Age-Related Alterations in Sleep Quality and Neuroendocrine Function: Title and subTitle BreakInterrelationships and Implications

Normal aging is associated with a number of subjective and objective alterations in sleep quality,^{1 - 2} as well as with a variety of changes in endocrine-metabolic functions.³ To date, a precise definition of the chronology of age-related changes in sleep patterns and their possible temporal relationships with changes in neuroendocrine function has yet to be elucidated in healthy men or women. Moreover, the possible clinical and functional significance of such sleep-endocrine changes in the elderly are of great medical and societal interest, even though incompletely understood.

Among the most common sleep disturbances detected in aging individuals are decreased deep (stages 3 and 4 or delta) slow wave (SW) sleep and an increased number and duration of nocturnal awakenings; in contrast, effects of aging on rapid eye movement (REM) sleep appear to be more variable and to occur later in life.⁴ Slow wave sleep is thought to be controlled primarily by a sleep-wake homeostatic process involving 1 or more as yet unidentified neural factors,⁵ whereas REM sleep is proposed to be regulated by the circadian pacemaker located in the hypothalamic suprachiasmatic nucleus.² The effects of aging on sleep regulation differ substantially between men and women. Thus, for example, pronounced decreases in SW sleep occur in early to mid adulthood in men, but not in women.⁵ Of note is that most prior studies of the effects of aging on sleep regulation have simply compared chronobiological outcomes in younger vs older individuals.

Growth hormone (GH) secretion is episodic and is modulated by the pulsatile release of hypothalamic GH-releasing hormone (GHRH) and by the intermittent inhibitory tone of somatostatin. Growth hormone secretion is maximal during puberty, occurring mostly during the night, especially during deep SW sleep. Beginning in early to mid adulthood (ages 20-39 and 40-59 years, respectively), GH production decreases at a rate of about 14% per decade, primarily as a result of a decrease in the amplitude of nocturnal GH pulses. By late adulthood (age 60 years and older), daily spontaneous GH secretion is decreased by 50% to 70% of that exhibited in the third to fourth decades of life.⁶ The acute secretory response of GH to pituitary stimulation with GHRH is reduced in healthy old vs healthy young individuals. This reduced GH secretory responsivity is thought to result from a decrease in the secretion or action of GHRH and from a rise in somatostatinergic tone.⁶ Several GH-releasing peptides (GHRPs) and related nonpeptide GH secretagogues have been synthesized that exert potent stimulatory effects on pulsatile GH release. These novel GH secretagogues exert antisomatostatinergic functional effects mediated via an endogenous GH secretagogue receptor, which, along with a naturally occurring ligand for this GH secretagogue receptor, has recently been identified.^{7 - 8} The physiological role(s) of the naturally occurring GH secretagogue(s), and their contribution to the age-related decline in GH production, remain to be elucidated.

Adrenocorticotropin and cortisol are normally secreted rhythmically in response to discrete pulses of hypothalamic corticotropin-releasing hormone (CRH). These pulses increase to maximal amplitude in the early morning hours and become smaller and less frequent during the late morning and afternoon, leading to a highly reproducible diurnal secretory pattern for adrenocorticotropin and cortisol, with a nadir in the afternoon and evening, and a peak in the early morning. Although earlier studies reported increased, unaltered, and decreased spontaneous cortisol secretion with aging, more recent investigations reveal an increase in cortisol production, especially at the nocturnal nadir, in healthy, nonstressed, nondepressed elderly individuals, perhaps in association with a decreased resiliency of cortisol secretion.⁹ In the elderly, several studies have shown no sex differences in cortisol secretion, whereas others have demonstrated significantly higher unstimulated cortisol production in elderly women, possibly owing in part to differences in cortisol-binding globulin, the principal binding protein for cortisol.⁹

The GHRH-GH and CRH-adrenocorticotropin-cortisol axes are functionally interrelated. Small increases in glucocorticoids stimulate, whereas excessive concentrations of glucocorticoids inhibit, spontaneous and secretagogue-stimulated GH secretion.¹⁰ Growth hormone production is reduced in glucocorticoid deficiency states, normalized after glucocorticoid replacement, and further augmented by glucocorticoid administration at levels equivalent to twice normal cortisol secretion.¹⁰ In contrast, in states of cortisol excess, such as Cushing syndrome or depression, there is an inverse relationship between the secretion of cortisol and that of GH. Aging, as well as pathological GH deficiency, glucocorticoid excess, or both occurring in nonelderly adults, is associated with osteopenia; loss of muscle mass, strength (ie, relative sarcopenia), and exercise capacity; increased total and intra-abdominal fat with glucose intolerance and dyslipidemia increased fragility of skin and blood vessels; poor healing of connective tissue; altered immune function (ie, increased susceptibility to bacterial infection); and diminished quality of life.^{6 ,9 ,11 - 12} Adverse consequences of cortisol excess are more evident at the nighttime nadir than during the daytime peak.¹³ Thus, it has been speculated that age-related decreases in GH and increases in nocturnal cortisol secretion separately and interactively contribute to many, if not all, of the aforementioned alterations in body composition and function in older persons.^{6 ,9}

Sleep-endocrine rhythms are closely interrelated, and sleep is an important regulator of neuroendocrine function.^{14 - 15} For example, it has been observed that sleep-wake homeostasis is the primary contributor to the temporal organization of GH secretion, whereas the influence of circadian rhythmicity is much less important.¹⁵ In young healthy men, nearly 60% to 70% of GH release occurs during delta (stages 3 and 4) SW sleep, whereas in young healthy women, SW sleep and GH release are less closely coupled, and daytime GH secretion is relatively more important than in men.¹⁵ The usual nocturnal sleep-onset GH secretory pulse results from a burst of hypothalamic GHRH release occurring simultaneously with a diminished somatostatin tone, although the nocturnal GH surge can occur prior to sleep onset in both sexes. Young adult men given agents that increase SW sleep, such as γ-hydroxybutyrate, exhibit concomitant increases in nocturnal and 24-hour GH secretion,¹⁶ suggesting that such drugs may be useful GH secretagogues.

Both normal and pathological sleep states in humans are strongly influenced by the reciprocal influences of GHRH and CRH, as well as by the effects of several other neuropeptides.¹⁴ In young healthy men, GHRH, especially when given episodically rather than continuously¹⁷ and at night rather than in the morning,¹⁸ stimulates SW sleep and GH secretion and inhibits cortisol release, whereas CRH exerts the opposite effects. These sleep-endocrine effects of GHRH are reduced in older individuals.¹⁹ Neither GH²⁰ nor GH secretagogues^{21 - 22} significantly alter SW sleep, although the exact effects of the secretagogues are dependent on the route and paradigm of administration. Pulsatile infusions of cortisol increase SW sleep and GH secretion and decrease REM sleep.^{23 - 24}

In this issue of THE JOURNAL, Van Cauter and colleagues²⁵ provide important new information regarding the influence of age on sleep-endocrine functions. In their cross-sectional survey of 149 healthy men aged 16 to 83 years, the largest such group reported to date, the authors convincingly demonstrate that the amount of deep SW sleep decreases by about 80% from young (age 16-25 years) to mid (age 36-50 years) adulthood, in temporal association with a nearly 75% reduction in spontaneous GH secretion, and that SW sleep, independent of age, is a major determinant of 24-hour and nocturnal GH release. During this same time period, the amount of light (stages 1 and 2) sleep increases, whereas REM sleep, time awake, total sleep, and cortisol secretion do not change. In contrast, Van Cauter et al find that REM sleep diminishes from mid to late adulthood (44-83 years), in synchrony with an increase in nocturnal (nadir) cortisol release, with a nonsignificant trend for age-independent effects of REM sleep on cortisol secretion. Concomitantly, the amounts of light sleep and total sleep decrease, while that of awakenings increases. Thus, in healthy men, there are distinctly different time courses in the development of age-related changes in sleep quality, and the latter are closely associated with separate and specific alterations in neuroendocrine function.

Several caveats seem warranted in evaluating the study by Van Cauter et al. First, the results are potentially subject to confounding by cohort and secular effects, which are commonly encountered in cross-sectional investigations of aging.²⁶ Confirmation of the findings in a longitudinal investigation would be of great interest. Second, GH assays of varying sensitivity were used in the data analysis, and GH secretion was expressed in terms of pulsatile rather than total (ie, pulsatile plus interpulse baseline) secretion. Some adjustment of the quantitative, if not qualitative, GH data and their relationships with SW sleep might be expected. Third, androgen status, especially of the older men, was not specified. Given the frequent (30%-50%) age-related decline in circulating total and free testosterone in men within the age group studied,³ the known interrelationships between androgens and the GH axis,²⁷ and the likely influence of androgens on sleep quality, any possible influence of testicular function on the results may have gone undetected. Fourth, the study population was composed of nonobese men in whom there was a significant inverse relationship of body mass index with both SW sleep and GH secretion during waking. The generalizability of the current findings to men with increased total and intra-abdominal fat remains to be established. Fifth, the possibility that sleep quality was influenced by the hormonal milieu, rather than vice versa, was not examined. Nonelderly adults with GH deficiency exhibit decreases in SW sleep that are reversed with GH treatment.²⁸ Moreover, aged individuals may be deficient in GHRH and respond to exogenous GHRH with increases in SW sleep. Finally, this study was performed only in healthy men, and other target populations, such as women and frail elderly persons, need to be studied similarly.

Despite these caveats, there are several implications of the investigation by Van Cauter et al: age-related alterations in sleep quality may contribute to concomitant changes in body composition and function via their influence on neuroendocrine functions; some of the "age changes" are firmly in place and therapeutic interventions should be assessed relatively early in adult life; and measures that augment SW sleep may potentially serve as useful therapeutic GH secretagogues. Clearly, more research into this promising area is indicated, as the potential benefits to the aging population are substantial.