Chapter 4 122 Associations between the CAR and sleep outcomes In this study, we also evaluated the association between the CAR and sleep. Moderation analyses show that a higher CAR is associated with better sleep quality in healthy men and fewer sleep disruptions in OC users, but neither association is present in NC women. Mediation analyses show that the CAR does not mediate the differences in sleep quality in OC users compared to NC women. The results from the mediation analysis are partially in line with Zhang et al. (2014), who found lower awakening cortisol levels in men with insomnia and higher awakening cortisol levels in women with insomnia. One possible explanation for the observed sex- and OC-specific associations could be found in differences in temporal sex hormone dynamics between NC women, OC users and men. In NC women, there could be a dynamic association between the HPA axis and estradiol: studies on the CAR throughout the menstrual cycle found that close to ovulation, when estradiol level are highest, the CAR is also the highest (Ozgocer et al., 2017; Wolfram et al., 2011). This indicates that high estradiol levels could support healthy cortisol dynamics. If this were the case, the association between CAR and sleep in NC women could be dependent on menstrual cycle phase which we could not capture with our present dataset or study design. In OC users, the finding of a lower CAR is in line with previous findings, which also show reduced daily cortisol levels and cortisol reactivity after stress tasks (Herrera et al., 2019). These findings paint the image of a reduction in overall cortisol dynamics in OC users. One mechanism through which salivary cortisol could be reduced in OC users is through an increased level of cortisol-binding globulin (CBG). In serum, the majority of cortisol is bound to CBG and not bioavailable (Lewis et al., 2005). Salivary cortisol samples only measure the small proportion of unbound (bioavailable) cortisol, and CBG levels can therefore significantly affect salivary cortisol measurements (Kumsta et al., 2007; Vining et al., 1983). OCs significantly increase the levels of CBG, resulting in higher CBG levels in OC users compared to naturally cycling women (van der Vange et al., 1990; Vos et al., 2017). Although OCrelated CBG changes could be more likely to influence absolute salivary cortisol levels rather than cortisol dynamics, it is possible that higher serum CBG levels in OC users reduced the amount of free cortisol, therefore lowering lower salivary cortisol levels and CAR in our study. We find that OC users with a lower CAR show higher probability of sleep disruptions, which
RkJQdWJsaXNoZXIy MTk4NDMw