Relationship of sleep with diurnal cortisol rhythm considering sleep measurement and cortisol sampling schemes

Abstract

Information on the relationships between the previous night's sleep and the next-day diurnal cortisol rhythm is inconsistent due to confounding factors such as sleep measurements (trait/state sleep and objective/subjective sleep) and cortisol sampling schemes. Therefore, this study aimed to investigate these relationships, considering the confounding factors. College students (n = 79) wore actigraphy for 3 days to undergo an evaluation of previous night-time sleep objectively and reported their subjective sleep parameters in a sleep diary. In addition, participants provided six salivary cortisol samples daily. Furthermore, six cortisol sampling schemes were created to reflect diurnal cortisol rhythms, and two different methods were used to calculate the index of diurnal cortisol slope (DCS). A multilevel model was created to examine the impact of both trait and state sleep on next-day diurnal cortisol rhythm. The results revealed that higher objective state sleep efficiency and longer objective state total sleep time were associated with a higher cortisol awakening response (CAR). Moreover, higher objective trait sleep efficiency and longer objective trait total sleep time were associated with higher waking cortisol levels and steeper DCS. In addition, a minimum of four saliva samples were required at different time points, including upon waking, 30 min after waking, 1 h after waking, and at bedtime, to explore the relationship of sleep efficiency/total sleep time with waking cortisol, CAR, and DCS. Furthermore, the index of the peak-to-bed slope was appropriately employed to examine the relationship between sleep efficiency and DCS, whereas the wake-to-bed slope was effective for examining the relationship between total sleep time and DCS. In summary, this study clarified the relationship between sleep and next-day diurnal cortisol rhythm and suggested a cost-effective cortisol sampling schedule and calculation methods.