Anticonvulsant blood levels: historical review with a pediatric focus

Epilepsy is heterogeneous and its treatment is often complicated by variable drug responses. Buchtal et al reported a close correlation between serum phenytoin levels, electroencephalographic findings, and clinical status in 1960. They suggested that physicians adjust dosage to attain a "therapeutic level." The concept was enthusiastically received. "Therapeutic serum levels" were proposed for most anticonvulsant drugs, and by 1975, most authorities believed that pharmacokinetic factors explained individual differences in drug response. However, Froscher found that measuring levels did not improve patient outcome. More recently, Schumacher's double-blind study found no correlation between phenytoin levels and seizure control or adverse effects. Pharmacodynamic variables (differences in drug responsiveness) are more important than pharmacokinetic factors for many drugs, especially receptor-active drugs. Pharmacokinetic variables were studied first, and led to a simplistic model. They are less significant than pharmacodynamic factors in the case of warfarin anticoagulation. Anticonvulsant levels can reveal noncompliance and pharmacokinetic differences. They say nothing about pharmacodynamics. Reports of "subtherapeutic levels" imply a need to increase dosage, but this is not supported by outcome data. We still lack evidence that specific drug levels are a valid intermediate target 40 years after Buchtal's paper. Responses to some anticonvulsants could depend primarily on pharmacokinetic factors, while pharmacodynamic factors could be supreme for others.