People Live Longer Limiting Sleep to Seven Hours, Study Says
Feb. 3, 2004 - The battle of the researchers over how much sleep humans need continues with the lastest study saying people live longer if they sleep seven hours, rather than the often recommended eight hours. - or less than 4.5 hours.
An editorial, titled “Do We Sleep Too Much?” in the February 2004 issue of the journal SLEEP by Daniel F. Kripke, M.D., professor of psychiatry, University of California, San Diego (UCSD) School of Medicine, comments on a study by A. Tamakoshi and Y. Ohno in the same issue of the journal, as well as the results of two other studies.
The Editorial from SLEEP
How Much Should An Adult Sleep?
The JACC study from Japan, reported in this issue of SLEEPby Tamakoshi, Ohno, and colleagues, suggested:
1) The best survival is experienced by those who sleep 6.5-7.5 hours on weekdays.
2) The mortality risk of those who sleep more than 7.5 hours is of more concern than the risk of those who sleep less than 6.5 hours.
Even those who reported sleeping 8 hours had greater mortality risk than those who slept an hour less. Although these conclusions might surprise clinicians, the JACC data are fully consistent with recently-reported results from the Nurses’ Health Study (NHS).2-4 and the Cancer Prevention Study II (CPSII).5
The consistency of results among these three enormous studies, conducted with varying methodologies on two continents and over two decades, suggests that they are likely to prove reliable. A dozen smaller studies have supported the general findings of the large studies. No persuasive epidemiologic evidence contradicts them.
A strength of the JACC study was that it specifically asked about average sleep duration on weekdays, clarifying what was meant. The average sleep duration was 7.58 hours for men and 7.12 hours for women and increased somewhat with age.
For men in JACC, the lowest mortality was 1 hour less than the mode, as was the case for men and women in CPSII. There were fewer nurses who slept 8 hours, possibly due to nursing duties, so the NHS mode of 7 hours was also the point of minimal mortality, as it was for women in JACC.
In JACC, those who slept 8 hours had very significantly higher mortality than those who slept 7 hours, regardless of the extent to which the data were adjusted for comorbidities.
The same was true in NHS and CPSII. The fully-adjusted mortality risk ratios for women who slept 8 hours (as compared to 7 hours) were 1.30 (1.12-1.52 95% C.I.) in JACC, 1.13 in NHS, and 1.13 (1.09-1.16 95% C.I.) in CPSII.
These estimates seem in excellent agreement, considering that the age ranges were somewhat different. Moreover, in JACC as in the other studies, those who slept 9 hours or 10 hours and more had progressively increasing mortality risks. The increases in mortality of those who slept 8 hours (7.5-8.5 hours) compared to 7 hours (6.5-7.5 hours) were quite small, but because of the large numbers, would represent a substantial proportion of total population mortality.
The results falsify the widely-circulated hypothesis that it is best to sleep at least 8 hours.
In the JACC models most fully adjusted for comorbidities, those who reported 5 or 6 hours sleep per week night did not have significantly increased mortality compared to those who slept 7 hours. Only those who reported less than 4.5 hours sleep (i.e., 4 hours or less) experienced significantly increased mortality. Data of the NHS were similar.
In the very large CPSII sample, the risk ratios of those who slept 5 and 6 hours were only slightly elevated compared to 7 hours, though statistically significant. The three studies were therefore consistent in showing more mortality risk above 7.5 hours sleep than below 6.5 hours sleep, especially since a higher proportion of the subjects reported sleeping 7.5 hours or more as compared to less than 6.5 hours.
In conclusion, there is more excess mortality associated with sleep above 7.5 hours than below 6.5 hours. Long sleep was associated epidemiologically with more of the population’s excess mortality risk than short sleep.
A strength of the JACC study was an attempt to control for mental stress and depression. Although the questions used to indicate depression did not correspond well with validated depression scales, the JACC scale was to some extent validated by its usefulness for predicting mortality.
In JACC, 2 or more depressive symptoms were observed least among those reporting 7 hours sleep, with depressive symptoms occurring more among both those reporting both shorter or longer sleep. A similar U-shaped relationship of depression to hours of sleep was observed in the Women’s Health Initiative.6 In contrast, in a linear fashion, the more subjects slept, the less they reported mental stress.
If the underlying cause of mortality associated with sleep durations were mental stress or depression, one would expect that control for mental stress and depression would reduce the risk ratios associated with sleep duration. To the contrary, control for mental stress and depression tended to increase the JACC risk ratios associated with short and long sleep, suggesting that reported sleep duration was not a proxy for emotional symptoms.
Similarly, in NHS, adjustment for depression had no substantial effect on risk ratios associated with sleep longer than 7 hours.4
In the JACC study, BMI (body mass index) had virtually no relationship to reported sleep duration. There was little relationship of body mass index to sleep duration in NHS also, though a more complex relationship was observed in CPSII subjects. Since body mass index is an extremely important correlate of sleep apnea (along with gender and age, all of which were controlled), it seems unlikely that sleep apnea could explain associations of reported sleep duration and mortality.
In NHS, the least regular snoring was observed among the 7 hour sleepers, but adjustment for snoring did not appreciably change the mortality risks of sleeping more than 7 hours.2,4
Another strength of the JACC study was an analysis excluding events which occurred within 2 years after entry. This tended to exclude subjects who were moribund at entry, and whose abnormal sleep duration might be attributed to a terminal illness. Because the risk ratios were not materially altered by this exclusion, it is unlikely that the association of sleep durations and mortality can be explained by terminal illnesses.
The JACC study also found that those reporting sleep longer than 7 hours were less educated, less likely to have a spouse, and more likely to have a history of stroke, myocardial infarction, or cancer. By these indices, long sleepers were evidently less well than those who slept 7 hours. Nevertheless, control for these comorbidities did not substantially alter the association of long reported sleep with increased mortality risk.
Were there evidence that those who sleep 8 hours or more had greater wellness, an argument favoring 8 hours sleep might be offered, but JACC found several forms of excess morbidity among those who slept 8 hours or more.
Too much credence has been given to results implying that sleep restriction disturbs glucose regulation, obtained from 11 subjects studied for a couple of weeks without control for viral innoculation effects or experimental order.7,8
The NHS evidence of 70,026 subjects studied for 10 years was more reassuring in the most extensive multivariate model. NHS found no significant relative risk of incident among those sleeping 5, 6, 7, or 8 hours, but a significant increase in relative risk among those sleeping 9 hours or more.3
In view of the mortality and morbidity findings of JACC, clinicians should not recommend that adults sleep 8 hours or more.
The causal mechanisms of mortality associated with reported sleep of 8 hours and more have not been explained. Large epidemiologic studies have all relied on self-reports of sleep duration.
Thus, we are not certain to what extent those who report 8, 9, or 10 hours sleep actually do sleep more. Perhaps respondents tended to report time-in-bed rather than time asleep. Since prolonged time-in-bed is associated with increasing sleep latencies and increasing wake-after-sleep-onset, those who report 8, 9, or 10 hours of sleep might actually experience little increase in physiological sleep.
When presented with evidence that there is increased mortality risk associated with 8, 9, or 10 hours sleep, many with long sleep wish to know if they should voluntarily restrict their sleep. This is an important unanswered question. Eventually, clinical trials of sleep restriction or other interventions focusing on long sleep will be needed, both to test the causal mechanisms for excess mortality and to find how to reverse them. Explaining the risks associated with long sleep will be a high priority.
• News Report on this study - Click Here
• Cancer Prevention Study - Click Here
Editorial—Kripke SLEEP, Vol. 27, No. 1, 2004 13 Do We Sleep Too Much? Comment on Tamakoshi A; Ohno Y. Self-reported sleep duration as a predictor of all-cause mortality: results from the JACC study, Japan. SLEEP 2004;27(1):51-4. Daniel F. Kripke, M.D.Professor of Psychiatry, University of California, San Diego Address correspondence to: Daniel F. Kripke, M.D., Professor of Psychiatry, University of California, San Diego, 9500 Gilman Drive, La Jolla, CA 92093- 0667, ph: 858-534-7131, fax: 858-534-7405, email: DKripke@UCSD.edu diabetes
REFERENCES 1. Tamakoshi A; Ohno Y. Self-reported sleep duration as a predictor of all-cause mortality: results from the JACC study, Japan. SLEEP 2004;27(1):51-4.
2. Ayas NT; White DP; Manson JE; Stampfer MJ; Speizer FE; Malhotra A; Hu FB. A prospective study of sleep duration and coronary heart disease in women. Arch Intern Med. 2003;163:205-209.
3. Ayas NT; White DP; Al-Delaimy WK; Manson JE; Stampfer MJ; Speizer FE; Patel S; Hu FB. A prospective study of self-reported sleep duration and incident diabetes in women. Diab Care. 2003;26:380-384.
4. Patel SR; Ayas NT; White DP; Speizer FE; Stampfter MJ; Hu FB A prospective study of sleep duration and mortality risk in women. Sleep 26 [Abstract Supplement], A184. 2003.
5. Kripke DF; Garfinkel L; Wingard DL; Klauber MR; Marler MR. Mortality associated with sleep duration and insomnia. Arch Gen Psychiatry. 2002;59:131-136.
6. Kripke DF; Brunner R; Freeman R; Hendrix S; Jackson RD; Masaki K; Carter RA. Sleep complaints of postmenopausal women. Clinical Journal of Women’s Health. 2001;1:244-252.
7. Spiegel K; Leproult R; Van Cauter E. Impact of sleep debt on metabolic and endocrine function. Lancet. 1999;354:1435-1439.
8. Spiegel K; Sheridan JF; Van Cauter E. Effect of sleep deprivation on response to immunization. JAMA. 2002;288:1471-1472.
