Background: Technical hurdles limit the characterization of key hormonal rhythms. Frequent sampling increases detection of changes in magnitude or circadian and ultradian patterns, but limits feasibility for clinical or research settings. These caveats are particularly pertinent for cortisol, a hormone that displays a prominent circadian rhythm and whose magnitude is tightly regulated in the absence of biobehavioural challenge. Ideally, one would like to obtain samples non-invasively from a matrix of interest at frequent intervals. While many investigations have reported a high correlation between serum and salivary cortisol assays, the degree to which salivary cortisol reflects the circadian patterns of circulating cortisol concentrations has not been established across a 24 h period.
Methods: We obtained hourly serum and salivary samples over a 24 h period in nine adults in an inpatient setting. The circadian patterns for serum and salivary cortisol were analysed by harmonic regression.
Results: For all but two subjects (both on oral contraceptives), the salivary cortisol concentration was synchronous with the serum concentration, indicating that the salivary assay could be substituted for the serum assay to assess circulating rhythmicity across the 24 h time frame.
Conclusions: This statistical model has distinct improvement over the correlational approach of examining serum and saliva cortisol relationships. Saliva cortisol appears to represent serum cortisol across the 24 h period, except for those on oral contraceptives.