How Much “UFC” Is Really Cortisol?

When urinary free cortisol (UFC) determinations became readily available for clinical use in 1968 (1), chromatographic methods were cumbersome. The first competitive protein-binding method (radiotransinassay) (1) used human corticosteroid-binding globulin (transcortin) as the binding protein. The specificity of the assay was enhanced by the use of Fuller’s earth as the agent to adsorb the unbound fraction because it takes up some of the competing steroids differentially (2). The reference interval was ∼10–100 μg/day (30–300 nmol/day). When radioimmunoassays, based on the same competitive binding principle but using antibodies raised to cortisol linked to albumin, became popular during the 1970s, it was assumed that these would be more specific for cortisol, but this assumption was not warranted. As pointed out recently (3), most of the articles published over the past 20 years have quoted even higher values, reflecting a significant lack of specificity. In 1976, Chattoraj et al. (4) found values for UFC after combined thin-layer and column chromatography that were approximately one-half those of the original method (1), as did Schöneshöfer et al. in 1980 (5) after HPLC, and Murphy et al. in 1981 (6) after Sephadex LH-20 chromatography. The last group (6) showed that of four commercial antibodies tested in samples from 20 patients, including patients with Cushing syndrome, obesity, and pregnancy, two appeared to be much less specific than the three transcortins (human, dog, and horse). One of the antibodies gave values four- to fivefold higher, but preliminary chromatography was recommended, which lowered the values to close to the transcortin values. None of the antibodies gave values substantially lower than the transcortins. Whereas the transcortins all gave values that correlated well with each other (r ≥0.90), two of the antibodies gave consistently higher values in pregnancy. The situation for the determination of cortisol in cord blood at delivery was similar: some antibodies gave extremely high values—up to 10-fold higher than the values after chromatography (7). As pointed out by Murphy in 1983 (7), cord blood contains many steroids that are not present in adult blood. Although it has been apparent for more than 20 years that many antibodies give falsely high values for cortisol, especially for “UFC”, this fact seems to have been systematically ignored. Only a few authors have bothered to validate their methods, although HPLC has been readily available during this period. Although some of the methods used seem to correlate well with the true values (8) and have demonstrated their clinical utility, is this good enough? Although such methods have been useful for the detection of Cushing syndrome, low values cannot be interpreted. Few commercial suppliers actually give normal means and ranges for UFC. Their specificity data are not very helpful because much of the competing material has not been identified and varies from one batch of antibody to another, even by the same supplier. If it is assumed, and it usually is, that the unidentified material consists of cortisol metabolites, does it really matter? Yes, it may be very important because this assumption requires that the metabolism of the cortisol must be normal—an assumption that is unwarranted and is rarely tested. Demitrack et al. in 1991 (9) provided evidence that patients with chronic fatigue syndrome have lower values of UFC than healthy subjects. Their values in healthy subjects had a mean of 200 nmol/day with a range from ∼100 to 300 nmol/day, approximately threefold higher than the “true” mean excretion rate of 57 nmol/day, with a range of 28–117 nmol/day, that had been established by Schöneshöfer et al. (5) using HPLC. Their patients with chronic fatigue syndrome had a mean excretion rate of 125 (range, 70–200) nmol/day, i.e., considerably higher than the values established with more nearly accurate methods but lower than in the control subjects studied with their method. So what do the lower “UFC” values in the chronic fatigue syndrome patients really mean? Such patients do not have true adrenal insufficiency as determined by adrenocorticotropic hormone or insulin hypoglycemia testing, and have been considered to have “central” adrenal insufficiency. However, because less than one-half of the “UFC” is really cortisol, interpretation of these results is impossible. Rather than low cortisol, might it not be some metabolite that is present in lesser quantities in the patients compared with the controls? And might this not reflect a difference in the way cortisol is handled by these patients? The same considerations apply to the interpretation of the results of Putignano et al. (10) in this issue. Because only approximately one-half of the material is cortisol (range, 16–88 μg/day or 44–244 nmol/day), it is difficult to interpret the results. We must assume that the unidentified material behaves the same as cortisol at all rates of water excretion, which may or may not be true. Their results differ from those of Mericq and Cutler in 1998 (11), whose values were approximately four times the true values. As the authors themselves point out, further studies using chromatographic methods are needed to resolve this issue. After more than 20 years since the true values for UFC were established, it is time we stopped using such inaccurate methodology and terminology with respect to “urinary free cortisol”.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.