Comparison of assay methods used to measure antiepileptic drugs in plasma.

The various techniques used to measure the concentration of antiepileptic drugs in plasma have been compared using data from the HEATH-CONTROL external quality control scheme. Assay methods were categorized as gas-liquid chromatography with and without derivatization (GLC + D, GLC-ND), radio- and enzyme multiplied immunoassay (RIA, EMIT), thin-layer and high-pressure liquid chromatography (TLC, HPLC), and spectrophotometry and colourimetry. The accuracy of the methods was determined from differences of measurements from the levels of drugs spiked into the samples. Most methods significantly underestimated the spiked drug concentrations. Comparisons of assay precision were made with the standard deviation of measurements or from the performance of individual laboratories in a ranking procedure. Comparable significant differences were found in both analyses. RIA followed by GLC-ND were the least variable methods for phenytoin; EMIT and then GLC-ND were the best for both phenobarbitone and primidone. Spectrophotometric measurements of phenobarbitone were unacceptably variable. TLC and colourimetry performed well for carbamazepine, whereas the GLC methods were the most variable. HPLC topped the list for ethosuximide, whereas EMIT had difficulty with this drug. Methods used for valproic acid were similar in precision. GLC + D was consistently less effective for all drugs when compared with GLC-ND.