Too little or too much sleep, that is sleeping less or more than 6 to 8 hours on average, is linked to premature death says a new study that pooled data on more than 1.3 million participants worldwide.

You can read about the meta analysis by researchers at the University of Warwick in the UK and the Federico II University Medical School in Naples, Italy, online in the 1 May issue of the journal Sleep.

First author Professor Francesco Cappuccio, who heads the Sleep, Health and Society Programme at the University of Warwick, and colleagues, found that sleeping less than 6 hours a night was linked to a 12 per cent higher chance of premature death compared to sleeping the recommended 6 to 8 hours.

They found that consistently sleeping 9 or more hours a night was also linked to an increased risk of death, but they concluded this was probably for different reasons than not sleeping enough.

In a press statement, Cappuccio, who is also Consultant Physician at the University Hospitals Coventry and Warwickshire NHS Trust, suggested that:

"Whilst short sleep may represent a cause of ill-health, long sleep is believed to represent more an indicator of ill-health.”

"Modern society has seen a gradual reduction in the average amount of sleep people take, and this pattern is more common amongst full-time workers, suggesting that it may be due to societal pressures for longer working hours and more shift-work," he added, explaining that on the other hand, diminishing health is often accompanied by longer sleep time.

For the study, Cappuccio and colleagues pooled data from 16 prospective studies from the US, UK and other European countries, and countries in in East Asia to examine the relationship between duration of sleep and all-cause mortality and to estimate the risk.

To find the studies they systematically searched published literature held in various databases (eg MEDLINE from 1966 to 2009), and from other sources, and included studies not published in English. They only selected prospective studies that followed participants for over three years, recorded amount of sleep at the start of the study, and counted deaths from all causes.

The 16 studies that met these criteria provided 27 independent cohort samples covering more than 1.3 million men and women followed for 4 to 25 years and recorded over 112,000 deaths. Sleep duration had been assessed by questionnaire and mortality from death certificates.

The researchers extracted the relative risks (RRs) and 95 per cent confidence intervals (CIs) and pooled them. They also took into account possible publication bias and heterogeneity (a measure of consistency of design in the studies: high heterogeneity means you should be careful about drawing overall conclusions from a pooled analysis, as in this study's findings on long duration sleep).

The pooled analysis showed that:

Short duration of sleep was linked to 12 per cent higher risk of death (RR: 1.12; 95 per cent CI from 1.06 to 1.18; P < 0. 01), with no evidence of publication bias, (P = 0.74), with some, but not significant, heterogeneity between studies (P = 0.02).

Long duration of sleep was also linked with a 30 per cent greater risk of death (1.30; 95 per cent CI from 1.22 to 1.38; P < 0.0001), with no evidence of publication bias (P = 0.18) but significant heterogeneity between studies (P < 0.0001).

The researchers concluded that:

“Both short and long duration of sleep are significant predictors of death in prospective population studies.

Cappuccio said that these findings suggest "consistently sleeping 6 to 8 hours a night may be optimal for health", but we need to do more research to find out why sleep is essential to health.

He recommended policy makers and implementers include sleep duration among the risk factors they consider when drawing up public health measures, and professionals cover it when advising or counselling people on how to change lifestyle and behaviour to improve health.

Some experts suggest it may not be a straightforward relationship but a combination of factors.

Professor Jim Horne, of the Loughborough Sleep Research Centre, told the BBC that sleep is "just the litmus paper", an indicator of health, and is affected by many diseases, including physical and mental conditions like depression.

He also suggested just getting more sleep may not help a person live longer, but agreed that under five hours a night was probably "not right", and may result in drowsiness during the day that increases one's risk of accidents, for example while driving or using machinery.