The effects of age, cognitive function (measured by Cambridge cognitive examination (CAM-COG) score); and dexamethasone (DEX) levels on the dexamethasone suppression test were studied in 33 healthy older subjects (age 51-96). Three subjects (9.1%) were nonsuppressors and were older and had lower CAMCOG scores than the 30 suppressors. Significant correlations were observed between natural log-transformed postdexamethasone cortisol (LNCOR) levels and age (r = 0.40) and CAMCOG score (r = -0.45). Multiple regression analysis was used to investigate the relationship between LNCOR, age, DEX levels, and CAMCOG score. Age and DEX combined explained 41% of the variance in LNCOR values, whereas CAMCOG score and DEX levels explained 44% variance. As age and CAMCOG were highly correlated (r = -0.72), both together did not significantly improve the fit of regression equation (47% variance explained). These findings suggest an association between advancing age, impaired glucocorticoid feedback, and cognitive dysfunction in healthy human subjects. Although any causal connection remains to be demonstrated, results would be consistent with the "glucocorticoid cascade" hypothesis of human aging.