Long Sleep a Greater Mortality Risk Than Short Sleep in Older Adults

To the Editor

We commend Mesas et al. for their excellent paper (JAGS 58: 1870-1877, 2010),¹ “Sleep Duration and Mortality According to Health Status in Older Adults”. While the sleep field and the popular literature has focused a great deal of attention on the risks of short sleep, the Mesas et al. study is consistent with many other epidemiologic studies which indicate that the mortality and morbidity risks are even greater for long sleep.^2,3,4 Moreover, as observed before,² Mesas et al. found a greater prevalence of long vs. short sleep. Indeed, the average respondent reported a sleep duration of 8 hr, which was associated with significant mortality, suggesting a greater public health relevance of long sleep vs. short sleep. Other strengths of the report are an inclusion of a representative sample, consideration of combined nocturnal sleeping and daytime napping, and an extensive in-home interview with participants, which allowed more definitive assessment of reported sleep duration and other factors (e.g., current medication use) than is typical in similar studies.

Notwithstanding rapidly accumulating evidence from perhaps 100 studies showing associations of long sleep with mortality/morbidity, discussion of similar studies has typically focused on the mortality/morbidity risks of short sleep, while dismissing the association of long sleep as some sort of artifact.^5,6 For example, it has been argued that the associations of long sleep with mortality/morbidity are likely to be explained by the opposite direction of putative causality, i.e., morbidity causing long sleep. Contrary to this argument are findings that the associations of long sleep with mortality/morbidity are just as apparent following control for multiple morbidities.² The Mesas et al. study goes one step further in showing that the association with mortality is observed even in samples restricted to apparently healthy individuals at initial assessment. Moreover, our recent research with long sleepers^7,8 found equal tolerance of chronic sleep restriction in apparently healthy individuals and those with existing morbidities, for whom more adverse effects would be expected if morbidity were truly the cause of their long sleep. Rather, the participants almost unanimously reported a long history of long sleep, which presumably preceded their morbidity.

Critics of the association of long sleep with mortality have also raised the issue that data are limited to self-reported sleep duration. However, recent studies have shown similar associations of objectively recorded sleep duration with mortality and morbidity.^9,10

We agree with Mesas et al. that the risks of long sleep might be related to time-in-bed and poor sleep quality. Studies which have distinguished between time-in-bed and sleep duration have found similar risks associated with these variables.^9,10

The authors caution that “21.3% of persons aged 60 and older in Spain sleep at least 10 hours per day”. Moreover, nearly 2/3 of the sample reported ≥ 8 hr of sleep, and 8 hr was associated with significant mortality. However, against the current zeitgeist that we are a sleep-deprived society, this point is far from the message that older adults commonly receive.¹¹ The message they hear is that 8 hr of sleep is necessary for health and that they ought to be able to sleep that long.¹¹

This fear might partly explain why older adults often spend more time-in-bed, but sleep less than young adults.^12,13 We know that excessive time-in-bed can lead to even more sleep fragmentation, which has been associated with morbidity both in epidemiologic studies¹⁴ as well as studies involving experimental induction of the fragmentation.¹⁵

The argument that the long sleep (or long time-in-bed) could not possibly be harmful is becoming more difficult to justify. The Mesas et al. result point to a need for more randomized controlled trials of sleep restriction or extension in older adults.