Accuracy and precision of methods for theophylline measurement in physicians' offices.

We tested the accuracy and precision of five theophylline methods intended for use in physicians' offices. The Syntex AccuLevel, Ames Seralyzer, and 3M Diagnostics TheoFAST methods were less reproducible (CVs 6.3% to 9.2%) than the Abbott Vision and Kodak DT-60 (CVs 2.2% to 3.3%). Caffeine interfered with the Vision, Seralyzer, and AccuLevel methods, and theobromine interfered with the Vision, Seralyzer, and TheoFAST methods. Only the DT-60 method was free from interference from any of the 24 compounds tested. Results by all methods correlated well with those by the HPLC comparison method (Clin Chem 1981;27:1931-3) and by the Abbott TDx method for assay of 100 serum (or, when appropriate, paired whole-blood) samples. The frequency of sample results differing from the comparison method by greater than 2.0 mg/L was as follows: TDx, 11%; Vision (serum), 12%; Vision (whole blood), 18%; DT-60, 14%; AccuLevel, 18%; Seralyzer, 25%; and TheoFAST, 31%. The Kodak DT-60 method was the most nearly accurate and precise among these physician's office methods. Some physician's office methods for theophylline analysis are not adequate to guide dosage adjustments.