Simple adaptation of a radioimmunoassay kit for serum cortisol used to measure salivary cortisol.

Drug immunoassays with use of reagent strips have been introduced by the Ames Co. for theophylline and phenytoin assays. Their phenobarbital assay is based on the ais (Apoenzyme Reactivation Immunoassay System) principle. Thirty microlitres of diluted serum is applied to the reagent pad and the strip is read in the Seralyzer reflectance photometer after 85 s. Fifty-three serum samples assayed by our “high-performance” liquid-chromatography (HPLC) were analyzed in duplicate with the reagent strips. Linear regression analysis gavey (Seralyzer) = 1.OOx + 1.65 p g/rnL, = 1.15 ,ug/mL, SDXJS X = 8.7, r = 0.994. Estimates of precision of duplicate analyses of patients’ samples gave a mean CV of 3.0% over a range of 6.2 49.8 jzg/mL, with none exceeding 6% for any individual pair. Quality-control samples gave within-batch CVs <2.8% and between-batch (n = 13) CVs of <3.9%. To check stability of calibration, we analyzed the HPLC controls with the Seralyzer strips over 33 days without recalibration, which gave between-batch CVs <3.4%. Assays of 18 samples from an external quality-assessment scheme with the Seralyzer strips gave linear regression estimates against the quoted overall trimmed mean values of: y (Seralyzer) = 1.02x 0.34 pg/mL, S = 1.07 j g/mL, SDX./SYX = 13.8, r = 0.998. Although the use of qualitycontrol and assessment material with dry-reagent chemistry systems has been queried, we found that test-strip results for the multiply-spiked lyophilized human samples showed no significant bias from the overall trimmed mean. The stability of calibration displayed is important in “stat” situations. The detection limit, 5 pg/mL, might be a problem in monitoring compliance, but would not be a problem where toxicity is the clinical query.