, at above, below and at average CL, and at each level of nonadherence. An area-under-the-curve (AUC) was computed beginning with the first missed dose following steady state, continuing through the 1, 2 or 3 days of nonadherence, and ending after a second day of full adherence. This AUC was compared to the AUC that would have occurred if the subject were fully compliant. Reduction in AUC ranged from 15-33%, 25-50%, and 33-60% following missing 1, 2 and 3 days of dosing, respectively. CBZ, LEV, OXC and TPM were remarkably similar in their results. Because of its longer half-life, ZNS consistently produced the smallest changes. The effect of enzyme induction had a more notable adverse effect on drugs with the longer half-lives. Some degree of nonadherence is certain in the clinical setting. Drugs with longer half-lives will maintain relatively greater exposures during periods of nonadherence. However, it will also take relatively longer for concentrations to increase back to steady state if the usual maintenance dose is resumed. In such cases, loading doses may permit more rapid attainment of steady state. This study demonstrates that drugs with longer half-lives will be associated with smaller reductions in drug exposure in situations of occasional nonadherence. (Supported by NIH NINDS P50-NS16308 and a grant from Elan Pharmaceuticals.)