A RAPID MONOCLONAL-ANTIBODY BLOOD THEOPHYLLINE ASSAY - LACK OF CROSS-REACTIVITY WITH ENPROFYLLINE

We evaluated a rapid monoclonal antibody theophylline assay for two reasons: (a) to determine its specificity with respect to the possible confounding influence of a structurally related xanthine, enprofylline, and (b) to assess its accuracy relative to high-performance liquid chromatography (HPLC). Blood samples were taken from 233 patients who had been randomized in double-blind fashion to receive either oral theophylline (n = 117) or enprofylline (n = 116) for the treatment of chronic reversible obstructive airways disease. Monoclonal antibody assays (MAAs) were performed in 10 clinical sites by 10 trained paramedical technicians. Three patients, who actually received enprofylline but not theophylline, had MAA theophylline values of greater-than-or-equal-to 3.2 mug/ml, giving a specificity of 97%. HPLC determination of simultaneous blood samples confirmed that theophylline levels were in fact <3.2 mug/ml and that theophylline was not being taken surreptitiously. Good correlation was observed between MAA and HPLC in patients taking theophylline (y = 1.07 x + 0.36; r = 0.93; standard error of the estimate (SEE) = 1.93). However, there was wide variability from technician to technician such that r values for individual sites ranged from 0.67 to 0.99. Based on the overall correlation, the prediction of an individual HPLC value from an individual MAA value had broad 95% confidence limits: when the MAA value was 10 mug/ml, the predicted HPLC value was 9.19 +/- 3.32; when MAA = 15 mug/ml; HPLC = 13.19 +/- 3.33; and when MAA = 20 mug/ml; HPLC = 17.19 +/- 3.36. Our results confirmed good specificity of the MAA with only occasional false positive readings in the presence of a structurally related xanthine. However, the accuracy of this assay may vary among trained technicians and, although accurate in a statistical sense, it may have limited usefulness in a clinical context as result of the broad confidence limits of individual predicted values.