Sleep and health: beyond sleep duration and sleepiness?

Numerous epidemiological studies have demonstrated associations between self-reported sleep duration and health outcomes, including cardiovascular disease, diabetes, weight gain and all-cause mortality (e.g. Cappuccio et al., 2010, 2011; Castro-Costa et al., 2011; Gallicchio and Kalesan, 2009; Lyytikainen et al., 2011; Magee et al., 2012). Although in some of these studies aspects of sleep quality were considered (Cappuccio et al., 2010; Lemola et al., 2011), and sometimes even statistically controlled for (Mesas et al., 2011), in general the emphasis has been on self-reported sleep duration as assessed by a single question. There is little doubt that sleep duration is an important sleep parameter, but in the context of sleep and health it is to be expected that other aspects of sleep are also important. This is illustrated clearly by considering sleep apnoea, where sleep duration is not necessarily compromised, and many negative health outcomes, including all-cause mortality, have been reported, including community-based studies (Marshall et al., 2008), although some paradoxical contributions of apnoea to survival have also been reported (Lavie and Lavie, 2009). In this issue of JSR, two papers specifically consider the contribution of sleep ‘quality’. Grandner et al. (2012) use the concept of general sleep disturbance, operationalized as difficulty falling asleep, staying asleep or sleeping too much, and applied this to a large data base to which 138 201 individuals provided complete data. When the data were separated based on the frequency of disturbed nights (fewer than six nights or six nights or more in the past 2 weeks), several associations between sleep disturbance and health outcomes, such as obesity, diabetes, coronary artery disease and previous myocardial infarction emerged, even when correcting for sociodemographic, socioeconomic and health risk covariates. Although this very large study emphasizes sleep disturbances rather than sleep duration, it cannot separate the effects of sleep duration from sleep disturbances and other attempts are required to understand more clearly which sleep parameter mediates the sleep–health association. Jackowska et al. (2012) took an alternative approach in investigating the contribution of sleep problems to health, as presented in their paper in the current issue. The primary question was whether sleep quality associates with heart rate variability during the working day. Although the sample was relatively small (199 working women), the measures were interesting and some of them somewhat novel. Sleep problems were assessed with the Jenkins Sleep Problems Scale and heart rate variability was assessed over a 24-h period, i.e. including both a sleep period and a working day. Other studies have shown that reduced heart rate variability is a predictor of cardiovascular disease and mortality. In order to assess to what extent the participants’ affective state may contribute to or confound any association between sleep problems and heart rate variability, the authors applied the day reconstruction method, which has advantages over simple questionnaires (Kahneman et al., 2004). The results were relatively clear: poorer sleep quality associates with lower heart rate variability during the working day (but not significantly so during the sleep period). Interestingly, controlling for general affect and stress measures did not negate this association, suggesting that sleep quality itself may be important. Whereas dependent variables such as obesity and cardiovascular disease are reported frequently in studies on sleep duration, and daytime sleepiness is an important dependent variable in many studies on sleep disorders, other outcome variables may also be extremely relevant. Recently reported measures include optimism in children (Lemola et al., 2011), flexible implementation of task goals in new parents (Plessow et al., 2011), falls in older adults (Mesas et al., 2011) and psychomotor slowness in the general population (Kronholm et al., 2011). In the current issue, we can read how performance on the Sustained Attention To Response Task (SART) does not differ over a wide range of disorders of excessive sleepiness (Van Schie et al., 2012). Interestingly, no correlation between overall error score on this performance task and multiple sleep latency was reported, suggesting that sustained attention to response and sleep propensity are, to some extent, separate phenomena. Nemeth et al. (2012) studied obstructive sleep apnoea patients and compared their performance on an implicit sequence learning task, which is a learning task thought to be dependent on basal ganglia, to the performance of a control group. They found no difference in performance on this task, even though the patients were impaired on a listening span task. This diversity of outcome measures and their differential responses in various conditions and sleep disorders will enhance our understanding of the nature of the consequences of insufficient and disrupted sleep. Unfortunately, there may not yet be an agreement on a standard set of behavioural outcome measures or waking performance measures in research studies in sleep disorder patients. Such a standard would greatly facilitate comparison across studies. Fortunately, some progress has been made on standardizing the procedures to be used on the diagnosis and treatment of sleep disorders in Europe, as we can read in the consensus statement published in the current issue (Fischer et al., 2012).