Evaluation of a radioimmunoassay of urinary cortisol without extraction

We evaluated the Kallestad "Quanticoat" cortisol RIA for direct (no extraction) measurement of urinary free cortisol, which requires no solvent extraction. An analytical-recovery study showed a linear regression of y (measured) = 0.65x (added) + 37.5 micrograms/L (Sy.x = 21.4 micrograms/L, r = 0.978, n = 48). Recovery appeared to vary with the urine used and with the concentrations of cortisol added. Within- and between-run CVs were less than or equal to 4.1% and less than or equal to 3.8%, respectively. Cross reactivities were low, except for prednisolone (20.5%). This no-extraction method gave higher values for urinary free cortisol than did either an RIA method involving extraction or an HPLC method. A comparison study with the HPLC (x) and with the method involving extraction (x') gave the following Deming-debiased regression equations: y = 1.60x + 68.8 (Sy.x = 34.4, n = 29) and y = 1.33x' + 0.69 (Sy.x = 40.3, n = 66), respectively. We conclude that the no-extraction method may give misleading results for patients' diagnosis or management if this cross reactivity is not taken into account.