Abstracts

Rationale: Adherence rates for chronic pediatric conditions are approximately 50%; however, rates of adherence vary by the measurement method. Few studies have documented adherence to antiepileptic drug (AED) therapy using multiple measures of adherence. The objectives of the current study were to: 1) document rates of AED adherence using both self-report and objective measures and 2) to examine the association between objective, electronically-monitored adherence and AED serum level. Self-reported adherence was hypothesized to be correlated, but significantly higher, than objective, electronically-monitored adherence. It was also hypothesized that patients who fell below the therapeutic AED serum levels would demonstrate lower objectively-measured adherence rates compared to patients within or above the therapeutic range. Methods: Participants included children with new-onset epilepsy and their caregivers. Adherence to treatment was electronically monitored with MEMS TrackCaps and a self-report questionnaire. Adherence was calculated for the week prior to a 4-month follow-up clinic visit. Serum levels for AEDs were obtained from medical chart review and were conducted during the clinic visit. Therapeutic levels were determined as follows based on laboratory protocol: carbamazapine 4-12 mcg/ml, valproic acid 50-100 mcg/ml, levetiracetam 5-45 mcg/ml, oxcarbazepine 15-35 mcg/ml, ethosuximide 40-100 mcg/ml. Patients were categorized into 3 categories: below, at, or above therapeutic level. Results: Participants included 63 children with new-onset epilepsy (Mean age = 7.1 years, 33% female, 73% Caucasian) and their caregivers. Fifty-seven percent of children were diagnosed with partial epilepsy, 32% with generalized epilepsy, and 11% with unclassified type. Mean one-week adherence for children with new-onset epilepsy was 95.4% (SD = 8.8) for self-report and 84% (SD = 18.7) for MEMS TrackCaps. Self-reported adherence was significantly correlated with MEMS TrackCaps (r = 0.40, p < .001). No significant differences were found by serum level group (e.g., below, at or above therapeutic level) for MEMS TrackCap adherence (F (2, 48) = 0.2, p = n.s.). Conclusions: Electronic monitoring of adherence remains the gold standard; however, it cannot be used in routine clinical practice due to its high expense. Although inexpensive and easy to obtain during clinic encounters, self-report is inflated to but correlated with objective, electronically-monitored adherence. Isolated serum levels that do not incorporate information about individual characteristics (e.g., pharmacokinetic variation, time of dose) are not a reliable proxy measure of nonadherence since they do not relate to objective adherence data. Future studies using population pharmacokinetic models interpretation of isolated serum levels may allow for improved assessment of adherence and permit better individualization and intervention to maximize seizure control. Funded by the National Institutes of Health - K23 Grant (K23HD057333-01)