| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Special Feature |
Section on Biological Rhythms, Mood and Anxiety Disorders Program (T.T.P., M.A.J., T.A.W.), National Institute of Mental Health, Bethesda, Maryland 20892; Division of Sleep Medicine, Department of Medicine (D.A.), Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts 02115; Department of Psychiatry (L.S.), Columbia University, New York, New York 10032; and Department of Psychiatry and Human Behavior (J.R.M.), University of Texas Medical Branch, Galveston, Texas 77555-0189
Address all correspondence and requests for reprints to: Daniel Aeschbach, Ph.D., Division of Sleep Medicine, Department of Medicine, Brigham and Women’s Hospital, Harvard Medical School, 221 Longwood Avenue, Boston, Massachusetts 02115. E-mail: daeschbach{at}hms.harvard.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The duration and timing of sleep are regulated by an interaction between the circadian pacemaker in the suprachiasmatic nucleus (SCN) of the hypothalamus, which programs daily cycles in sleep propensity, and the sleep homeostat, which tracks the increase of sleep pressure (i.e. sleep debt) during waking and its decrease during sleep (4, 5, 6, 7). Conceivably, either one of these regulatory systems might differ among individuals. Recently, quantitative analyses of the electroencephalogram during sleep and wakefulness indicated that although natural short sleepers (habitual sleep duration <6 h) manage to live under higher sleep pressure, they do not differ from long sleepers (>9 h) with regard to the kinetics of the sleep-wake dependent variations in sleep pressure (8, 9).
Given that the sleep homeostat appears to operate normally in long sleepers and short sleepers, we hypothesized that the temporal program of the circadian pacemaker may be different in the two groups. The circadian pacemaker programs the daily cyclic changes in the body’s internal milieu, thereby creating a biological day and biological night (10, 11, 12). The two intervals are delimited by relatively abrupt changes in neuroendocrine function, body temperature, and arousal (10) that appear to result from abrupt changes in the firing rate of cells in the SCN (13, 14). During the biological night, melatonin is secreted, cortisol concentration in the blood increases, core body temperature is lowered, and sleepiness and sleep propensity increase (6, 10, 11, 12). We hypothesized that the duration of the biological night that is programmed by the circadian pacemaker is longer in long sleepers than short sleepers. To allow for an unmasked measurement of the biological night, we subjected young, healthy long sleepers and short sleepers to a constant routine (CR) protocol (15) and measured the duration of the nocturnal interval of the circadian rhythms in plasma melatonin, plasma cortisol, core body temperature, and subjective sleepiness.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Ten long sleepers (20–30 yr; five women and five men) who habitually slept more than 9 h, and 14 short sleepers (21–34 yr; eight women and six men) who habitually slept less than 6 h were studied. Subjects were recruited over a period of 3.5 yr through newspaper and radio advertisements in the Washington, D.C., metropolitan area. Two- to four-week sleep logs and wrist motor activity recordings were used to identify individuals who met the criteria. Special attention was paid to the weekends: Individuals who slept less than 6 h on average but more than 7 h on weekends were excluded because we suspected that their weekdays’ sleep duration deviated from their natural sleep need. Subjects reported that their habitual sleep duration had not changed within 1 yr before the study. They reported no sleep problems or shift work and no use of medications, drugs, or tobacco. The reported daily consumption of caffeinated beverages (coffee, tea, cola) did not differ (t test) between long sleepers (mean ± SEM, 1.1 ± 0.5) and short sleepers (1.3 ± 0.4). Subjects were in good health as determined by medical history, structured clinical interview, physical exam, blood and urine tests, electrocardiography, and all-night polysomnography. Subjects were free of any current or past Axis I psychiatric disorder. There were no significant differences (t tests) between groups in plasma glucose levels (long sleepers vs. short sleepers: 93.1 ± 1.4 vs. 91.6 ± 4.4 mg/dl), total cholesterol (158.4 ± 5.5 vs. 150.1 ± 10.0 mg/dl), systolic (114.2 ± 3.8 vs. 116.6 ± 4.2 mm Hg), and diastolic blood pressure (63.2 ± 3.44 vs. 66.8 ± 2.6 mm Hg).
All-night polysomnography showed that subjects had fewer than two 10-sec apneas per hour of sleep (long sleepers vs. short sleepers: 0.37 ± 0.15 vs. 0.26 ± 0.13, n.s., t test) and fewer than 11 periodic leg movements per hour of sleep (1.45 ± 0.42 vs. 2.22 ± 0.91, n.s.). Women reported no current use of oral contraceptives. They were studied during the follicular phase of the menstrual cycle as determined by ovulation kits and logs. For 1 wk before the study, subjects were asked to refrain from alcohol and caffeinated products and maintain their habitual bed times and wake-up times. Compliance to the latter instruction was checked with wrist-worn light/activity monitors (Actillume, Ambulatory Monitoring, Ardsley, NY). The average duration of the solar day during the month before the study did not differ (t test) between long sleepers (12.62 ± 0.64 h) and short sleepers (12.38 ± 0.36 h).
Subjects gave written informed consent for their participation in the study. The protocol was approved by the Institutional Review Board of the National Institute of Mental Health.
Protocol
Subjects were admitted to a research ward for 2 nights of sleep, an approximately 40-h CR protocol, and a period for recovery sleep. The first 2 nights were scheduled to match each individual’s average weekday bed time and wake-up time that were derived from the sleep logs and activity recordings. On average, time in bed was from 2312 h ± 14 min (SEM) to 0859 h ± 19 min in the long sleepers and from 0100 h ± 16 min to 0623 h ± 13 min in the short sleepers. Total sleep time as determined by polysomnography (16) in the second night’s sleep was 8.62 ± 0.25 h in the long sleepers and 4.98 ± 0.19 h in the short sleepers. The CR protocol began after the second night’s sleep and ended at 2300 h on the following day. During this time, subjects stayed awake in bed in a propped-up position in dim light (<10 lux) without having access to clocks. Fluids and isocaloric meals were given at room temperature every hour and every 2 h, respectively. A staff member attended the subjects continuously to prevent them from falling asleep and ensure adherence to the protocol. The purpose of the CR protocol was to minimize or distribute evenly over time the masking influences of sleep, posture, light, and meals on the circadian rhythms.
Data collection
Blood samples were collected through an indwelling iv catheter at half-hourly intervals for 24 h (1400–1400 h) during the CR. Plasma melatonin concentrations were measured with a RIA that had a detection limit of 2.8 pg/ml (Stockgrand, Surrey, UK). The intraassay coefficients of variation were 7.2% (n = 10), 5.6% (n = 10), 6.2% (n = 10), and 4.5% (n = 10) at means of 22, 35, 50, and 203 pg/ml, respectively, and the interassay coefficient of variation were 11.8% (n = 10), 7.6% (n = 10), 8.5% (n = 10), and 3.8% (n = 10) at means of 22, 34, 50, and 205 pg/ml, respectively. Plasma cortisol concentrations were measured with a RIA that had a detection limit of 0.5 µg/dl (Covance Laboratories, Inc., Vienna, VA). The intraassay coefficients of variation were 8.4% (n = 10), 5.3% (n = 10), and 11.1% (n = 10) at means of 5.01, 19.21, and 33.05 µg/dl, respectively, and the interassay coefficients of variation were 9.1% (n = 10), 8.9% (n = 10), and 11.5% (n = 10) at means of 4.83, 19.18, and 35.69 µg/dl, respectively.
Core body temperature was recorded at 1-min intervals with an indwelling rectal probe that was connected to an electronic thermometer (Iso-thermex, Columbus Instruments, Columbus, OH). After exclusion of visually identified artifacts, the data were averaged to yield half-hourly values.
Subjective sleepiness was assessed half-hourly with bipolar 100-mm visual analog scales.
Data analysis
The nocturnal interval of high plasma melatonin levels was defined in each individual as the interval between the onset of secretion of melatonin and its disappearance from the plasma. Onset was defined as the half-way point between the last undetectable and first detectable plasma concentration. Disappearance was defined as the half-way point between the last detectable and first undetectable concentration.
The nocturnal interval of increasing plasma cortisol levels was defined as the interval between the minimum and maximum that was identified after subjecting the cortisol time series to a two-time 5-point moving average.
The nocturnal interval of low body temperature was defined as the interval between the two midrange crossing points of the smoothed (5-point moving average) temperature data. The midrange was defined as the half-way point between the maximum during the first 24 h of the CR and the nocturnal minimum.
The nocturnal interval of increasing sleepiness was determined as follows. In each subject, a function that represents the sum of a linear component and two cosine components with periods of 24 h and 12 h (17) was fitted to the smoothed (two-time 5-point moving average) half-hourly sleepiness rating (SAS Institute, Inc., Cary, NC). Next, the linear component that was introduced to approximate the wake-dependent increase in sleepiness was subtracted from the smoothed data. Two points were determined in the resulting time series: the maximum, and last local minimum before the maximum that remained within or below the range of all local daytime (i.e. before 2400 h) minima. The nocturnal interval of increasing sleepiness was then defined as the interval between the last local minimum and the maximum.
Statistics
Based on our knowledge that long sleepers go to sleep earlier and arise later than short sleepers, we hypothesized that in the former the circadian rhythms’ nocturnal intervals begin earlier and end later. Therefore, one-tailed unpaired t tests were used to compare between groups onset time, offset time, and duration of these intervals. If the variances within the two groups were unequal, an approximate t statistic that was based on the Satterthwaite’s approximation for the degrees of freedom was used. The phase relationship between habitual bed time and onset of the circadian rhythms’ nocturnal intervals and between habitual wake-up time and offset of the circadian rhythms’ nocturnal intervals were analyzed with two-tailed paired and unpaired t tests.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
, top) was longer (P = 0.024) in long sleepers than in short sleepers. The amplitude of melatonin secretion as measured by the maximum of the smoothed (5-point moving average) time series did not differ (P > 0.15, two-tailed t test) between long sleepers (63.3 ± 8.0 pg/ml) and short sleepers (50.5 ± 4.6 pg/ml). The duration of the nocturnal interval of high melatonin levels did not correlate with the amplitude of melatonin secretion (long sleepers: r = 0.27, P > 0.4, n = 9; short sleepers: r = 0.02, P > 0.9, n = 13; pooled: r = 0.20, P > 0.3, n = 22). The difference between groups in the duration of the nocturnal interval of high plasma melatonin levels remained statistically significant (P = 0.033) after correction for differences in amplitude that was based on the linear regression of the two variables.
|
, bottom) was longer (P = 0.028) in long sleepers than in short sleepers. The amplitude of the circadian rhythm of body temperature (i.e. half of the difference between the maximum and minimum of a fitted function that constitutes the sum of two cosine components with periods 24 h and 12 h) did not differ (P > 0.7, two-tailed t test) between long sleepers (0.29 ± 0.01 C) and short sleepers (0.28 ± 0.03 C).
The nocturnal interval of increasing plasma cortisol levels (Fig. 2) was longer (P = 0.034) in long sleepers than in short sleepers. In neither group did the timing of the plasma cortisol’s maximum differ from habitual wake-up time (P > 0.4).
|
) was longer (P = 0.021) in long sleepers than in short sleepers. In neither group did the timing of the maximum of sleepiness differ from habitual wake-up time (P > 0.4).
|
. The longer duration of the circadian rhythms’ nocturnal intervals in long sleepers than short sleepers was due to a later offset time in the morning (melatonin P = 0.053; other variables P < 0.05) and not due to an earlier onset time in the evening (P > 0.8). The phase angle between the onset of the circadian rhythms’ nocturnal intervals and habitual bed time was larger in the short sleepers than in the long sleepers (sleepiness P = 0.11; other variables P < 0.05). The phase angle between wake-up time and offset of the circadian rhythms’ nocturnal intervals did not differ between the two groups (P > 0.4), except for melatonin, for which it was larger in the short sleepers (P = 0.038).
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The waveforms of the circadian rhythm in sleepiness suggest differences between long and short sleepers in the duration of a circadian signal that promotes sleep. Such a signal is thought to originate in the SCN and be transmitted through efferent projections and possibly by output molecules to hypothalamic nuclei implicated in arousal state control (18, 19). The neuroendocrine and temperature variables that we studied may not simply be markers of a circadian signal that promotes sleep but may also be mediators of such a signal. Presence of melatonin in the blood and low core body temperature have previously been associated with increased sleep propensity and sleep consolidation (20, 21, 22, 23). Thus, extended periods of high plasma melatonin levels and low body temperature may directly facilitate long sleep. Moreover, the coincidence of the maximum in plasma cortisol with habitual wake-up time in both groups is consistent with the notion that the hypothalamo-pituitary-adrenal axis plays an important role in and even anticipates the sleep-waking transition (24). In this regard, the importance of glucocorticoids in conveying circadian time information from the SCN to peripheral tissues was shown recently (25).
The present data provide a physiological basis for the interindividual variability and the intraindividual stability of habitual sleep duration. The persistence or inertia of a circadian pacemaker’s individual program as was revealed in constant conditions provides a mechanism that stabilizes individual sleep duration. This mechanism is habit promoting and could explain the commonly experienced difficulty of changing habitual sleep duration willfully. From a functional point of view, a built-in mechanism that stabilizes an individual’s sleep duration may serve to prevent deviations from an individual’s sleep need, which appears to have pathophysiological consequences (2, 26).
We cannot from this study determine the origin of the individual differences in the circadian pacemaker’s program. Circadian rhythms arise from autoregulatory transcriptional and translational feedback cycles of clock genes in cells of the SCN (27). It is conceivable that polymorphisms in one or several of these genes result in differences in the duration of the circadian rhythms’ nocturnal interval and thus in the duration of sleep. Studies in both animals (28, 29, 30) and humans (31, 32) suggest a genetic basis for sleep duration. Specifically, decreased sleep duration was recently reported for the mouse Clock mutation (33). Thus, long and short sleepers may represent promising phenotypes for the study of the function of clock genes in human sleep. On the other hand, one needs to keep in mind that in both animals and humans, a circadian pacemaker’s temporal program is not completely rigid and can adjust to chronic changes in the duration of the photoperiod to which the organism has been exposed (10, 11, 34, 35). Although the duration of the solar day before the study did not differ between long sleepers and short sleepers, differences in the exposure to artificial light may have been effective (11). Thus, prolonged exposure to short photoperiods, as occurs when sleep is chronically long, or long photoperiods, as occurs when sleep is chronically short, may contribute to shaping the circadian pacemaker’s program, which in turn could perpetuate a long sleep duration in some individuals and a short sleep duration in others. This possibility may explain the observation that imposed chronic sleep restriction has a long-term effect on habitual sleep duration after the sleep restriction is lifted (36). Hence, although the origin of the differences between long and short sleepers is unknown, the circadian pacemaker appears to be important in mediating and perpetuating such differences.
All four markers of the biological night indicated that its longer duration in the long sleepers was due to a delayed offset in the morning and not an advanced onset in the evening. This asymmetry was not reflected in habitual bed time and wake-up time, which suggests that the two events are controlled by different factors. The observation that short sleepers go to sleep later relative to the onset of the circadian rhythms’ nocturnal interval indicates that their tolerance of the rising circadian pressure to sleep is higher than in long sleepers. This adds to a previous observation that short sleepers tolerate a higher homeostatic sleep pressure than long sleepers (8, 9). Taken together the data indicate that the differences between long and short sleepers in wake-up time are related to differences in the timing of the offset of a circadian signal that promotes sleep, whereas the differences in bed time are related to differences in the tolerance or sensitivity to the circadian and homeostatic drives for sleep. In addition, the decision to go to bed may be influenced by an individual’s use of artificial evening light, which has acute potency to reduce sleepiness (37).
There is an ongoing debate as to how much sleep is enough to avoid adverse effects on health. Recent studies (26) have indicated that subchronic curtailment of sleep has negative effects on carbohydrate metabolism, sympathovagal balance, and evening cortisol concentrations. Chronic sleep loss may therefore contribute to some age-related pathologies, such as diabetes, hypertension, and memory loss. The present study was not designed to test these hypotheses. It is worth noting, however, that the tests during the screening phase of this study did not reveal significant differences between groups in plasma glucose concentration (insulin was not measured), total cholesterol concentration, and blood pressure (see Subjects). Furthermore, during the study there was no difference (P > 0.7) in plasma cortisol concentrations in the evening (1600–2100 h) between long sleepers (4.70 ± 0.55 µg/dl) and short sleepers (5.07 ± 0.75 µg/dl). The observation that the circadian pacemaker appears to facilitate short sleep in some individuals and long sleep in others indicates that sleep need may not be uniform among individuals and therefore it may not be possible to define one healthy sleep duration for an entire population.
In conclusion, we postulate that differences in the circadian pacemaker’s program may contribute to the disparities in sleep duration among individuals. The present results are relevant not only for understanding differences in sleep duration among healthy people but also for indicating possible mechanisms underlying pathological changes in sleep duration, such as occur during mania (shortened sleep) and depression (extended sleep) in patients with manic-depressive illness (38). Finally, the observation of increased melatonin levels in the morning in long sleepers may serve as a potentially useful paradigm for the treatment with melatonin of early morning awakenings, a chief complaint in elderly people (23, 39).
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Abbreviations: CR, Constant routine; SCN, suprachiasmatic nucleus.
Received May 29, 2002.
Accepted September 16, 2002.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  O. G. Jenni, L. Molinari, J. A. Caflisch, and R. H. Largo Sleep Duration From Ages 1 to 10 Years: Variability and Stability in Comparison With Growth Pediatrics, October 1, 2007; 120(4): e769 - e776. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. P. Pinheiro, E. S. Schernhammer, S. S. Tworoger, and K. B. Michels A prospective study on habitual duration of sleep and incidence of breast cancer in a large cohort of women. Cancer Res., May 15, 2006; 66(10): 5521 - 5525. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. K. Verkasalo, K. Lillberg, R. G. Stevens, C. Hublin, M. Partinen, M. Koskenvuo, and J. Kaprio Sleep Duration and Breast Cancer: A Prospective Cohort Study Cancer Res., October 15, 2005; 65(20): 9595 - 9600. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D.-J. Dijk and M. von Schantz Timing and Consolidation of Human Sleep, Wakefulness, and Performance by a Symphony of Oscillators J Biol Rhythms, August 1, 2005; 20(4): 279 - 290. [Abstract] [PDF]
|
|
|
|
 |
|
 L. Sher Type D personality: the heart, stress, and cortisol QJM, May 1, 2005; 98(5): 323 - 329. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
O. G. Jenni and B. B. O'Connor Children's Sleep: An Interplay Between Culture and Biology Pediatrics, January 1, 2005; 115(1/S1): 204 - 216. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. M. W Rajaratnam, B. Middleton, B. M Stone, J. Arendt, and D.-J. Dijk Melatonin advances the circadian timing of EEG sleep and directly facilitates sleep without altering its duration in extended sleep opportunities in humans J. Physiol., November 15, 2004; 561(1): 339 - 351. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. Wilkoff, O. G. Jenni, I. Iglowstein, L. Molinari, and R. H. Largo Sleep Need in Children Pediatrics, December 1, 2003; 112(6): 1463 - 1464. [Full Text] [PDF]
|
|
|
|
 |
|
 S. A. Rivkees Time to Wake-Up to the Individual Variation in Sleep Needs J. Clin. Endocrinol. Metab., January 1, 2003; 88(1): 24 - 25. [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 |
0.05, t test).