Study objectives: To present normative data of sleep-wake characteristics and to examine risk factors associated with extreme values (i.e., in the 5 lower and upper percentiles of the distribution) in older adults.
Design: Cross-sectional telephone survey
Setting: The metropolitan area of Paris, France.
Participants: A total of 7010 randomly selected households were contacted. Among them, 1264 households included at least 1 resident 60 years of age or older; 1026 subjects agreed to participate (participation rate: 80.9%).
Interventions: None.
Measurements and results: Subjects were interviewed with the Sleep-EVAL System about their sleeping habits and sleep and psychiatric disorders. In addition, the system administered to all the participants the Psychological General Well-Being Schedule, the Cognitive Difficulties Scale (Mac Nair-R), and an independent living scale. The median nighttime sleep duration was 7 hours without significant difference between the age groups. Factors positively associated with the 5 percentile (4 hours 30 minutes or less) of nighttime sleep duration were obesity, poor health, insomnia, and insomnia accompanied by daytime sleepiness and cognitive impairment. At the other extremity (95th percentile), long sleep (9 hours 30 minutes or more) was associated with organic disease, lack of physical exercise, and lower education. A daytime sleep duration of 1 hour or more (95th percentile) was associated with being a man, cognitive impairment, high blood pressure, obesity, and insomnia. Long sleep latency (95th percentile at 80 minutes) was associated with anxiety, lower education, poor health, insomnia without excessive daytime sleepiness, and obstructive sleep apnea syndrome. Obesity and loss of autonomy in activities of daily living was associated with both early (9 PM or earlier) and late bedtime (1 AM or later) and early (< or = 5 AM) and late (> or = 9 AM) wake-up time.
Conclusions: This study illustrates the usefulness of normal distributions of sleep parameters in the general population to calculate different risk factors associated with extreme values of the normal distribution.