Suppression of the hypothalamic-pituitary-adrenal (HPA) axis is an accepted indicator of potential side effects from inhaled corticosteroids. Although cortisol monitoring is frequently used to detect changes in HPA axis activity, the optimal method for identifying the subset of asthma patients on inhaled steroids who experience severe cortisol suppression of potential clinical significance has not been established. The objective of this study was to compare several methods for assessing HPA axis activity in asthma patients taking inhaled corticosteroids. After screening, 153 patients with mild to moderate asthma were randomly assigned to receive inhaled fluticasone propionate (110, 220, 330, or 440 microg bid), flunisolide (500 microg or 1000 microg bid), or one of two control regimens (prednisone or placebo) for 21 days. Salivary (8 a.m.) and urinary (24-h) cortisol determinations were compared against 22-hour area under the serum cortisol concentration-time curve (AUC0-22 h) measured at baseline and on day 21. Comparisons were also made against 8 a.m. serum cortisol. A significant positive correlation was found between AUC0-22 h of serum cortisol and 8 a.m. serum cortisol level (r = 0.5140; p = 0.0001). The AUC0-22 h of serum cortisol was weakly correlated with 24-hour urinary cortisol levels, both corrected (r = 0.4388; p = 0.0001) and uncorrected (r = 0.3511; p = 0.0001) for creatinine excretion. The 8 a.m. salivary cortisol level correlated positively with the 8 a.m. serum cortisol level (r = 0.5460; p = 0.0001). Salivary cortisol was both sensitive and specific for the detection of a 50% decline in AUC0-22 h of serum cortisol. Cortisol reductions of this magnitude have been observed following repeated use of inhaled steroids. Because it is noninvasive, salivary cortisol measurement offers distinct advantages as a screening method for detecting pronounced HPA axis suppression in asthma patients receiving corticosteroid therapy.