The present study was undertaken to examine the effect of plasma dexamethasone (dex) concentration on the dexamethasone suppression test (DST) in patients with major depressive episodes. We measured the plasma dex and postdex cortisol concentrations at 4:00 pm in a group of 50 depressed patients, 28 anxiety patients, and 33 normal subjects who were given 1 mg oral overnight DST. Nonsuppressor status was defined by any postdex plasma cortisol concentration > 5 μg/dl. When DST nonsuppressors and suppressors were compared irrespective of subject groups, plasma dex concentrations were significantly lower (t = 4.90, p < 0.001) in the DST nonsuppressors. When the subjects were classified by diagnostic category, there were significantly lower plasma dex concentration in the depressed nonsuppressors versus suppressors (t = -3.54, p < 0.001) and there was a strong, but not statistically significant, trend toward lower plasma dex concentrations in anxiety and normal nonsuppressors versus suppressors. There was a significant negative correlation between plasma dex level between postdex cortisol level in the depressed, anxiety or total group respectively (n = 48, r = -0.213, p < 0.05; n = 28, r = -0.382, p < 0.05; n = 104, r = -0.286, p < 0.05). These relationships disappeared when we restricted our analyses to an empirical range of plasma dex concentrations (0-0.75 ng/ml) within which the DST results were considered valid. Comparing our results with those of Johnson et al (1987) and Ritchie et al (1990), we suggest that Chinese depressives may have a lower limit of plasma dex window and further investigation is needed to determine the pharmacokinetics of dex in Chinese.