SEATTLE—Getting eight hours of sleep each night may not only be unnecessary, it might be bad for your health. According to data from the Cancer Prevention Study II, individuals who average seven hours of sleep each night have a lower mortality rate than do those who sleep eight hours or more.

“The lowest [mortality] group actually slept between six and one-half to seven and one-half hours,” said Daniel F. Kripke, MD, a Professor of Psychiatry at the University of California, San Diego. “The major mortality risk associated with habitual sleep duration is among long sleepers, by which I mean those sleeping eight hours or more. Incidentally, calling them long sleepers is accurate, because from the new data we know that the average American on weekdays sleeps about six and one-half hours.”

Specifically, he said, individuals who slept eight hours per night were 12% more likely to die within six years than were those who slept seven hours. The highest mortality rates occurred among men and women who slept more than eight and one-half hours. This group had at least a 15% increased risk of death, said Dr. Kripke. As for short sleep, men who slept less than four and one-half hours and women who slept less than three and one-half hours also had at least a 15% increased risk of death. Dr. Kripke presented his findings at the 16th Annual Meeting of the Associated Professional Sleep Societies.

THE BIG SLEEP

The American Cancer Society collected data for its initial study in 1959. Preliminary results from the Cancer Prevention Study I, published in 1964, included the finding that the best survival rates were experienced by men and women who slept seven hours a night. “It was a very highly consistent and significant result,” said Dr. Kripke. However, he noted, “this study did not receive much attention from the sleep community at the time, because it was published in the cancer literature.” In 1979, after reanalyzing the original study data and controlling for a number of comorbidities, Dr. Kripke and colleagues found that the lowest mortality was again at seven hours, with some increase in mortality associated with short sleep and an even steeper increase with long sleep. Still, he said, many questions remained regarding other risk factors that were not taken into account.

In 1982, Dr. Kripke began collecting data for the Cancer Prevention Study II, which questioned 1.1 million participants between ages 30 and 102 about their sleep duration and frequency of insomnia. The investigators then conducted a six-year follow-up of 98% of the original study cohort, and Dr. Kripke remarked that it has taken essentially since then to study all the data to control for the various comorbidities.

VARIABLE COVARIATES

“In the new study, we were able to control simultaneously for 32 risk factors,” he said. “It was a very big job to do Cox proportional hazard models in a million people with 32 covariates.” These covariates included age, race, education, marital status, and various habits, as well as sleep duration, insomnia, and sleeping pill use. The investigators also examined health predictors such as diabetes and hypertension, along with patients’ use of medications. “These were chosen partly because of univariate analyses,” said Dr. Kripke. “Of course, the studies were mainly designed to look at cancer predictors, so they didn’t have every variable that we might wish,” he added.

After studying all the covariates, Dr. Kripke said that the mortality rates went down. “Control for other risk factors reduced the risk associated with short sleep more than with long sleep,” he said. “The mortality rises more steeply on the long side than on the short side.” Furthermore, a slightly lower mortality rate was found among women sleeping four hours per night than among those sleeping eight hours, though “probably not significantly so,” he added.

IS INSOMNIA GOOD FOR YOU?

“Oddly enough, we found mortality ratios of less than 1.0 with increasing reported insomnia per month,” Dr. Kripke reported. “That is, those who reported insomnia, when we controlled for other covariates, lived longer. I don’t know if I believe that this is a reliable protective effect of insomnia, but that is what the data suggest.” Also noteworthy, he observed, is that “when we controlled for sleep duration and insomnia, those who said they took sleeping pills had a significantly higher mortality.”

Dr. Kripke pointed out that a number of other studies have shown similar results regarding a higher mortality among long sleepers. The studies “used a variety of techniques to measure comorbidity,” he said. “I can assure you that the increased mortality associated with long sleep epidemiologically is a very, very reliable finding. It’s been available in the literature since 1964, and since then nobody has been able to do a study that could explain this association by comorbidities.”

Dr. Kripke acknowledged that his findings may be difficult to reconcile with other studies regarding the risks associated with short sleep duration. He pointed out that most of these other studies have been conducted in the laboratory, however. “Maybe we should respect the epidemiology and look more critically on the other kinds of data,” he said. “It’s interesting that mammals with shorter sleep live longer.”

NR

—Colby Stong

Suggested Reading
Buysse DJ, Ganguli M. Can sleep be bad for you? Can insomnia be good? Arch Gen Psychiatry. 2002;59:137-138.

Kripke DF, Garfinkel L, Wingard DL, et al. Mortality associated with sleep duration and insomnia. Arch Gen Psychiatry. 2002;59:131-136.

Return to table of contents