Sleepless in San Diego
March 15, 2002
Sleepless in San Diego
Abstract & Commentary
Synopsis: Mortality rates have a U-shaped relationship with sleep duration, with lowest mortality for those who slept 7 hours. Self-reported insomnia does not appear to increase mortality.
Source: Kripke DF, et al. Arch Gen Psychiatry. 2002;59:131-136.
This is a prospective, explorative study of mortality rates of about 1 million American Cancer Society (ACS) volunteers aged 30-102 years (mean, 57.5) in 1982. Three questions related to sleep were asked on this extensive questionnaire:
- "On the average, how many hours do you sleep each night?"
- "On the average, how many times a month do you have insomnia?"
- Some measure of past month use of "prescription sleeping pills."
Six years later, data about death or survival were available for 98% of the sample. The study controlled for demographic risk factors, habits, health, and medication use, and undertook standard statistical analyses.
At follow-up, 9.4% of the men and 5.1% of the women had died. The causes of death were overwhelmingly cardiovascular (ischemic heart disease, other heart disease, and cerebrovascular accidents), with breast and colon (not lung) being the leading causes of cancer deaths. The modal sleep duration was 8 hours, and almost half the sample reported a sleep duration of 7.5 hours or more. Only 4.3% of women and 2.6% of men reported insomnia 10 or more times per month. More frequent insomnia and sleeping pill use were reported both by those sleeping less than 7 hours and by those sleeping more than 8 hours a night. Heavier men tended to sleep less, but there was a U-shaped relationship between sleep duration and body mass index for women.
The "big finding" was a U-shaped relationship between sleep duration and survival, with the best survival for those who slept 7 hours. Sleeping longer than 8.5 hours or less than 4.5 hours (for men) or 3.5 hours (for women) was associated with an increased mortality risk of 15%. Broken down for cause of mortality, only cerebrovascular deaths were increased for those getting 8 or more hours of sleep. Participants reporting insomnia had a reduced hazard ratio compared to those who did not. Participants who used prescription sleeping pills had increased mortality hazards compared with those who did not.
Comment by Barbara A. Phillips, MD, MSPH
This paper, which has been widely quoted in the lay press and media, has the sleep community all abuzz. As one colleague put it, "Kripke has just undone 10 years of hard work and education." It’s easy to criticize this paper: the sample is not representative of the entire population (the causes of death and rates of insomnia are different from the population at large, for example), many important variables are undefined (eg, insomnia, kinds of sleeping pills, geographic and racial backgrounds of the participants), naps are not accounted for, and the data are nearly 14 years old. The likelihood remains that long (and short) sleep are markers for underlying health, psychosocial, or lifestyle issues that are the primary cause of mortality. The National Sleep Foundation’s web site points out that "substantial research serves as the basis for the recommendation that adults obtain an average of 7-9 hours of sleep each night, recognizing that each person has an individual requirement for his/her own amount of sleep. If someone sleeps 8.5 hours a night and feels alert and energetic, it would be incorrect to reduce sleep time based on the Kripke article."1
What do we tell our patients and how do we manage our own sleep on the basis of this paper? First, it appears quite unlikely that insomnia per se is going to kill anybody. Insomnia is by far the most common sleep complaint (certainly more prevalent than was reported in this article), and this paper gives us grounds for reassurance of some of our insomniacs. Second, the relationship between sleep duration and mortality was U shaped, not linear. In other words, short sleep (less than 4.5 hours for men or 3.5 hours for women) was also associated with increased mortality. We need sleep! As an aside, the 3.5 to 4.5 hour lower limit is remarkably consistent with the minimum requirement of sleep cited by house officers that they need to function well,2 and with the "core sleep" concept advanced by Jewett and colleagues.3 Further, the "prescription sleeping pills" used in this study (and associated with increased mortality) were likely benzodiazepines, sedating antidepressants, and antihistamines.4 Now that we have safer, more effective hypnotics, this issue needs to be carefully re-examined. Finally, we need to emphasize that sleepiness impairs performance and quality of life. Patients need to know that daily alarm clock use, a difference of more than an hour of sleep on nonwork nights compared with work nights, excessive caffeine use, and lack of alertness during the day are symptoms of inadequate sleep or sleep disorders.
This is a huge study about a universally applicable condition (sleep), and it is difficult to ignore it. Dr. Kripke has done us a favor by stimulating interest, discussion, and (hopefully) carefully-designed studies to further address this issue. Perhaps one of the benefits of this paper will be the result urged by the National Sleep Foundation, "Patients should report to their doctors about the quantity and quality of sleep, and whether they experience daytime sleepiness, insomnia, snoring, pauses in breathing, or excessive movements during sleep."1
Dr. Phillips, University of Kentucky; Director, Sleep Disorders Center, Samaritan Hospital, Lexington, KY, is Associate Editor of Internal Medicine Alert.
References
2. Baldwin, DC. Presented to Sleep, Fatigue and Medical Training workshop. October 28, 2001. Alexandria, Va.
3. Jewett ME, et al. Sleep. 1999;22:171-179.
4. Walsh JK, Schweitzer PK. Sleep. 1999;22:371-375.
You have reached your article limit for the month. Subscribe now to access this article plus other member-only content.
- Award-winning Medical Content
- Latest Advances & Development in Medicine
- Unbiased Content