Sleep hours 'linked to death'

“Women who get between five and six-and-a-half hours sleep a night could live longer,” reports the Daily Express. The news is based on the results of a long-running study of 459 women aged 50 to 81.

This research initially assessed the women’s sleep patterns over one week, using wrist-mounted activity monitor worn at night. Researchers then followed the women for up to 14 years to see how their sleep patterns might have influenced their chances of survival. The scientists found that those women who had slept longer or shorter hours were more likely to have died compared to women who had slept a moderate length. However, as with all studies of this type, it is difficult to say for certain whether sleep directly caused differences in survival rates. Also, the study’s findings may not apply to men or to people under 50 years old.

Although this and other studies seem to suggest that sleeping too much or too little is associated with an increased risk of death, it is not possible to say why this might be, nor whether changing your sleep patterns could influence your longevity.

Where did the story come from?

The study was carried out by researchers from the University of California and the Jackson Hole Center for Preventive Medicine in Wyoming. It was funded by the US National Institutes of Health, and it was published in the peer-reviewed journal Sleep Medicine.

The Daily Express and Daily Mail reported this study. Both summarised the results accurately.

What kind of research was this?

This was an analysis of data from a long-running cohort study called the Women’s Health Initiative (WHI). It aimed to assess the relationship between length of sleep and the risk of death during the follow-up period.

Several previous studies have suggested that sleeping for a long time (e.g. more than 7.5 hours) or for a short time (e.g. less than 6.5 hours) are associated with an increased risk of death compared with an intermediate amount of sleep. However, these studies mainly used subjective reports of sleep, where a person reports their own sleep patterns. The current study wanted to assess whether objectively measured sleep length was also related to risk of death. A previous study looking at objectively measured sleep did not find any evidence of a link.

This type of study design is appropriate for addressing the question asked, as it would not be feasible to use a study design that would randomly assign people to differing amounts of sleep over a long period of time.

As with all observational studies, the main danger is that factors other than the one being studied may be influencing the results. For example, if people who slept for only short periods or long periods also had unhealthier lifestyles, this could be influencing their risk of death rather than their sleep patterns. It is important that this type of study takes this possibility into account.

What did the research involve?

Between October 1995 and June 1999, the researchers asked 451 women taking part in the WHI study to attach to their wrists an activity monitor called an ‘actigraph’. They used the data from these actigraphs to identify the periods when the women were asleep. They then followed the women for up to 14 years to see who died and, if so, when. The researchers then carried out analyses to see if there was a relationship between death and sleep duration.

The researchers selected a sample of women to take part. It included a high proportion of older women, and of women reporting six hours sleep or less, or eight hours sleep or more. This was intended to increase the likelihood that they could detect an effect of sleep length on risk of death if there was one. The women’s average age at the start of the study was 67.6 years (range 50 to 81 years).

The women filled in sleep questionnaires and had psychiatric interviews. The researchers also had access to the questionnaires the women had filled at the start of the original WHI study about their health and lifestyles. All participants wore the actigraph on their wrist for seven days and nights. They also completed a sleep diary and estimated their sleep duration over these seven days. Most participants had urine samples collected over 24 hours, and they wore an oxygen saturation monitor for three nights to identify any sleep apnoea.

The women were mailed annual questionnaires and contacted by telephone until 2005. Any deaths were indentified in this way. In 2009, any additional deaths were identified using the Social Security Death Index. The final set of analyses used follow-up data available for 444 women, 98% of the study population.

The analyses compared the survival of women who slept for different lengths of time. For example, they compared women with 300–390 min of sleep against women who had less than 300 min or longer than 390 min. The researchers took into account factors that could affect results (potential confounders). This included age, history of high blood pressure, diabetes, heart attack, cancer and major depression.

What were the basic results?

Among the 444 women with follow-up data, 86 women had died over an average of 10.5 years of follow-up. According to actigraph readings, the women slept an average of about six hours, which was shorter than the average length of sleep according to the women’s estimates in their sleep diaries, which was 6.88 hours.

The researchers estimated that women who slept less than five hours a night had a 61% chance of survival at the end of follow-up, compared with 78% among those who slept longer than 6.5 hours, and 90% for those who slept between five and 6.5 hours a night. When taking into account all potential confounders, there was a significant link between sleep duration and risk of death. However, the link was only just statistically significant when looking at sleep duration as a continuous outcome, i.e. the relationship seen across all sleep durations.

Medical factors such as high blood pressure, diabetes, having previous heart attack or cancer, or having major depression at the start of the study appeared to have a greater influence on risk of death.

How did the researchers interpret the results?

The researchers concluded that their study confirmed the “U-shaped” relationship between sleep duration and risk of death. Short and long sleep durations were both associated with an increased chance of death compared with intermediate durations. They say that more research is needed to identify approaches that might prevent this increased risk of death.

Conclusion

This study’s strengths include the use of an objective measure of sleep, and the long period of follow-up. It supports the findings of other studies with subjective sleep measures, which suggest that very long or very short sleep duration is linked to an increased risk of an earlier death. There are some important points to note:

• The data is for women aged 50-and-over only, and may not apply to men or to younger age groups.
• The actigraph measures movement, so it would not be able to tell the difference between someone lying still while awake and someone lying still while asleep. This may have caused some inaccuracy in the estimation of how long people were asleep, although possibly less than if a subjective measure of sleep was used. The researchers acknowledge that using actigraphs may over- or under-estimate sleep compared to the gold standard way of measuring sleep (called polysomnography).
• The study only measured sleep using the actigraph for one week at the start of the study. Women’s sleep in this period may not have been representative of their sleep patterns over a lifetime.
• It is not possible to say whether duration of sleep is itself affecting the risk of death, or whether another unknown factor is behind both the different sleep patterns and risk of death. Although researchers took into account a number of factors that could affect risk of death, these or other unmeasured factors could still be having an effect.
• Although studies subjectively measuring sleep duration have found a link, a study objectively measuring sleep using the gold standard objective (called polysomnography) did not find a link between short sleep and increased risk of death.

How long we sleep may be affected by many factors, including our internal body clock, our jobs and families, lifestyles, the environment in which we sleep, and stress levels.

Although this and other studies have suggested that sleep duration is linked to risk of death, it is not yet possible to say whether you can improve your longevity by simply modifying how long you sleep, without changing any other factors.