Salivary cortisol and sleep problems among civil servants.

OBJECTIVE: The present study used information from a field study conducted among 4489 civil servants (70% women) in Denmark in 2007. The purpose was to examine the association between sleep problems and salivary cortisol by using a cross-sectional design with repeated measures in a subsample three-month later.
METHODS: Sleep problems during the past night and the past 4 weeks were assessed by a self-administered questionnaire on overall sleep quality, disturbed sleep, sleep length and awakening problems. Saliva samples were collected in a single day, using cotton tubes, 30 min after awakening and again at 2000 h. A subsample of 387 participants collected saliva samples three-month later at awakening, +20 min and +40 min after awakening and at 2000 h. We adjusted for confounders related to sampling time, life style and personal characteristics, socioeconomic status and work aspects.
RESULTS: Sleep problems during the past four weeks were associated with low morning and evening saliva cortisol concentrations: [-3.1% per score of disturbed sleep (p=.009); and -4.7% per score of awakening problems (p<.001)]. Whereas sleep problems were not related with slope (the morning to evening change in cortisol levels). Awakening problems predicted lower cortisol (-7.51% per score; p=.003) three-month later. Cortisol awakening response (CAR) and slope three-month later were significantly associated with disturbed sleep (-7.84% and -8.24%) and awakening problems (-6.93). Area under the curve (AUC(morning)) increased with disturbed sleep (3.77%).
CONCLUSION: Surprisingly, low morning cortisol was associated with increased sleep problems during a four-week period prior to sampling among 4066 Danish civil servants. At follow-up three-month later, those with sleep problems had a flattened cortisol profile. Those with awakening problems also had low salivary cortisol in general.