What does the influence of sleep duration have on mortality and morbidity?

Although it is commonly assumed that 8 hours of sleep per night is optimal for health and well-being, epidemiologic studies have consistently shown a “U-shaped” association of sleep duration with mortality.1 Across various studies, the lowest mortality rates have been associated with 6 to 7 hours of sleep, with progressively higher mortality associated with sleeping lesser or greater amounts. These associations have been established in approximately 20 studies over a 40-year period, and have included samples of more than 1 million people, statistical control for up to 32 variables, and follow-up assessments of 10 to 20 years.

A similar U-shaped association of sleep duration with morbidity has been found. Long sleep, as well as short sleep, has been associated with a higher prevalence of cardiovascular disease, obesity, myocardial infarction, stroke, cancer, and depression.2

Whereas there has been tremendous recent research focus on the potential negative health implications of short sleep and underlying mechanisms, the issue of how long sleep might be hazardous has been essentially ignored. One apparent reason for this is that the notion that excessive sleep can be harmful is counterintuitive, particularly in the context of current publicity suggesting that modern industrialized societies are chronically and dangerously sleep-deprived. There is no compelling scientific evidence to support these fears. As for the plausibility that long sleep could be harmful, consider that many health-related behaviors can be harmful in excessive amounts, including exercise, sunlight exposure, and caloric and water consumption.

Despite the predominant focus on short sleep, there is reason to believe that long sleep is a more important risk, at least with respect to mortality. For example, in a sample of 636,095 women, after adjustment for 32 covariates, sleeping >7.5 hours was associated with higher mortality than sleeping as little as 3.5 to 4.5 hours.3 Furthermore, a far greater proportion of the sample reported sleeping >8 hours than sleeping short durations. Of excess deaths related to sleep duration other than 6.5 to 7.5 hours, approximately 80% were associated with sleeping >7.5 hours. Indeed, if long sleep was truly the cause of all deaths with which it has been associated in epidemiologic studies, it would be the fourth leading cause of death in the United States.4

Notwithstanding associations of sleep duration with health and longevity, it is commonly implied that at least 8 hours of sleep per night is needed for well-being and daytime functioning; however, studies have failed to support this assumption. For example, a recent analysis found that 6- to 8-hour sleepers were more “successful” by several criteria, including salary, family, and job compared with people who slept >8 hours. There is no compelling evidence that sleep duration is associated with quality of well-being.5,6 Even complaints about sleep are more common among both long and short sleepers.7

Of course, epidemiologic associations cannot prove causality. The associations might be explained by some other heretofore uncontrolled factor. On the other hand, extensive control could result in underestimation of risk. For example, if chronically spending too much time in bed elicits depression (we know that acute sleep extension can leave us feeling poor), control for depression could underestimate the risk of long sleep. Randomized studies of moderate sleep restriction in long sleepers are needed; however, before large-scale studies of this sort can begin, there should be pilot studies demonstrating feasibility of sleep restriction in long sleepers.

The Facts
Meanwhile, evidence for increased mortality and morbidity associated with long and short sleep continues to accumulate. Several recent studies presented at the June 2005 Association of Professional Sleep Societies conference in Denver will be briefly reviewed.

Hall et al reported the results of a prospective 5-year study of 3,075 people participating in the Health, Aging, and Body Composition Study.8 A unique aspect of the study compared with other epidemiologic investigations of this topic was that 42% of the sample was black. Cox proportional hazard models, adjusted for age, gender, race, socioeconomic status, BMI, and health status, indicated that those who reported sleeping <6 hours (hazard ratio 1.56; CI, 1.11-2.18), and particularly those who reported sleeping >8 hours (hazard ratio 1.93; CI, 1.28-2.79), had significantly higher mortality than those sleeping 6 to 8 hours. Moreover, sleeping <6 hours and >8 hours was associated with a significantly greater prevalence of obesity, diabetes, and cardiovascular disease.

In a sample of 3,127 older women (83.6±3.8 years) derived from the Study of Osteoporotic Fractures, Goldman et al9 found that those in the highest (>7.6 hours) and lowest quartiles (<6 hours) of actigraphically recorded total sleep time reported greater difficulty in performing activities of daily living than those in the middle quartiles for total sleep time. Another interesting outcome was that the longest and shortest sleepers demonstrated slower walking speeds than those with average total sleep time. Slower walking speed is a significant predictor of falls in older adults. These associations were found after controlling for age, race, and other covariates associated with sleep or function. The study is noteworthy because it confirms an association of morbidity with objectively recorded sleep, whereas other studies of this topic have relied on self-reported sleep.

In a smaller sample (n=373) of nondemented older adults (ages 75-85 years), Harris et al10 found that reported sleep durations of <5 hours or >9 hours were significantly associated with several measures of impaired neuropsychological functioning including nonverbal ability, visual-spatial ability, processing speed, and motor ability, though this association was no longer significant after statistical control for depression.

Brassington et al11 studied college students who report satisfaction with their sleep. Compared with long sleepers (n=110 who slept >8 hours), short sleepers (n=31 who slept <6 hours) reported significantly less depression, anger, and hostility, and significantly more vigor on the Profile of Mood States (POMS) questionnaire. The authors argued that short sleep might elicit increased energy and drive. This argument is consistent with evidence that both acute and chronic long sleep is associated with lethargy.

In a study of 1,412 older adults ages 55 to 84, Britz12 found that those who were still working full time reported significantly less sleep on work nights (mean 6.8 hours) vs weekends (mean 7.2 hours), whereas those who were retired or working part time had no differences; however, the full-time workers reported significantly less difficulty falling asleep and waking up during the night. Retirement could be an important explanation for why many older adults spend more time in bed, but sleep worse than younger adults. It is known that excessive time in bed can result in sleep fragmentation; that sleep restriction can be a very effective tool for chronic insomnia; and that this effect has been particularly well documented in older insomniacs. Thus, older adults might be the most appropriate target for examining the influence of moderate sleep restriction on morbidity and mortality.

Shawn D. Youngstedt, PhD, is assistant professor at the Department of Exercise Science, Norman J. Arnold School of Public Health, University of South Carolina, Columbia.

References
1. Youngstedt SD, Kripke DF. Long sleep and mortality: rationale for sleep restriction. Sleep Med Rev. 2004;8:159-174.
2. Tamakoshi A, Ohno Y. Self-reported sleep duration as a predictor of all-cause mortality: results from the JACC study. Sleep. 2004;27:51-54.
3. Kripke DF, Garfinkel L, Wingard DL, Klauber MR, Marler MR. Mortality associated with sleep duration and insomnia. Arch Gen Psychiatry. 2002;59:131-136.
4. Minino AM, Smith BL. Deaths: preliminary data for 2000. CDC Natl Vit Stat Rep. 2001;49:1-40.
5. Pilcher JJ, Ott ES. The relationships between sleep and measures of health and well-being in college students: a repeated measures approach. Behav Med. 1198;23:170-178.
6. Jean-Louis J, Kripke DF, Ancoli-Israel S. Sleep and quality of well-being. Sleep. 2000;23:1115-1121.
7. Grandner MA, Kripke DF. Self-reported sleep complaints with long and short sleep: a nationally representative sample. Psychosom Med. 2004;66:239-241.
8. Hall M, Naydeck BL, Stone KL, et al. Association between sleep duration and mortality is partly mediated by higher levels of inflammation in older adults. Sleep. 2005;28S:A118.
9. Goldman SE, Stone KL, Ancoli-Israel S, et al. Actigraphic measures of sleep are associated with decreased daytime functioning in elderly women. Sleep. 2005;28S:A110.
10. Harris S, Schmutte T, Levin R, Zweig R. Nonverbal correlates of poor sleep in older adults without dementia. Sleep. 2005;28S:A113.
11. Brassington GS, Tucker J, Wooley M, Picola J, Posell-Wilson J, Hicks RA. Habitual sleep duration and mood states in college students who are satisfied with their sleep. Sleep. 2005;28S:A315.
12. Britz PA. Retirement and sleep behavior: what’s work got to do with it? Sleep. 2005;28S:A117.