| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Neuropsychiatry, Akita University School of Medicine 1-1-1 Hondo, Akita City, Akita 010-8543, Japan
Address all correspondence and requests for reprints to: Kazuo Mishima, M.D., Department of Neuropsychiatry, Akita University School of Medicine 1-1-1 Hondo, Akita City, Akita 010-8543, Japan. E-mail: mishima{at}psy.med.akita-u.ac.jp.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Several lines of evidence indicate that the age-related advance in the biological clock phase (BCP), which is determined by the primary circadian pacemaker of the hypothalamic suprachiasmatic nucleus (SCN), produces earlier nocturnal sleep in elderly subjects. The circadian timing of various physiological functions, including autonomic, neuroendocrine, metabolic, immune, and behavioral systems, is regulated by the SCN via neuronal and hormonal BCP signals. The timing and duration of human sleep are also believed to be regulated in part by the SCN, being in balance with homeostatic influences (3), such that physiological functions that promote or inhibit human sleep are kept in adequate phases relative to sleep timing. Among these various physiological functions, core body temperature (BT) and serum melatonin (MLT) rhythms are often used as indirect markers of BCP. It is well established that the occurrence and continuity of human sleep are intimately related to the circadian phase of MLT (4, 5) and core BT (4, 6, 7, 8) rhythms; sleep propensity increases on the rising limb and decreases on the falling limb of the MLT rhythm phase, and a reversed relationship is also observed between sleep propensity and core BT curve. Numerous studies have reported that earlier clock time for nocturnal sleep in elderly persons is closely related to the phase advances of MLT (9, 10, 11) rhythms as well as the core BT (12, 13, 14, 15, 16, 17).
Given that the advanced shift of BCP that occurs with aging induces early bedtime and wake time in the elderly, it is still difficult to explain the vulnerability of the elderly to sleep maintenance disturbances, such as decreased sleep efficiency and increased nocturnal awakening, especially in the second half of the nocturnal sleep episode. It has been hypothesized that aging induces not only the clock time advance of BCP and associated sleep timing, but that it also alters the relationship between them, such that sleep is delayed relative to the BCP. Theoretically, such delayed sleep timing loads the latter part of sleep time in the elderly on the falling limb of the MLT curve or the rising limb of the core BT curve, during which times sleep propensity generally decreases. Previous studies evaluating sleep timing relative to the BCP, however, have failed to report consistent findings; shift advances, delays, and even marked variations have been reported against MLT (9, 11, 18, 19, 20, 21) and core BT (13, 14, 15, 16, 17, 22, 23) rhythms. An important possible explanation for these inconsistencies is the masking effects on MLT rhythm by the environmental light conditions (24, 25) as well as on core BT rhythm parameters by sleep (15, 17), which could lead to BCP miscalculations. The aim of this study was to estimate the relationship between the BCP and sleep timing as well as sleep maintenance ability in elderly subjects using the MLT secretion rhythm under dim light conditions as a reliable marker of BCP.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Subjects in the present study were 42 Japanese healthy men and women over 60 yr of age (mean age, 68.8 yr; male/female ratio, 16/26) and 27 Japanese healthy males under 30 yr of age (mean age, 22.5 yr), who gave written informed consent. All study subjects underwent rigorous physical and psychological evaluations by 3 physicians. Subjects were screened for the following: no irregularity in sleep-wake pattern (determined on the basis of a self-reported sleep diary) for 2 wk, no history of physical disease (including arthritis, asthma, and chronic obstructive pulmonary disease) that could affect sleep states, no history of psychiatric disease [determined on the basis of the Mini-International Neuropsychiatric Interview (26), a structured diagnostic psychiatric interview that screens for DSM-IV disorders (27) including depression and dementia], and no use of medications such as ß-blockers, antidepressants, estrogens, or glucocorticoids that have been reported to modify sleep states or MLT secretion levels during the prior 3 months. Twenty-eight of 42 elderly subjects and 2 of 27 young subjects took medication, including antihypertensives, antihyperlipidemics, iron supplement, etc. The subjects had experienced no jet lag during the past 6 months. Menopause was present in all elderly female subjects. Although we had no information about onset age of menopause for 2 of 26 female subjects, for the remaining 24 female subjects, the average onset age (±SD) of menopause was 51.0 ± 3.7 yr (range, 45–59), and the average postmenopausal period before this study was 17.7 ± 4.3 yr (range, 10–23), respectively. All patients had normal hematology and urinalysis data. All study subjects underwent overnight polysomnography to rule out moderate to severe sleep apnea or periodic limb movement according to the International Classification of Sleep Disorders criteria (28).
For 14 d before the start of the study, excessive exercise and alcohol were forbidden. During this prestudy period, sleep quality and regularity were assessed with an actigraph (AMI, Inc., Ardsley, NY) fitted to the nondominant wrist of each subject. We collected actigraph data and estimated sleep-wake state using Cole’s algorithm with optimal parameters (29). The average sleep onset time in the prestudy period was determined by actigraph data for each subject and was referred to as her/his 00 h. After the prestudy period, each subject was asked to maintain usual sleep habits (<10 lux) from 00 h for 7 d, and the sleep state was monitored with an actigraph and sleep diary (sleep-monitoring period). On the day after the sleep-monitoring period ended, each subject entered the sleep laboratory at 14 h before 00 h (-14 h). Before -12 h, the subject donned a cotton gown provided for the study, and an indwelling catheter (heparin lock) for painless blood collection was inserted into a peripheral vein in the left forearm. At this point and for 26 h until the end of the study at 14 h on the next day, subjects were required to recline on a seat during the daytime. They slept on a bed from 00–8 h, and sleep was forbidden outside this period. Some physical movement, walking to the toilet in the next room, and occasional stretching of the limbs, was allowed. At -6 to 5 h, 8–9 h, and 12–13 h, a 500- to 750-Cal meal and as much water as desired were given to each subject. Laboratory illumination was maintained at less than 10 lux during sleep and at 100 lux near the subject’s eyes at other times. The ambient temperature in the sleep laboratory was maintained at 23 ± 1 C throughout the study period.
Evaluation of sleep properties
Features of nocturnal sleep were evaluated on the basis of the sleep diary and actigraph data obtained during the sleep-monitoring period. Nighttime sleep parameters were determined per subject as follows: bedtime (clock time for going to bed), sleep onset time (clock time for the onset of sleep), wake time (clock time for end of major night sleep after which subject did not fall asleep again), total sleep time (the total time asleep), midpoint of sleep (the midpoint between sleep onset and wake time), sleep efficiency (SE; the total time asleep as a percentage of the total time in bed), and wake time after sleep onset (WASO: the accumulated time awake after the onset of sleep). SE and WASO were calculated separately for the first and second halves of sleep time divided at the midpoint of sleep for each subject.
Evaluation of MLT secretion rhythm
Blood sampling of serum MLT was collected every hour from -10 h for 24 h (25 points) via an iv catheter under dim light (<100 lux while awake and <10 lux during sleep) to reduce the masking effect on MLT levels as much as possible (25). Blood samples were immediately centrifuged (3000 rpm for 15 min), and serum was separated and frozen (below -80 C) before RIA. Parameters for MLT rhythm were determined per subject as follows: total amount of MLT secretion (AUC; the area under the MLT time-concentration curve), peak MLT concentration (maximum value of 24-h MLT secretion), amount of MLT secretion for the second half of sleep time (AUChalf; the AUC for the second half of sleep time), dim light MLT onset time [DLMOn; the evening time at which serum MLT concentrations reached 8.1 pg/ml (30), which was 3-fold the detection limit of the RIA kit used in this study], dim light MLT offset time [DLMOff; the morning time at which serum MLT concentrations decreased <8.1 pg/ml (30)], and midpoint (mBCP; the midpoint between DLMOn and off time as a marker of BCP). As blood sampling was performed every hour relative to average sleep onset time (00 h) for each subject, the individual DLMOn, DLMOff, or mBCP was converted to clock time for further analysis of relationship between sleep timing and mBCP.
Statistics
We used paired and unpaired t tests and one-way ANOVA, followed by Bonferroni’s post hoc analysis, to identify significant intra- and intergroup differences for sleep and MLT parameters. The F test was applied to determine whether a significant intergroup difference existed in variability of the distribution of the time interval between mBCP and wake time. Results are shown as the mean and SEM unless otherwise stated. P < 0.05 was considered significant.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Properties of sleep timing in the present subjects are shown in Table 1
. There was no statistically significant gender difference in either sleep onset or wake time in the elderly subjects, although the female elderly subjects showed a tendency to clock time advance in the midpoint of sleep (P = 0.062) compared with that in the male elderly subjects. The clock bedtime (mean difference, 84 min), clock sleep onset time (76 min), clock midpoint of sleep (84 min), and clock wake time (93 min) in the whole elderly group were significantly advanced compared with the corresponding values in the young group.
|
Raw data plots of daily serum MLT secretion in the young and elderly subjects are illustrated in Fig. 1, A and B. All subjects in both age groups showed obvious diurnal variation of serum MLT level; it peaked during the nocturnal sleep time and decreased below the detection limit (2.7 pg/ml) during the daytime. This made it possible to determine the authentic phase of DLMOn and DLMOff in each subject. There was no statistically significant gender difference in the DLMOn, DLMOff, or mBCP in the elderly subjects (Table 1
). The clock DLMOn (mean difference, 78 min), clock DLMOff (53 min), and clock mBCP (66 min) in the entire elderly group were significantly advanced compared with the corresponding values in the young group.
|
Properties of the relationship between mBCP and sleep timing in the present subjects are shown in Table 1 and illustrated in Fig. 1, C and D. There was no statistically significant gender difference in relationship between the mBCP and sleep onset, midpoint of sleep, or wake time in the elderly subjects. The interval between the mBCP and either sleep onset or midpoint of sleep did not differ significantly between the young and entire elderly groups. The interval between wake time and mBCP tended to decrease in the entire elderly group compared with the corresponding value in the young group; however, it did not reach a statistically significant level (P = 0.061). Figure 2 shows the distribution of the interval between mBCP and wake time in the two age groups. There was no significant difference in the variability of distribution of the interval between the two groups (F = 1.08; P = 0.83).
|
Properties of sleep quality in the subjects are shown in Table 1. There was no statistically significant gender difference in sleep quality parameters in the elderly subjects, although the female elderly subjects showed a tendency to decreased sleep efficiency for the second half of the sleep time (P = 0.086) compared with the male elderly subjects. The entire elderly group showed a significantly lower SE value as well as a significantly higher WASO value for the first and second halves of the sleep time compared with the corresponding values in the young group. We classified the elderly subjects into three subgroups, defined as possessing a longer (n = 14), intermediate (n = 16), or shorter (n = 12) interval between mBCP and wake time, with the average ± 0.5 SD as cut-off points. The three subgroups of elderly subjects showed significant decreases in SE (F = 21.17; df = 3; P < 0.001; Fig. 3A) and increases in WASO (F = 23.22; df = 3; P < 0.001; Fig. 3b) for the second half of the sleep time compared with that in the young subjects depending on the degree of reduction in the interval between mBCP and wake time.
|
Properties of MLT secretion volume in the subjects are shown in Table 1. There was no statistically significant gender difference in the MLT secretion parameters in the elderly subjects. Both AUC and peak MLT concentrations in the entire elderly group were significantly reduced compared with the corresponding values in the young group. However, AUChalf did not differ between the groups and subgroups (F = 1.96; df = 3; P = 0.129; Fig. 3C). The AUC (r = 0.21; P = 0.19) and the AUChalf (r = 0.22; P = 0.18) showed no significant correlation to SE for the second half of sleep time in the elderly subjects. Similarly, the AUC (r = 0.21; P = 0.17) and AUChalf (r = 0.28; P = 0.17) showed no significant correlation to WASO for the second half of sleep time in the elderly subjects.
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Our elderly subjects showed marked clock time advances in mBCP and associated sleep timing compared with the young subjects, consistent with the findings of numerous previous studies (9, 10, 11, 12, 13, 14, 15, 16, 17). However, we found no significant age-related change in the interval between mBCP and either sleep onset (mean difference, 11 min between the elderly and young groups) or midpoint of sleep (mean difference, 19 min) in our subjects. Duffy et al. (9) demonstrated that aging is associated with earlier sleep timing in both bedtime and wake time relative to MLT rhythm phase. The elderly subjects in our study also showed a decreasing tendency only in the interval between the wake time and mBCP (mean difference, 32 min; P = 0.061); however, it did not reach a statistically significant level. These findings seem to suggest that the sleep timing in the elderly is substantially stable against the BCP even compared with the young healthy controls. In this study we used 8.1 pg/ml as a cut-off point for determination of mBCP according to the findings of a previous study (30). Although we also applied the 24-h mean value of MLT secretion used in the study by Duffy et al. (9) as well as the value of 10 pg/ml advanced by Lewy et al. (31), we observed no significant difference in the interval between mBCP and sleep timing in either of these analyses. In some of our study subjects the peak MLT secretion curves showed irregular patterns: for instance, multipeaks or flattened shapes, which are often observed especially in elderly subjects. In these subjects the time of the MLT peak seemed inadequate as a phase marker considering their entire secretion profiles or the results using other analyses, mentioned above. The finding that the distribution of the interval between mBCP and wake time was nearly the same between the elderly and young subjects also suggests the stability of circadian regulation of BCP and associated sleep timing throughout the aging process.
It has been reported that elderly subjects with sleep maintenance deterioration show delayed sleep timing relative to BCP (11, 13, 14, 22), which theoretically loads the latter part of the sleep period onto the circadian phase during which sleep propensity generally decreases. This hypothesis is obliquely supported by the fact that timed exposure of bright light (32, 33) or passive body heating (23), which induces a phase delay of core BT rhythm and realigns the time relationship between sleep period and core BT rhythm closer to that of a younger person, could alleviate sleep maintenance disturbances in the elderly. However, many of these previous studies did not successfully exclude the masking effect induced by sleep itself on core BT rhythm used in these studies as a BCP marker. We speculate that such incomplete demasking treatment could result in the previous inconsistent findings.
In the present study we found a significantly decreased ability for sleep maintenance, such as lower SE and increased WASO in the elderly, which was dominant in the second half of the sleep time, compared with those in young controls. The sleep properties in the present subjects were consistent with the knowledge that aging is commonly associated with reduced sleep quality, especially a decrease in sleep maintenance (1). Although the influence of gender difference on sleep parameters has been reported (34), the present elderly subjects showed no statistical significant gender differences in sleep parameters, except that female elderly subjects showed a tendency for decreased sleep efficiency in the second half of the sleep time. Interestingly, decreased SE or increased WASO for the second half of the sleep time in the elderly subjects became prominently dependent on the degree of reduction in the interval between mBCP and wake time. This suggests that elderly subjects with an earlier wake time possess not only a simple advanced timing of sleep termination relative to the BCP, but also a decreased ability for sleep maintenance. This finding seems contrary to the well established data concerning the relationship between BCP and sleep propensity in young people (4, 5, 6, 7, 8). Namely, a decreased ability for sleep maintenance in the present elderly subjects appeared in a BCP for which sleep propensity remained sufficient to sustain sleep. One possible explanation is that diminished nighttime MLT secretion, observed in our elderly group, may have caused the deterioration in sleep maintenance. MLT has been reported to have some effects on human sleep, including acute hypothermic (35), hypnogenic (36), and phase-shifting effects (37) in healthy young controls when administered at appropriate circadian phase. Numerous previous studies have also indicated age-related decreases in MLT secretion during the sleep period (11, 19, 38, 39), especially in elderly insomniacs (30, 40). According to these data, some investigators have tried supplementary administration of exogenous MLT in elderly subjects with sleep maintenance disturbances and reported its favorable effects (41, 42). However, these data do not necessarily certify the relation of cause and effect between the diminished MLT secretion and deteriorated sleep maintenance in the elderly. Actually, we found no significant relation between the AUChalf and either SE or WASO in the elderly group, suggesting that at least a diminished level of absolute amount of MLT secretion in the elderly did not simply account for their decreased ability to maintain sleep.
We believe that future studies should focus on the following three unresolved issues to investigate the mechanism responsible for age-related deterioration in sleep maintenance. First, the relationship between sleep timing and other physiological functions that could promote or inhibit human sleep, including HPA axis or glucose metabolism, should be determined. Recent studies of clock gene regulation revealed that each peripheral organ has its own circadian oscillation with a different BCP (43, 44). Peripheral organs in an individual could exhibit different mutual phase relationships from those under physiological conditions when exposed to particular environmental light or feeding due to, for instance, jet lag or restricted meal time (44, 45). Although we adopted the MLT phase as a BCP marker in this study, the present findings do not necessarily mean that aging also induces no significant changes in relationship between sleep timing and other physiological functions that might possess sleep-regulating effects. Second, we should consider the involvement of homeostatic process in aging sleep. It has been successfully hypothesized that human sleep propensity is regulated by two independent processes, circadian process and sleep-dependent homeostatic process (3). It has been revealed that the homeostatic drive of sleep propensity increases in dependence on the degree of sleep loss or prior wakefulness in human. Although the neurobiological entity of the homeostatic process remains unclear, the age-related decrease in homeostatic drive during the latter part of the sleep time might be a primary mechanism of sleep maintenance disturbances in the elderly. Third, the present study has not focused on age-related changes in sensitivity to BCP signals, including MLT, bright light, thermoregulation, glucocorticoids, etc. Some previous studies have reported an age-related decrease in responsiveness to synchronizing signals induced by light (46, 47, 48, 49), activity (50), and exogenously administered MLT (51). A recent study has also shown increased sensitivity to the arousal-producing effect of the CRH in elderly subjects (52). These findings let us assume that such a change in sensitivity to the various BCP signals might be a factor that accelerates the decreased homeostatic drive of sleep propensity with advancing age. Given that the altered sensitivity to other BCP signals commonly occurs with aging, deterioration in sleep maintenance could be induced without any shift in nocturnal sleep timing in the elderly.
Conclusion
In this study we have shown that the decrease in sleep maintenance that occurs with age could appear without a significant phase delay of sleep timing against the circadian signal of pineal hormone MLT, which is a stable circadian phase marker. The present study raises important questions concerning the etiology of age-related changes in human sleep quality and the physiological significance of age-related changes in circadian regulation represented by marked phase advances.
|
Footnotes |
|---|
Abbreviations: AUC, Area under the curve; AUChalf, area under the curve for the second half of sleep time; BCP, biological clock phase; BT, body temperature; DLMOff, dim light melatonin offset time; DLMOn, dim light melatonin onset time; mBCP, midpoint of biological clock phase; MLT, melatonin; SCN, suprachiasmatic nucleus; SE, sleep efficiency; WASO, wake time after sleep onset.
Received January 29, 2003.
Accepted June 16, 2003.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
and •, Average MLT secretion profiles in young and elderly subjects, respectively; 
and 
, average sleep time in young and elderly subjects, respectively. The horizontal axis in D indicates the time relative to BCP. 
and 
(DLMOff), Corresponding values in young and elderly subjects, respectively. Columns with a center line indicate sleep time and the midpoint of sleep in young and elderly subjects. Values are shown as the mean ± SEM.
