Serum cortisol in adrenal hirsutism as estimated by five different methods

Serum cortisol had been estimated in 152 hirsute women complaining of fertility problems, of whom 36 were subsequently diagnosed as having adrenal hirsutism and 10 as having congenital adrenal hyperplasia (steroid 21-hydroxylase deficiency), using five methods: an in-house tritium radioimmunoassay after extraction with ethanol; the Diagnostic Products Corp. "Coat-a-count" iodinated direct radioimmunoassay; the Pharmacia-LKB "DELFIA" lanthanum-enhanced fluoroimmunoassay; the Amersham "Amerlite" luminescence immunoassay; and the Walker "Synelisa" enzyme-linked immunoassay. Although stripped pool serum samples containing weighed amounts of cortisol produced acceptable values in all assays, the patient samples showed a number of high results, much greater than the accepted normal upper limit of 250 ng/ml (25 micrograms/dl, 690 nmol/l). This was especially so in 21-hydroxylase deficiency, when cortisol values should be very low. Only the luminescence and iodinated assays produced very low values after dexamethasone suppression. After the outliers had been excluded, only the iodinated assay showed a good statistical agreement with the more elaborate tritium assay. The most specific assay was the luminescence method, which produced generally lower results in most cases. This was selected as the new routine method. The unreliable cortisol results in adrenal hirsutism are attributed to high cross-reaction of the antiserum in each of the assays with 17-hydroxyprogesterone, progesterone and 21-deoxyderivatives of cortisol and deoxycorticosterone. In general, all standard and commercially available cortisol assays appears to be unsuitable for cortisol estimation in 21-hydroxylase deficiency, and probably also for neonates.