OBJECTIVES: To compare relationships between the sleep-wake cycle and endogenous circadian rhythms in young and older adults and to examine correlates between evening naps and circadian rhythms in older adults.
DESIGN: For 1 week of home recording, subjects wore wrist-activity monitors and kept daily sleep logs. After the home monitoring, subjects entered the laboratory on a 90-minute sleep-wake schedule and were monitored on this schedule for at least 30 hours.
SETTING: Community living and laboratory.
PARTICIPANTS: Sixty-seven young adults, aged 18 to 32, and 56 older adults, aged 60 to 75, who were healthy and had few sleep complaints.
MEASUREMENTS: Times of nocturnal sleep, out-of-bed napping, and illumination were obtained at home. Sleep propensity and oral body temperature (OBT) were measured in the laboratory, along with circadian rhythms of cortisol and 6-sulfatoxymelatonin (aMT6s, assayed from urine samples collected every 90 minutes).
RESULTS: Home sleep times and illumination acrophases (fitted peak times) were advanced in older adults. The phase angles (time intervals) between onset of aMT6s and sleep onset were not changed in older adults, but sleep offset was more advanced than acrophase and offset of aMT6s with aging. Acrophases of cortisol and sleep propensity were advanced in older adults to the same extent as sleep times, but OBT was less advanced than sleep times. Older adults who took evening naps showed more advanced sleep offset and circadian rhythms of aMT6s, but there were no differences in the phase angles of sleep onset and circadian rhythms of aMT6s and cortisol compared with older adults who did not take evening naps.
CONCLUSION: Measuring different circadian markers suggested different phase relationships between the sleep-wake cycle and endogenous circadian rhythms in aging. Early awakening in older adults cannot be explained simply by a relative phase advance of the circadian system. Evening naps and advanced illumination may play a role in the advance of the circadian system in aging.