Salivary cortisol and sleep disruption among early-stage breast cancer patients undergoing a stress management and relaxation training intervention

Date of Award




Degree Name

Doctor of Philosophy (Ph.D.)

First Committee Member

Charles S. Carver, Committee Chair

Second Committee Member

Michael H. Antoni, Committee Member


This study investigated the relationships between sleep, mood, and salivary cortisol changes among newly diagnosed, early-stage breast cancer patients undergoing a group-based stress management and relaxation intervention. This study was largely exploratory since little research has been conducted on cortisol production and sleep variables in the context of cancer patients and in response to a relaxation intervention. No cortisol differences were found between the Intervention (N = 40) and Control (N = 12) participants; both groups had significant increases of pre-session cortisol over a 10-week period. Most analyses were conducted on the Intervention participants who attended 10 weekly groups. Predictor variables included psychological variables (Life Orientation Test, Sickness Impact Profile, expectations for recurrence, and locus of control about remaining free of cancer), compliance and relaxation practice, and sleep variables (quality, duration, and time to fall asleep). Outcome measures included weekly pre-, post-, and within-session changes in cortisol, as well as measures of negative mood, confusion, and fatigue using the Incredibly Short POMS (ISP). Cortisol decreased significantly within weekly sessions. Pre- and post-session levels followed an inverted-U quadratic pattern over the 10 weeks. This pattern of cortisol over time may be due to initial increased awareness of stress and HPA activation over time, followed by psychological skills to effectively cope with stressors. Negative mood also decreased within sessions and increased over time. As expected, momentary negative mood was related to greater cortisol levels cross-sectionally. Relaxation home practice was related to initial within-session decreases in cortisol, but greater increases over 10 weeks. Attendance was unrelated to cortisol measures. Linear increases and inverted-U patterns of post-session cortisol across the 10 weeks were related to a more external locus of control regarding remaining free of cancer. Poor sleep quality ratings were related to initially greater cortisol levels, but also to greater declines in cortisol over 10 weeks. Poor sleep quality was also related to greater negative mood and fatigue. Exploratory analyses revealed differences in cortisol patterns between cohorts of groups. While this investigation may include interesting findings with regard to the relationships between sleep, mood, and cortisol, the study design contains important limitations; generalizability of the results should be done with caution.


Psychology, Clinical; Psychology, Physiological; Health Sciences, Oncology

Link to Full Text