Stress management reduces intraindividual cortisol variability, while not impacting other measures of cortisol rhythm, in a group of women at risk for breast cancer

Abstract

Objective: The stress hormone cortisol exhibits a diurnal rhythm throughout the day, as well as within person variability. Recent statistical approaches allow for the estimation of intraindividual cortisol variability ("ICV") and a greater ICV has been observed in some mood disorders (major depression, remitted bipolar disorder); however, ICV has not been examined following stress management. In this secondary analyses of an efficacious randomized clinical trial, we examine how ICV may change after cognitive behavioral stress management (CBSM) among healthy stressed women at risk for breast cancer. Second, we concurrently compare other calculations of cortisol that may change following CBSM.

Methods: Multilevel modeling (MLM) was applied to estimate ICV and to test for a group by time interaction from baseline, post-intervention, to 1 month following CBSM. Forty-four women were randomized to the CBSM; 47 to the comparison group; mean age of the entire group was 44.2 (SD=10.27).

Results: After controlling for relevant covariates, a significant time by group interaction emerged (β estimate=-.070; p<.05), such that CBSM participants demonstrated a lower ICV following CBSM compared to the comparison group. The interaction for cortisol slope and cortisol output (area under the curve) approached significance (β estimates=-.10 and -.062, respectively; p's<.08), while other cortisol outcomes tested were not significantly changed following CBSM.

Conclusion: ICV may represent a novel index of cortisol dysregulation that is impacted by CBSM and may represent a more malleable within-person calculation than other, widely applied cortisol outcomes. Future research should examine these relationships in larger samples, and examine ICV and health outcomes.

Clinical trials identifier: NCT01048528.

Keywords: Cortisol; Modeling; Stress; Stress management; Variability.

Published by Elsevier Inc.

Figures

Fig. 1

Participant flow chart. *The multilevel models applied in the current analyses accommodate missing data in longitudinal models, thus participants providing initial cortisol sample (n = 91) were included in the final analysis.