Solid-phase extraction as an alternative to dichloromethane in analysis of urine free cortisol.

Abstract

Measurement of daily free cortisol excretion is used in the assessment of hypercortisolism. Because of the large number of interferents present in urine, three-quarters of laboratories extract urine prior to assay. The most commonly used method involves a cumbersome and timeconsuming procedure with dichloromethane. Recently, high-performance liquid chromatography (HPLC) methods have been developed for the analysis of urine free cortisol (UFC). Rather than dichloromethane extraction, more practical solid-phase extraction methods were used. In this study, we compared the performance of a solid-phase and a dichloromethane extraction method prior to analysis of UFC. Urine free cortisol was measured in 72 samples using competitive chemiluminescent immunoassay techniques. Samples were analysed by the Immulite (DPC Ltd, Llanberis, UK) after extraction with dichloromethane. Measurements were also undertaken using the ACS:180 (Chiron Diagnostics, Halstead, Essex, UK) without extraction and after solid-phase extraction using C8 columns (200mg/10mL, International Sorbent Technology, Hengoed, Mid Glamorgan, UK). Brie y, columns were activated by washing with 3mL of methanol and borate buffer (25mmol/L, pH9). Subsequently, urine (300mL) was alowed to adsorb on to the columns under reduced pressure. After washing sequentially with 261mL borate buffer, 261mL acetone and 1mL hexane, the columns were allowed to dry under vacuum. Cortisol was eluted with 261mL ethylacetate and samples evaporated to dryness at 408C under air prior to resuspension in cortisol diluent and analysis. Results after solid-phase extraction showed better precision than after dichloromethane extraction. Within-assay precision for the solidphase extraction method was 4 6, 5 1 and 3 1 percentage coef®cient of variation (%CV) for cortisol concentrations of 164nmol/L, 347nmol/L and 904nmol/L, respectively (n=10). Between-assay precision was 7 2, 4 8 and 5 3%CV (171, 298 and 572nmol/L cortisol, respectively). This compared with values of 14, 13 and 8 3%CV after dichloromethane extraction (cortisol concentrations of 149, 526 and 1085nmol/L). Recovery experiments were performed by adding different amounts of urine from a patient with Cushing’s disease to three different urine samples. The quoted recoveries are the means for three independent determinations of each cortisol concentration. Recoveries for the solid-phase extraction method were 89, 90 and 80% for cortisol concentrations of 90, 195 and 390nmol/L. Values for the dichloromethane extraction method were 85, 85 and 81% at the same cortisol concentrations. The solid-phase and dichloromethane extraction methods showed good agreement. Comparing results from the ACS:180 (after solid-phase extraction) with those from the Immulite (after dichloromethane extraction), the former showed a 35% positive bias (P<0.001, n=72, paired Student’s t-test). This difference was independent of cortisol concentration (see Fig. 1). The correlation coef®cient was 0 987 and the regression line had a gradient of 0 94 (95% CI 0 88± 0 98) and intercept 18 nmol/L. A component of this bias was the inherent differences between the two immunoassay methods used. When results from the ACS:180 (unextracted) were compared with those from the Immulite (after dichloromethane extraction), the former showed a 58% positive bias (P=0 007, n=32, paired Student’s t-test). The correlation coef®cient was 0 921 and the regression line had a gradient of 0 97 (95% con®dence interval (CI) 0 77±1 2) and intercept 98 nmol/L. This suggests that the solid-phase method Short Report Ann Clin Biochem 2001; 38: 67±68