Sleep duration and mortality: the effect of short or long sleep duration on cardiovascular and all-cause mortality in working men and women

Abstract

Background: There is evidence to suggest that insufficient sleep may have an adverse effect on physical and psychological health. Previous studies have reported that when adjusting for major risk factors for cardiovascular disease and a number of demographic and social variables, sleeping 7–8 h each night is associated with lower mortality. These studies, however, have excluded any consideration of stress, which is known to be related to a number of behavioural risk factors for disease and, like sleep, may influence neurochemical, hormonal and immunological functioning.

Methods: This study revisits the associations between sleep duration, cardiovascular disease risk factors and mortality, taking into account the perceived stress of individuals. The data come from a cohort of working Scottish men and women recruited between 1970 and 1973; approximately half of the cohort was screened for a second time, 4–7 years after the baseline examination.

Results: For both men and women, higher self-perceived stress was associated with a reduction in the hours of sleep reported. The pattern of mortality from all causes and the pattern of mortality from cardiovascular disease were consistent for both men and women. When sleep was measured on one occasion only, the risk of dying was reduced for men sleeping more than 8 h in every 24 h compared with those sleeping 7–8 h over the same period. This was after adjustment had been made for age, marital status, social class, cardiovascular risk factors and stress. The risk of dying was increased for women sleeping less than 7 h in every 24 h compared with those sleeping 7–8 h over the same period, after similar adjustments. When the data from the 1st and 2nd screening were considered longitudinally, both men and women who reported that they slept less than 7 h on both occasions that they were questioned, had a greater risk of dying from any cause than those who had reported sleeping 7–8 h at both screenings, after adjusting for age, marital status, social class and stress.

Conclusions: Short sleep over a prolonged period may be associated with an increased risk of mortality: men and women who reported sleeping fewer than 7 h in 24 on two occasions between 4 and 7 years apart, had greater risk of dying from any cause over a 25 year period than those who reported sleeping 7–8 h on both occasions that they were questioned.

Introduction

We sleep approximately 1 1/2 h less today than people did a century ago, with ‘normal’ sleep duration now being an average of 7 1/2 h in each 24. In contemporary society, many people choose to forego sleep in order to accommodate changes in their daily schedules or to prioritise other activities. Sleep duration, for some people, has come to represent a lifestyle factor [1], modifiable in much the same way as other health-related behaviours.

There is evidence to suggest that insufficient sleep may have adverse effects on cardiovascular, endocrine and immune function, as well as having a negative impact on mood and affect [1], [2], [3], [4], [5], [6], [7]. Further, studies have reported both short and long-sleep duration to increase total mortality risk [8], [9], [10], [11], [12]. Ref. [10] for example, reanalysed data from the epidemiological study of the American Cancer Society, which included over one million American subjects, and reported that under 5 h of sleep, or 10 h or more, were almost as strong predictors of mortality as reports of having ‘ever had’ diabetes, heart disease, stroke and/or high blood pressure. The increased mortality rates were seen in both sexes and in all age groups and remained stable after adjustments were made for reported insomnia, sleeping pill use and having ever had four major illnesses. The authors noted that the increased mortality associated with reported long sleep was more pronounced than that associated with reported short sleep, but concluded that further research was necessary to determine the causal factors. Of note is the fact that sleep patterns over a period of time were not analysed.

The majority of research has focused on the mortality risk of short or long sleep as a single factor. An exception to this has been the 9-year mortality follow-up of the Alameda County Study [9] which analysed a number of co-variables simultaneously and together in a summary index. This analysis showed that when adjusting for the other major risk factors for cardiovascular disease, and a number of demographic and social variables, never smoking, regular physical activity, low alcohol consumption, average weight status and sleeping 7–8 h each night were all associated with lower mortality from all causes.

An important omission from the Alameda County Study, however, has been any consideration of stress. Psychological stress is known to be related to a number of behavioural risk factors for disease [13], [14], [15] including poor sleep [5] and, like sleep, stress may influence neurochemical, hormonal and immunological functioning [16]. In this paper, therefore, we explore the relationship between conventional risk factors for disease, sleep, stress and mortality within a workplace-based study of Scottish men and women followed up over a 25 year period.