Skip to main content

11-04-2013 | Geriatrics | Article

Poor health underlies sleep duration and mortality link


Free abstract

medwireNews: Mortality appears to be increased in patients with insomnia only if they have poor pre-existing health, say researchers.

This finding provides some evidence against previously suggested causal mechanisms to explain the relationship between sleep duration and mortally, the researchers note.

Rather, they suggest that "the relationship between sleep duration and mortality is due to residual confounding with poor health at baseline."

Christopher Magee (University of Wollongong, Australia) and colleagues assessed health status and demographic factors in 227,815 adults aged 45 years and older through a self-report questionnaire.

Most of the participants reported sleep durations of 7 hours (23.9%) or 8 hours (40.7%) per night, while 3.6% reported short sleep durations of less than 6 hours and 7.7% long sleep durations of 10 hours or more per night.

Over the 5-year study period, there were 8782 deaths from all causes. The risk for mortality was significantly higher in those reporting long or short sleep durations (after taking into account age and gender), with hazard ratios of 1.43 and 1.27, respectively.

After further adjustment for factors such as marital status, smoking, body mass index, and physical activity, the associations were attenuated but remained significant, at hazard ratios of 1.13 for short sleep durations and 1.26 for long sleep durations.

But when the relationships were separately examined in individuals classified as healthy or less healthy, the latter defined as having one or more chronic conditions at baseline or a physical functioning score on the Medical Outcomes Study scale of less than 75 , they were only significant for the less healthy group. For these individuals, the hazard ratios were 1.20 for short sleep durations and 1.38 for long sleep durations.

The researchers suggest in Sleep Medicine that residual confounding may not have been effectively addressed in previous studies due to health status assessment being limited to the presence or absence of a health condition, and also not considering the severity of that condition.

"Our research is better able to address this, as we assessed health status in relation to preexisting illness and functional limitations, which are more indicative of an individual's overall health," they write.

"Therefore, our results provided supportive evidence that the nature of the association varies by health status and that residual confounding may well account for previous associations between sleep duration and mortality."

medwireNews ( is an independent clinical news service provided by Springer Healthcare Limited. © Springer Healthcare Ltd; 2013

By Lucy Piper, Senior medwireNews Reporter

Related topics