Sleep quality and the biological stress system during an internet-based intervention for major depressive disorder.

Introduction: Poor sleep quality is a persistent and debilitating symptom of major depressive disorder (MDD), with dysregulations in the biological stress system constituting a potential underlying physiological mechanism. Accordingly, a psychotherapeutic intervention may affect the interplay between sleep quality, MDD and the biological stress system.We examined how basal cortisol, and alpha-amylase levels correspond to perceived sleep quality during an internet-based intervention for MDD. Furthermore, we investigated how changes in sleep quality during the intervention relate to changes in these biological stress system markers. We hypothesized that: 1) short-term and long-term sleep quality would improve during the intervention, 2a) across assessment time points, poor sleep quality would be associated with higher cortisol and alpha-amylase concentrations, and 2b) pre-to-post intervention improvements in sleep quality (treatment response) would be associated with pre-to-post decreases in both biological markers, compared to non-response.

Methods: We analyzed forty-one participants (age: 35 ± 12y; females: 82.6 %) suffering from mild to moderate MDD. The cognitive behavioral internet-based intervention consisted of seven weekly writing-based modules with individualized feedback. Participants collected 12 saliva samples at home over two consecutive weekdays at pre-, mid-, and post-intervention. Outcome parameters of the cortisol and alpha-amylase diurnal profiles were the awakening responses, the total diurnal output, and the diurnal slopes. Self-reported sleep quality was retrospectively assessed for the night before (short-term) and for the two-week period preceding saliva collection (long-term). Treatment response was determined using the reliable change index of the pre-to-post, two-week sleep quality difference scores. Hypotheses 1 and 2a were tested using random intercept hierarchical linear models, Hypothesis 2b was tested using linear regressions with age, biological sex, BMI and medication use on the day of sampling as covariates.

Results: Long-term sleep quality increased significantly from pre-to post-intervention (d = 0.78; p < 0.001), partially confirming Hypothesis 1. Contrary to the expected effect of Hypothesis 2a, poor long-term sleep quality at pre-intervention was associated with a blunted cortisol awakening response (CAR; p < 0.05). Post-hoc analyses showed an association of pre-to-post CAR changes and pre-intervention sleep quality (p < 0.01) indicating that individuals with higher pre-intervention sleep problems, on average, exhibited a pre-to-post increase in the CAR. The responder analyses showed that individuals with a marked pre-to-post sleep quality increase (i.e., responders) showed a higher increase in the CAR, compared to non-responders (p < 0.05), which again ran contrary to the effect proposed in Hypothesis 2b.

Discussion: Prior to psychotherapeutic treatment MDD patients with poor sleep quality showed a blunted CAR, pointing to hypocortisolemia in these individuals. Furthermore, intervention-induced changes in sleep quality may lead to a normalization of the CAR.