PUBMED FOR HANDHELDS

Search MEDLINE/PubMed


  • Title: Neuroendocrine aspects of primary endogenous depression: XII. Receiver operating characteristic and kappa analyses of serum and urine cortisol measures in patients and matched controls.
    Author: Thompson LM, Rubin RT, McCracken JT.
    Journal: Psychoneuroendocrinology; 1992 Oct; 17(5):507-15. PubMed ID: 1484917.
    Abstract:
    The majority of studies investigating the diagnostic utility of hypothalamo-pituitary-adrenal (HPA) axis measures in major depression have focused on the dexamethasone (DEX) suppression test (DST). The DST correlates well, but imperfectly, with other measures of HPA activity. Fewer studies have considered the ability of basal cortisol measures to discriminate depressives from non-depressed patients or normal subjects. We used receiver operating characteristic (ROC) analysis to examine the mean 24-hr serum cortisol concentration, our benchmark of basal HPA axis activity, compared to smaller segments of the 24-hr cortisol profile, post-DEX serum cortisol values, and pre- and post-DEX 24-hr urinary free cortisol (UFC) levels in 40 primary endogenous major depressives compared to 40 matched normal control subjects. No statistically significant differences in ROC curves were found between mean 24-hr cortisol and the other cortisol measures. The mean 24-hr, 1300h-1600h, 1600h-1900h, and 16-hr post-DEX serum cortisol concentrations and the post-DEX UFC level all appeared to be comparable estimators of HPA activity. The single 2300h pre-DEX serum cortisol concentration and the pre-DEX 24-hr UFC level performed notably poorer than did the other measures. We additionally calculated kappa statistics to determine the optimally sensitive and specific discriminators of the cortisol measures between the depressives and the normal controls. The 2300h post-DEX serum cortisol measure was optimally sensitive, and the 1500h post-DEX serum cortisol was optimally specific. The 0700h, 1500h, and 2300h post-DEX serum cortisols were very close together as the optimally efficient measures (best combination of sensitivity and specificity).
    [Abstract] [Full Text] [Related] [New Search]