Interlaboratory variability in drug assay: a comparison of quality control data with reanalysis of routine patient samples. I: Anticonvulsant drugs and theophylline. Clinical Pharmacology and Toxicology Study Group, Italian Society for Clinical Biochemistry.

The interlaboratory variability in the assay of theophylline and various anticonvulsant drugs was evaluated in a 16-month prospective study by using two independent experimental approaches: (a) a conventional method based on the distribution of quality control (QC) lyophilized human sera spiked with the test drugs (phenytoin, phenobarbital, carbamazepine, valproic acid, primidone, theophylline); and (b) reanalysis by two to four reference laboratories of patients' samples randomly selected among those routinely assayed for phenytoin and theophylline by the 52 participating laboratories. For all tested drugs, precision as assessed by results on spiked QC samples was generally satisfactory, interlaboratory coefficients of variation (CV) being below 18.5% for all drugs tested except primidone (CV = 24.1%) at clinically relevant drug concentrations. For both phenytoin and theophylline, there was a good agreement between results obtained by individual reference laboratories (reference results) on routine samples. The correlation coefficients relating the original values to the means of the reference results were greater than 0.94. It is concluded that interlaboratory variability in the assay of the drugs included in this survey is relatively low not only when assessed by using conventional QC material, but also under strictly routine conditions.