Sex and age differences in the relationship between sleep duration and hypertension

Quality and duration of nocturnal sleep have been reported as factors able to affect the health status in a population [1]. However, the specific relevance of sleep patterns in relation to the cardiovascular risk profile has rarely been considered, in spite of recent data emphasizing the impact of sleep disturbances on the prevalence and severity of cardiovascular diseases [2]. In particular, sleep features and sleep disorders seem to play an important role in determining blood pressure (BP) levels both in the office and over the 24 h, and in modulating the day–night BP profile [3–6], which can have an impact on the prognosis of hypertensive patients [7–9]. Recent evidence indicates that both increased and reduced sleep duration (i.e. a sleeping period equal or shorter than 6 h and equal or longer than 9 h, respectively) may be associated with an increased risk of cardiovascular disease and in particular hypertension [10–12]. Moreover, the combination of an abnormal sleep duration with an alteration of the physiological day–night BP fluctuations is reported to further increase cardiovascular risk. In particular, the association of a BP increase at night (so called ‘riser’ pattern, defined as mean night-time SBP exceeding daytime SBP) with a short sleep duration seems to be a strong predictor of future cardiovascular disease, independent of average ambulatory BP levels [10]. Data are available that experimentally and acutely induced sleep deprivation can raise BP in normotensive animals and individuals, and worsen the degree of hypertension when already present, although these findings are still a matter of debate [13–19]. In particular, the relationship between usual sleep duration and the occurrence of chronic hypertension is still controversial [20,21]. The Sleep Heart Health Study [22] showed that usual sleep duration both above or below the median range of 7–8 h per night is associated with an increased prevalence of hypertension, an association which became particularly frequent when sleep time was less than 6 h per night. However, not only changes in sleep duration but also changes in sleep quality and in the patterns of breathing during sleep may affect the prevalence of cardiovascular problems and in particular of hypertension. Indeed, a progressively poor quality of nocturnal sleep and an increasing severity of an obstructive sleep apnea (OSA) condition have been found associated with a more frequent nondipper behavior in the hypertensive population [2]. Moreover, a significantly higher prevalence of OSA and apnea-related sleep fragmentation has been reported in hypertensive patients and in particular in drug-resistant hypertensive patients [23]. It is noteworthy that OSA patients with daytime sleepiness may be particularly exposed to cardiovascular problems because of reduced baroreflex sensitivity at night, which is associated with increased sympathetic activity and may thus contribute to an increase in BP levels [24,25]. Very recent data, however, suggest that patients with resistant hypertension also have shorter total and rapid eye movement (REM) sleep times and lower sleep efficiency independently of OSA [26], emphasizing the important role of sleep characteristics per se, with or without sleep-related breathing disorders, in relation to the severity of hypertension. The mechanisms through which sleep disturbances, including a reduced sleep duration, may lead to an increase in BP levels and alterations in the 24-h BP profile are still partly unknown. A link between sleep deprivation/fragmentation and metabolic syndrome has been hypothesized to play a role in this context [27,28]. In fact, sleep restriction may be associated with an impaired glucose tolerance, dyslipidemia and increased prevalence of obesity, together with overactivity of the renin–angiotensin–aldosterone system, overactivity of the sympathetic cardiovascular drive, blunted autonomic cardiovascular modulation over 24 h, renal impairment, endothelial dysfunction and increase in inflammatory indices [16,29–34]. An interesting observation made in two recent studies [35,36] is that sleep deprivation might have a stronger effect on BP levels and cardiovascular and metabolic problems in women than in men. In fact, the relationship between self-reported sleep duration and body composition, with a larger degree of fat content, may be stronger in women than in men, an inverse relations between sleep duration and BMI or waist circumference being observed only in women [37]. In addition to sex, age also seems to affect the cardiovascular changes induced by alterations in sleep patterns. Sleep deprivation (quantified as a sleep duration equal or lower than 5 h per night) was reported to be associated with a higher risk of hypertension in middle-aged adults but not in children or elderly individuals [20,38,39]. In this issue of the Journal of Hypertension, Stranges et al. [40] provide additional data on the complex interaction between sex, age, menopausal status and the cross-sectional association between sleep duration and prevalence of hypertension. This is done in a large population-based sample, including 3027 white men (43.5%) and women (56.5%)) from the Western New York Health Study, performed between 1996 and 2001. Hypertension was defined as a SBP or DBP of at least 140 or at least 90 mmHg, respectively, or based on the regular use of antihypertensive medication. On a multivariate analysis, short sleep duration (<6 h of night sleep) was associated with a significantly increased risk of hypertension compared with a sleep duration of at least 6 h per night, in women [odds ratio (OR) = 1.66, 95% confidence interval (CI) 1.09–2.53], but not in men (OR = 0.93, 95% CI 0.62–1.41). The additional novel contribution of this study is that, when considering also menopausal status in subgroup analyses, the effect of short sleep duration was stronger among premenopausal (OR = 3.25, 95% CI 1.37–7.76) than among postmenopausal women (OR = 1.49, 95% CI 0.92–2.41). Such a stronger association between short sleep duration and hypertension in premenopausal women was independent of socioeconomic status, traditional cardiovascular risk factors and psychiatric comorbidities [40]. It has to be acknowledged that exploring the link of sleep duration with cardiovascular problems, in particular with BP elevation, is a difficult task because of several methodological problems. An important limitation of most of the studies exploring the link between sleep duration and hypertension is that sleep duration was generally quantified on the basis of subjective and retrospective reporting by the individuals under scrutiny, and this was the case also for the study by Stranges et al. [40]. Indeed, a direct quantitative detection of sleep duration and poor sleep quality (in particular of sleep fragmentation) has rarely been obtained in the available studies. An additional problem with subjective sleep reports is that they are unable to define the determinants of an altered sleep duration, which may include insomnia, movement disorders during sleep, psychiatric disorders, sleep-related breathing disorders or all. Scanty information is also available on the long-term reproducibility of the reported sleep duration, when assessed over a short observation period only. In particular, little is known on how stable sleep duration might be over time, and whether different sleep patterns characterize the nights following working days as compared with weekends. In this context, the possible effects on night sleep quality and duration exerted by the type of work and by the occurrence of shift work have rarely been quantified with appropriate methods. Finally, sleep duration may have different BP effects in different individuals, also because of the large heterogeneity of individual sleep behaviors in a population. In fact, a short sleep duration may have a different impact on BP in individuals characterized by being usually long sleepers as compared with usual short sleepers. On the background of these difficulties, the findings reported by Stranges et al. [40] should be regarded as observational associations, which do not imply a causal link between sleep patterns and BP levels. As a consequence of this, the observed association between short sleep duration and increase in BP levels should not be taken as to necessarily conclude that declining trends in the average sleep length of a general population might be among the causal factors of a general worsening in health conditions and of an increasing prevalence of hypertension. The need for such a cautious approach is further underlined by the finding, in previous studies [10–12], of a similar increase in cardiovascular risk associated with opposite sleep behaviors, that is, both with too short and too long sleep durations. Another difficulty in interpreting the results obtained in this field comes from the fact that the cardiovascular effects of different sleep patterns have been reported to depend not only on total duration of sleep but also on sleep quality and in particular on the occurrence of sleep disordered breathing. However, these additional features are rarely assessed in studies on the cardiovascular effects of different sleep patterns because this would require a polysomnographic examination to be performed, which is difficult to implement when investigating a large group of individuals, as in the case of the study by Stranges et al. [40]. The methodological difficulties characterizing the studies exploring the link between sleep duration and BP levels are not only related to the assessment of sleep features but also to the approach selected to measure BP. Most studies, including that by Stranges et al. [40], only focused on clinic BP measurements, which are known to suffer from important limitations, such as intrinsic inaccuracy of the technique, observer's bias and digit preference, ability to provide only spot assessment of a parameter such as BP, characterized by a pronounced intrinsic variability, and possible interference by a white-coat effect [41]. In fact, use of isolated clinic BP measurement may sometimes lead to an incorrect classification of patients' BP conditions. Moreover, this approach does not allow any information on night-time BP to be obtained, and may thus fail to identify the BP changes accompanying different sleep patterns at night. Conversely, out-of-clinic BP monitoring techniques, which include home and 24-h ambulatory BP monitoring, have been shown to offer a more precise and reproducible estimate of the BP levels prevailing in a given individual. In particular, ambulatory BP monitoring may represent the method of choice for studies between sleep patterns and BP behavior because it offers the unique advantage of providing a description of the whole 24-h BP profile, including BP during night sleep [41]. Failure to adopt out-of-clinic BP monitoring techniques may thus have reduced the accuracy of the results obtained by a number of studies investigating the relation between sleep features and BP levels, including the study by Stranges et al. [40], in which the diagnosis of hypertension was defined either on the basis of elevated clinic BP (average of three measurements) or on the regular use of antihypertensive medications. Other possible limitations of the study of Stranges et al. [40] are related to the suboptimal participation rate (59.5%), which may have introduced a bias in patients' selection and restricted the generalizability of their findings to an unselected population. Moreover, also the finding of a stronger relation between sleep duration and BP levels in premenopausal than in postmenopausal women needs to be interpreted with caution because of the possible interference by a number of confounders such as the higher prevalence of hypertension, OSA and insomnia in the latter group. In spite of the above-mentioned methodological issues, however, the study by Stranges et al. [40] has a clinical relevance, insofar it provides a significant contribution to the ongoing discussion on the relation between sleep features and the level of cardiovascular risk, and might pave the way for future investigations in this field. In fact, some ground still has to be covered before definite links between short sleep duration and cardiovascular problems such as hypertension could be confirmed. Issues to address while exploring the link between sleep deprivation and blood pressure (BP) levels are: acute vs. chronic effects, effects of sex, effects of age and menopausal status, effects of comorbidities (other cardiovascular risk factors, interference by sleep-related breathing disorders), methods of BP assessment (clinic vs. home or 24-h ambulatory BP measurements) and subjective reports on sleep habits vs. objective polysomnographic examination. Clinically important issues to address would be a better understanding of the interindividual variation in the vulnerability to short sleep, which might be at least in part genetically determined, an improved understanding of what the influence of sleep recovery could be in this context and a better definition of the role exerted by sleep disorders. In particular, future studies should explore whether insomnia, a disorder frequently referred to as a state of increased arousability, could be a state marker for a hyperaroused autonomic nervous system, as part of a complex syndrome associated with hypertension.