Lack of Specificity of Urinary Free Cortisol Determinations: Why Does It Continue?

Values for urinary free (i.e. unconjugated) cortisol (UFC) vary widely in the literature (Table 1). Although it was shown 19 years ago that the true mean value in healthy subjects is about 20–30μ g/day (2, 3), three recent articles in this journal demonstrate that the gap between reported values is widening rather than narrowing. In their article “Obesity and gender influence cortisol secretion and metabolism in man” (12), Andrew et al. report the highest values yet! They found mean levels of urinary free cortisol in urine of 95 μg/day for men and 218 μg/day for women—values that far exceed the established values of about 45 μg/day for the crude assay (1, 5, 6, 8)—and 25 μg/day after Sephadex LH-20 (3) or high-performance liquid chromatography (HPLC) (2, 9). Their data probably include conjugated cortisol in addition to UFC. If there are such large discrepancies between the cortisol conjugates in men and women, these are of interest and should be explored further. In the initial paper describing the human CBG competitive binding assay, no differences in UFC were found between obese and nonobese subjects, either men or women (1). Because the difference between the crude assays (1, 5, 6) and the specific ones (2, 3, 7) is mainly due to cross-reactions with still unidentified polar metabolites, it is important to assess UFC in obese subjects using a specific method. Most studies have found that UFC in men exceeds that found in women (1, 3, 5, 7, 8) in a ratio of about 1.4:1. Why do the reported UFC levels vary so widely? One reason is that the producers of commercial assays have continued to use the term UFC without pointing out the fact that half or less of what is being measured is actually cortisol. Indeed, clinical relevance has been demonstrated for what is measured in some assays (1, 3, 9). Lin et al. (9) compared competitive protein-binding to human corticosterone-binding globulin (1, 2) with an HPLC method specific for cortisol and found them to be equally effective in diagnosing Cushing’s syndrome. However, that the cross-reacting material is behaving like cortisol in all radioimmunoassays is an assumption; although kit manufacturers provide specific data for known steroid competitors, the bulk of the competing material is never identified. While this assumption may hold true for some diseases, it may not hold true for others. Recently, Mericq et al. (10) have raised the possibility that high fluid intake increases UFC excretion. Because their values at normal urine volumes (77 μg/day) are about four times higher than those after HPLC, as they themselves pointed out, this should be investigated with more specific assays. Although the crude assays have been shown to be clinically useful provided they are carefully validated (1, 3, 9), all competitive binding assays (including all radioimmunoassays) overestimate true UFC, and in order to measure cortisol alone, it is necessary to employ chromatography. Laboratories that rely on crude assays should designate their material by some term other than UFC, such as urinary free corticoids (suggested in ref. 3), cortisol-like material, RIA-reactive material, or some such, and should not express it as nmol/L, as the molecular weights of the other reacting materials are not known. Specific assays are particularly important when exploring physiological problems such as obesity (1, 12) or pathological entities such as those of fibromyalgia (11), depression (6), and chronic fatigue syndrome (4). Values for urinary free cortisol: mean ± sd, μg/day in healthy subjects The mean ± sd is used here because, with the exception of Lamb et al. (5), who used the median, and Schöneschöfer et al. (2), who used mean and range, that is the way authors have expressed their data. CPB, competitive protein-binding; CBG, corticosteroid-binding globulin; HPLC, high performance liquid chromatography; chrom., chromatography. Values for urinary free cortisol: mean ± sd, μg/day in healthy subjects The mean ± sd is used here because, with the exception of Lamb et al. (5), who used the median, and Schöneschöfer et al. (2), who used mean and range, that is the way authors have expressed their data. CPB, competitive protein-binding; CBG, corticosteroid-binding globulin; HPLC, high performance liquid chromatography; chrom., chromatography.