0w~1~972x/92/7506-1431$03.00/0 Journal of Clinical Endocrinology and Metabolism Copyright 8 1992 by The Endocrine Society
Vol. 15, No. 6 Prmted in U.S.A.
Nocturnal Adrenocorticotropin and Cortisol Secretion Depends on Sleep Duration and Decreases in Association with Spontaneous Awakening in the Morning* ERNST HORST
SP.&TH-SCHWALBE, L. FEHM, AND JAN
THILO BORN
SCHijLLER,
Klinik fiir Innere Medizin, Universittit zu Liibeck, Universittit Bamberg, 8600 Bamberg; Germany
WERNER
2400 Liibeck; and Abteilung
KERN, Psychophysiologie,
ABSTRACT It is still discussed controversially to what extent the nocturnal activity of the hypothalamus-pituitary-adrenocortical system depends on sleep and awakening in the morning. Therefore, we investigated the association of plasma ACTH and cortisol levels with undisturbed nocturnal sleep and spontaneous awakening in 14 healthy male subjects (between 2300 h and 1100 h). Between sleep onset and 476.9 min after sleep onset mean plasma cortisol level was significantly (P < 0.01) higher (210 + 15 us. 155 f 9 nmol/L) in the group with a shorter (476.9 + 15.0 min; n = 7; mean + SEM) than in the group with a longer total sleep time (596.9 -t 14.4 min; n = 7). Spontaneous awakening in the morning was not linked to the presence of any specific sleep stage or
to rising plasma ACTH and cortisol levels. However, spontaneous awakening was followed by a brief rise in plasma ACTH and cortisol in both groups. Thereafter, during wakefulness plasma ACTH and cortisol abruptly declined in all subjects irrespective of the time of awakening. The slope of the plasma ACTH and cortisol curves differed significantly (ACTH: P < 0.001; cortisol: P < 0.002, for all subjects) comparing the time after awakening (until 1100 h) with a time interval of identical length before awakening. We conclude that the duration of sleep and nocturnal ACTH and cortisol secretion are interrelated. Furthermore, the data suggest that the endogenous early morning activation of the hypothalamus-pituitary-adrenocortical system is terminated by mechanisms closely associated with awakening. (J Clin Endocrinol Metab 75: 1431-1435,1992)
E
night (the fourth or fifth) tended to occur during increasing cortisol concentrations, i.e. during increased HPA activity (12). Thus, given that cortisol has been found to decrease REM sleep (9) it was hypothesized that increased HPA activation during the very last REM sleep epochs represents a trigger to wake up spontaneously. We expected that the probability of spontaneous awakenings in the morning is increased when REM sleep occurs in temporal association with increasing plasma ACTH and cortisol levels. On the basisof previous studies (3, 11) we also expected that wakefulness then may initiate mechanisms inhibiting further ACTH and cortisol secretion. Consequently, we expected that the duration of the early morning releaseof ACTH and cortisol depends on the time of the subject’s spontaneous awakening, and thus also on the subject’s sleeping time. To test these hypotheses, in the present study we investigated plasma ACTH and cortisol levels during undisturbed sleep and spontaneous awakening in the morning. Special attention was paid to a comparison of ACTH and cortisol profiles in subjects with long zts.short sleeping time.
VIDENCE has accumulated for relationships between central nervous sleep processesand the secretory activity of the hypothalamus-pituitary-adrenocortical (HPA) system (l-3). Close associations have been found between ultradian rhythms of sleep and episodic ACTH and cortisol secretion (4, 5). The sleep-wake cycle appears to play also an important role in the regulation of the circadian adrenocortical secretory patterns (6-8). Several studies suggestedthat these connections are bidirectional. In a recent study administration of ACTH or cortisol substantially reduced rapid eye movement (REM) sleep, and simultaneously increased time spent awake (9). Furthermore, during undisturbed sleeping conditions increasing endogenous plasma cortisol levels were found to be significantly associatedwith light sleep or wakefulness (4, 10). On the other hand, sleep manipulations which increased the time awake during the night have been found to temporarily reduce ACTH and cortisol secretion (11). The last hours of nocturnal sleep time are characterized by an activation of the hypophyseal-adrenocortical axis. In a previous study, REM sleep appeared to be associatedwith reduced HPA activity (4, 10). However, since the period length of nocturnal HPA activity exceeded that of the nonREM-REM cycles, the very last REM sleep epochs of the Received January 22, 1992. Address all correspondence and requests for reprints to: Dr. E. SpithSchwalbe at this present address: Universitatsklinik Ulm, Medizinische Klinik, Innere III, Robert-Koch-Strasse 8, D-7900 Ulm Germany. * Supported by a grant from Deutsche Forschungsgemeinschaft (to J. B.).
Subjects
and Methods
Fourteen healthy men, aged 20-34 yr (mean 24.9), were studied. The subjects did not suffer from sleep disturbances and were not taking any medication. It was ascertained that they had normal sleep-wake rhythms for at least 2 weeks before the experiments were carried out. The subjects were acclimatized to the experimental setting by an adaptation night under the conditions of the experiment. They were requested to get up between 0700 h and 0800 h before the experimental nights and not to take any naps throughout the day. The experimental nights were carried out in an air-conditioned elec-
1431
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 16 February 2015. at 03:56 For personal use only. No other uses without permission. . All rights reserved.
SPATH-SCHWALBE trically shielded room. Upon arrival at the laboratory at 2100 h the subjects were prepared for polysomnographic recordings and blood sampling. They were informed that they could sleep as long as they wanted, and that they should stay in bed after awakening until 1100 h. They were allowed to turn on the lights after awakening in the morning. Continuous recordings were obtained from 2300 h, the time when lights were turned off, until 1100 h the next morning. Blood for determination of plasma cortisol and ACTH was collected every 15 min from an adjacent room via an iv forearm catheter connected to a long thin tube. To prevent clotting about 500 ml saline (0.9%) were infused throughout the study period.
Recording
and data analysis
Sleep stages were determined from the electroencephalographic, electrooculographic, and electromyographic recordings, which were scored off line according to the criteria described by Rechtschaffen and Kales (13). Blood samples were centrifuged immediately after collection, and plasma was stored at -20 C until assay. Plasma ACTH was measured by immunoradiometric assay (Euro-Diagnostics BV, Apeldoorn, The Netherlands), using two specific antibodies directed against ACTH (216) and ACTH (34-39), respectively (14). Cross-reaction of the corticotropin fragments l-24 and CLIP in the two-site immunoradiometric assay were not observed. The assay detects only intact ACTH. Plasma cortisol was measured by RIA (Hermann Biermann GmbH, Bad Nauheim, Germany). The sensitivities of ACTH and cortisol assays were 1.7 pmol/L and 5.5 nmol/L, respectively. The intraassay precision ranged from 10% at 3.3 pmol/L to 4% at 11 pmol/L for ACTH, and 5% at 27.6 nmol/L to 3% at 415 nmol/L for cortisol. The intraassay coefficients of variation were below 10%. All samples from a subject were analyzed in duplicate in the same assay. For each night, the total sleep time as well as the percentage of time spent in the different sleep stages (wakefulness, W; sleep stage l-4, Sl54; and REM) were calculated. The total sleep time lasted from sleep onset (SO), defined as the onset of the first Sl epoch followed by S2 sleep until the definitive awakening in the morning, when no sleep (defined as any sleep stage Sl-S4 or REM) followed. SO latencies were computed with reference to 2300 h. The total subjects sample was divided into two groups according to their sleep time. Subjects of the short sleep group (SSG) displayed sleep times less than 540 min whereas the subject’s sleep time of the long sleep group (LSG) exceeded 540 min. Both groups consisted of seven subjects. For comparisons between SSG and LSG the individual data were referenced to SO. This presentation served to illustrate the differences in mean hormone levels between the SSG and LSG after SO. The significance of spontaneous awakening in the morning on subsequent hormone secretion was evaluated by comparing the mean levels during a 2-h period preceding and after spontaneous awakening. To assess nocturnal ACTH and cortisol release, the mean plasma concentrations were calculated from the concentrations measured every 15 min. Then, integrated ACTH and cortisol release was determined by calculating the area under the hormone curves. To assess the effect of awakening on plasma hormone levels, for each subject the average slope of the individual hormone curves was calculated for a time interval between awakening and 1100 h, and a time interval of equal duration immediately before awakening. Statistical evaluation was performed with nonparametric tests (Kruskal-Wallis with subsequent Mann-Whitney’s U test and Wilcoxon’s t test).
ET AL.
JCE & M. 1992 .vol75.No6 .
in REM sleep tended to be longer in the LSG without reaching statistical significance (130.1 + 9.4 US. 106.3 + 16.8 min; Fig. 1). No significant differences in the percentage of time spent in the different sleep stages were observed between both groups (Table 1). The mean of the ACTH (Fig. 2) levels between SO and 476.9 min (average sleep time of the SSG) tended to be higher in the LSG compared with the SSG (2.8 + 0.4 VS. 1.9 + 0.4 pmol/L), but this difference remained nonsignificant. Plasma cortisol appeared to rise earlier in the SSG than the LSG. The mean of plasma cortisol (Fig. 3) levels between SO and 476.9 min was significantly higher in the SSG compared with the LSG (210 + 15 US. 155 + 9 nmol/L; P < 0.01). During this time interval integrated cortisol release also was significantly higher (P < 0.01) in the SSG. In contrast to ACTH, all mean cortisol levels of the samples taken between SO and 5 12 min were higher in the SSG than the corresponding cortisol levels in the LSG. However, no significant differences were found, when mean ACTH and cortisol levels were calculated for the total sleep time (LSG IIS. SSG; 3.0 + 0.3 US. 1.9 f 0.4 pmol/L for ACTH, and 201 + 9 US. 210 +15 nmol/L for cortisol). Spontaneous awakening in the morning was not clearly linked to any specific sleep stage or to rising plasma ACTH and cortisol levels. When the average slopes of ACTH and cortisol profiles were compared in all subjects for a time interval between awakening and 1100 h with average slopes during a time interval of identical length immediately before awakening (ending with awakening) positive slopes of ACTH (0.011 +
W
1
REM
sws
2
FIG. 1. Sleep stage amounts (mean + SEM) for LSG (dotted column, n = 7), and SSG (hatched column, n = 7). W, Wake; S 1, 2, stages 1, 2; SWS, slow wave sleep. Asterisk indicates significant (P < 0.05) differences between both groups. TABLE 1. Percent (mean -t SEM)
of total
sleep time
spent
in each stage of sleep
Results Parameter
SO was similar in both groups (18.4 f 5.6 min in the SSG; 15.0 + 4.8 min in the LSG). Mean total sleep time was 476.9 + 15.0 min in the SSG (range 432-518 min), and 596.9 + 14.4 min in the LSG (range 563-642 min). The time spent in S 2 sleep was significantly longer in the LSG than in the SSG (289.7 + 12.0 US. 221.8 + 8.2 min; P < 0.01). Time spent
W Sl s2 sws REM W, Wake;
All (n = 14) 5.5 13.3 47.5 11.7 22.0
(1.3) (0.9) (1.1) (1.3) (1.6)
S 1, 2, stage 1, 2; SWS,
LSG (n = 7) 5.1 13.7 46.5 12.5 22.3
SSG (n = 71
(1.8) (1.4) (1.5) (1.8) (3.1)
slow wave
6.0 12.8 48.5 10.9 21.8 sleep.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 16 February 2015. at 03:56 For personal use only. No other uses without permission. . All rights reserved.
(1.9) (1.2) (1.9) (2.0) (1.1)
SLEEP,
AWAKENING,
AND
ACTH,
CORTISOL
.I20
.I80
-240
-120
-180
0240
SECRETION
ACTH [pmol/Ll
FIG. 2. Plasma ACTH levels between SO and 660 min. Hatched area corresponds to mean + SEM, SSG (n = 7); dotted urea, LSG (n = 7).
,
so
-60
.300
-360
.420
-480
440
-600
Cminl
-420
-480
.540
.600
Cminl
Cods01 Cn mo 500
400
3. Plasma cortisol levels between SO and 660 min. Hatched area, SSG; dotted area, LSG. FIG.
300
200
100
C
0
-60
0.003 pmol/L X min) and cortisol (0.64 + 0.27 nmol/L X min) were observed before awakening, but negative slopes of ACTH (-0.014 f 0.003 pmol/L X min) and cortisol(-1.10 + 0.13 nmol/L X min) were observed after awakening (Fig. 4). Comparison of the slopes before and after awakening showed significant differences for ACTH (P < 0.001) and cortisol (P < 0.002), indicating that the early morning activation of the HPA system is terminated in close association with awakening.
Discussion In the present study nocturnal hormone patterns were investigated during undisturbed sleep in subjects who, in
-300
*360
contrast to most comparable studies, were allowed to sleep “ad libitum.” This has the advantage that sleep is not interrupted artificially, since as shown previously, sleep manipulations have profound effects on hormone secretion (3). As in earlier studies (15, 16), our groups of short and long sleepers differed significantly in the duration of stage 2 sleep. There was also a tendency towards a longer time spent in REM sleep in long sleepers. Mean cortisol levels in the total sample of 14 subjects during sleep and the latency of the first cortisol rise were comparable to the data from a sample of 25 male subjects in a recent study (17). When the subjects in the present study were divided into two groups according to the duration of sleep we observed significantly higher plasma cortisol levels in short sleepers
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 16 February 2015. at 03:56 For personal use only. No other uses without permission. . All rights reserved.
SPATH-SCHWALBE
1434
JCE & M .1992 Vol75.No6
ET AL. ACTH Cpmol/Ll
ACTH [pmol/Ll 1
7I
01, -120
-60
0
-60
420
[mid
Corlisol Cnmol/Ll
C0h.d
Cnmol/Ll 600 7
1 ssc
1 LSG
0 1, -120
4. Mean of spontaneous
FIG.
40
0
4.0
plasma ACTH and cortisol levels referenced awakening (denoted 0 min) for all subjects
420
[mid
to the time (n = 14).
during their sleepperiod compared with long sleepersduring the same time period. Remarkably, already at the time of sleep onset mean plasma cortisol levels were higher and remained higher during the complete sleepperiod in the SSG compared with the corresponding levels in the LSG. In contrast to cortisol plasma ACTH levels were not reduced but even tended to be higher in the LSG. This reversed relationship between plasma ACTH and cortisol levels may be explained by a negative feedback effect of the elevated cortisol levels in the SSG (18, 19). However, there is also evidence from recent studies for diurnal changes in the adrenocortical responsiveness to ACTH (20, 21), and for substantial interindividual differences in the adrenocortical sensitivity to ACTH (22). Differences in the adrenocortical sensitivity to ACTH, hence, could also account for the reversed relationship between plasma ACTH and cortisol, with a higher sensitivity in the SSG than in the LSG. We failed to observe that spontaneous awakenings predominantly occurred during REM sleep. There was also no close association between rising ACTH and cortisol levels and spontaneousawakening. However, this doesnot exclude that the activation of the HPA axis in the early morning
-120
-60
0
40
420
480
Cminl
plasma ACTH and cortisol levels referenced to the time of awakening (indicated by arrows) for SSG, and LSG. Curves are presented according to the mean difference between the time of awakening in both groups, which was 120 min. FIG.
5. Mean
hours lowers the threshold for awakenings. Since most short spontaneous awakenings during the night were observed after the beginning of the endogenous rise in ACTH and cortisol. Also, in previous studieselevated plasmaACTH and cortisol levels during the night were accompanied by an increase of light sleep and wakefulness (9, 10, 17). Spontaneous awakenings in the morning, followed by wakefulness, mostly induced brief risesin plasma ACTH and cortisol. Similar increasesin ACTH and cortisol releasehave been observed after experimentally induced awakenings in recent studies (3, 7). Therefore we conclude that awakening, i. e. the transition from sleep to wakefulness in the morning, irrespective whether spontaneous or induced, stimulates ACTH and cortisol release.Also the switch from darknessto lights on after awakening may have contributed to the awakening-dependent ACTH and cortisol peak, as has been sug-
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 16 February 2015. at 03:56 For personal use only. No other uses without permission. . All rights reserved.
SLEEP,
AWAKENING,
AND
gested recently (23). Comparing both groups of the present study, the increase of ACTH and cortisol due to awakening appeared to be more pronounced in the LSG than in the SSG. Possibly, this resulted from the lower cortisol levels prior to awakening in the LSG (Fig. 5). This interpretation derives from the finding that prestimulus levels of ACTH and cortisol have a modulatory effect on stimulated ACTH and cortisol release (24). Shortly after awakening, while the subjects kept lying still in bed, the elevated plasma ACTH and cortisol concentrations fell abruptly. This fall during undisturbed wakefulness suggests an important role of the wake phase of the sleepwake cycle in terminating the endogenous early morning activation of the HPA system. Correspondingly, prolonged sleep in the morning was accompanied by sustained endogenous activation of the HPA system. The decline of plasma ACTH and cortisol during wakefulness supports the view that quietly lying awake inhibits adrenocortical activity in the morning, as already suggested by results from previous studies (3, 11). In sum, our data suggest differences in the nocturnal ACTH and cortisol secretion between short and long sleepers. The transition from sleep to wakefulness in the morning is associated with a transition from enhanced to inhibited endogenous HPA secretory activity under undisturbed conditions.
ACTH,
8.
9.
10.
11.
12.
13.
14.
15. 16. 17.
18.
References 1. Kupfer DJ, Bulik CM, Jarret DB. Night time plasma cortisol secretion and EEG sleep-are they associated? Psychiatry Res. 1983;10:191-199, 2. Weitzman ED, Zimmerman JC, Czeisler CA, Ronda J. Cortisol secretion is inhibited during sleep in normal man. J Clin Endocrinol Metab. 1983;56:353-358. 3. Spath-Schwalbe E, Gofferje M, Kern W, Born J, Fehm HL. Sleep disruption alters nocturnal ACTH and cortisol secretory patterns. Biol Psychiatry. 1991;29:575-584. 4. Born J, Kern W, Bieber K, Fehm-Wolfsdorf G, Schiebe M, Fehm HL. Night-time plasma cortisol secretion is associated with specific sleep stages. Biol Psychiatry. 1986;12:1415-1424. 5. Follenius M, Brandenberger G, Simon C, Schlienger JL. REM sleep in humans begins during decreased secretory activity of the anterior pituitary. Sleep. 1988;11:546-555. 6. Orth DN, Island DP, Liddle GW. Experimental alterations of the circadian rhythm in plasma cortisol concentrations in man. J Clin Endocrinol. 1967;27:549-555. 7. Weitzman ED, Nogeire C, Perlow M et al. Effects of a prolonged 3-hour sleep-wake cycle on sleep stages, plasma cortisol, growth
19.
20.
21.
22.
23.
24.
CORTISOL
SECRETION
1435
hormone and body temperature in man. J Clin Endocrinol Metab. 1974;38:1018-1030. Born J, Muth S, Fehm HL. The significance of sleep onset and slow wave sleep for nocturnal release of growth hormone (GH) and cortisol. Psychoneuroendocrinology. 1988;13:233-243. Born J, Spath-Schwalbe E, Schwakenhofer H, Kern W, Fehm HL. Influences of corticotropin-releasing hormone, adrenocorticotropin, and cortisol on sleep in normal man. J Clin Endocrinol Metab. 1989;68:904-911. Fehm HL, Bieber K, Benkowitsch R, Fehm-Wolfsdorf G, Voigt KH, Born J. Relationships between sleep stages and plasma cortisol: a single case study. Acta Endocrinol (Copenh). 1986;111:264-270. Born J, Schenk U, Spath-Schwalbe E, Fehm HL. Influence of partial REM sleep deprivation and awakenings on nocturnal cortisol release. Biol Psychiatry. 1988;24:801-811. Fehm HL, Born J. Evidence for entrainment of nocturnal cortisol secretion to sleep processes in human beings. Neuroendocrinology. 1991;53:171-176. Rechtschaffen A, Kales A. A manual of standardized terminology, techniques, and scoring system for sleep stages of human subjects. Washington DC: National Institute of Neurological Diseases and Blindness, Neurological Information Network: 1968. Hodgkinson SC, Allolio B, Landon J, Lowry PJ. Development of a non-extracted “two-site” immunoradiometric assay for corticotropin utilizing extreme aminoand carboxy-terminally directed antibodies. Biochem J. 1984;218:703-711. Webb WB, Agnew HW. Sleep stage characteristics of long and short sleepers. Science. 1970;168:146-147. Webb WB, Friel J. Sleep stage and personality characteristics of “natural” long and short sleepers. Science. 1971;171:587-588. Steiger A, Herth T, Holsboer F. Sleep-electroencephalography and the secretion of cortisol and growth hormone in normal controls. Acta Endocrinol (Copenh). 1978;116:36-42. Reader SCJ, Alaghbad-Zadeh J, Daly JR, Robertson WR. Negative, rate-sensitive feedback effects on adrenocorticotrophin secretion by cortisol in normal subjects. J Endocr. 1982;92:443-448. Won JGS, Tap TS, Chang SC, Ching KN, Chiang BN. Evidence for a delayed, .integral, and proportional phase of giucocorticoid feedback on ACTH secretion in normal human volunteers. Metabolism. 1986;35:254-259. Dallman MF, Engeland WC, Rose JC, Wilkinson CW. Shinsako J, Siedenburg F. Nycthemeral rhythm in adrenal responsiveness to ACTH. Am J Physiol. 1978;235:R210-R218. Ferrari E, Boss010 PA, Vailate A, Baldi F, Carnevale Schianca GP, Fioravanti M, Romona S. Chrono-sensitivity of the adrenal cortex to exogenous corticotropin. Int J Chronobiol. 1980;7:47-54. Roberts NA, Barton RN, Horan MA. Ageing and the sensitivity of the adrenal gland to physiological doses of ACTH in man. J Endocrinol. 1990;126:507-513. Orth DN, Island DP. Light synchronization of the circadian rhythm in plasma cortisol(17-OHCS) concentration in man. J Clin Endocrinol. 1969;29:479-486. Hermus ARMM, Pieters GFFM, Smals AGH, Benraad TJ, Kloppenborg PWC. Plasma adrenocorticotropin, cortisol, and aldosterone responses to corticotropin-releasing factor: modulatory effects of basal cortisol levels. J Clin Endocrinol Metab. 1984;58:187-191.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 16 February 2015. at 03:56 For personal use only. No other uses without permission. . All rights reserved.
Vol. 15, No. 6 Prmted in U.S.A.
Nocturnal Adrenocorticotropin and Cortisol Secretion Depends on Sleep Duration and Decreases in Association with Spontaneous Awakening in the Morning* ERNST HORST
SP.&TH-SCHWALBE, L. FEHM, AND JAN
THILO BORN
SCHijLLER,
Klinik fiir Innere Medizin, Universittit zu Liibeck, Universittit Bamberg, 8600 Bamberg; Germany
WERNER
2400 Liibeck; and Abteilung
KERN, Psychophysiologie,
ABSTRACT It is still discussed controversially to what extent the nocturnal activity of the hypothalamus-pituitary-adrenocortical system depends on sleep and awakening in the morning. Therefore, we investigated the association of plasma ACTH and cortisol levels with undisturbed nocturnal sleep and spontaneous awakening in 14 healthy male subjects (between 2300 h and 1100 h). Between sleep onset and 476.9 min after sleep onset mean plasma cortisol level was significantly (P < 0.01) higher (210 + 15 us. 155 f 9 nmol/L) in the group with a shorter (476.9 + 15.0 min; n = 7; mean + SEM) than in the group with a longer total sleep time (596.9 -t 14.4 min; n = 7). Spontaneous awakening in the morning was not linked to the presence of any specific sleep stage or
to rising plasma ACTH and cortisol levels. However, spontaneous awakening was followed by a brief rise in plasma ACTH and cortisol in both groups. Thereafter, during wakefulness plasma ACTH and cortisol abruptly declined in all subjects irrespective of the time of awakening. The slope of the plasma ACTH and cortisol curves differed significantly (ACTH: P < 0.001; cortisol: P < 0.002, for all subjects) comparing the time after awakening (until 1100 h) with a time interval of identical length before awakening. We conclude that the duration of sleep and nocturnal ACTH and cortisol secretion are interrelated. Furthermore, the data suggest that the endogenous early morning activation of the hypothalamus-pituitary-adrenocortical system is terminated by mechanisms closely associated with awakening. (J Clin Endocrinol Metab 75: 1431-1435,1992)
E
night (the fourth or fifth) tended to occur during increasing cortisol concentrations, i.e. during increased HPA activity (12). Thus, given that cortisol has been found to decrease REM sleep (9) it was hypothesized that increased HPA activation during the very last REM sleep epochs represents a trigger to wake up spontaneously. We expected that the probability of spontaneous awakenings in the morning is increased when REM sleep occurs in temporal association with increasing plasma ACTH and cortisol levels. On the basisof previous studies (3, 11) we also expected that wakefulness then may initiate mechanisms inhibiting further ACTH and cortisol secretion. Consequently, we expected that the duration of the early morning releaseof ACTH and cortisol depends on the time of the subject’s spontaneous awakening, and thus also on the subject’s sleeping time. To test these hypotheses, in the present study we investigated plasma ACTH and cortisol levels during undisturbed sleep and spontaneous awakening in the morning. Special attention was paid to a comparison of ACTH and cortisol profiles in subjects with long zts.short sleeping time.
VIDENCE has accumulated for relationships between central nervous sleep processesand the secretory activity of the hypothalamus-pituitary-adrenocortical (HPA) system (l-3). Close associations have been found between ultradian rhythms of sleep and episodic ACTH and cortisol secretion (4, 5). The sleep-wake cycle appears to play also an important role in the regulation of the circadian adrenocortical secretory patterns (6-8). Several studies suggestedthat these connections are bidirectional. In a recent study administration of ACTH or cortisol substantially reduced rapid eye movement (REM) sleep, and simultaneously increased time spent awake (9). Furthermore, during undisturbed sleeping conditions increasing endogenous plasma cortisol levels were found to be significantly associatedwith light sleep or wakefulness (4, 10). On the other hand, sleep manipulations which increased the time awake during the night have been found to temporarily reduce ACTH and cortisol secretion (11). The last hours of nocturnal sleep time are characterized by an activation of the hypophyseal-adrenocortical axis. In a previous study, REM sleep appeared to be associatedwith reduced HPA activity (4, 10). However, since the period length of nocturnal HPA activity exceeded that of the nonREM-REM cycles, the very last REM sleep epochs of the Received January 22, 1992. Address all correspondence and requests for reprints to: Dr. E. SpithSchwalbe at this present address: Universitatsklinik Ulm, Medizinische Klinik, Innere III, Robert-Koch-Strasse 8, D-7900 Ulm Germany. * Supported by a grant from Deutsche Forschungsgemeinschaft (to J. B.).
Subjects
and Methods
Fourteen healthy men, aged 20-34 yr (mean 24.9), were studied. The subjects did not suffer from sleep disturbances and were not taking any medication. It was ascertained that they had normal sleep-wake rhythms for at least 2 weeks before the experiments were carried out. The subjects were acclimatized to the experimental setting by an adaptation night under the conditions of the experiment. They were requested to get up between 0700 h and 0800 h before the experimental nights and not to take any naps throughout the day. The experimental nights were carried out in an air-conditioned elec-
1431
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 16 February 2015. at 03:56 For personal use only. No other uses without permission. . All rights reserved.
SPATH-SCHWALBE trically shielded room. Upon arrival at the laboratory at 2100 h the subjects were prepared for polysomnographic recordings and blood sampling. They were informed that they could sleep as long as they wanted, and that they should stay in bed after awakening until 1100 h. They were allowed to turn on the lights after awakening in the morning. Continuous recordings were obtained from 2300 h, the time when lights were turned off, until 1100 h the next morning. Blood for determination of plasma cortisol and ACTH was collected every 15 min from an adjacent room via an iv forearm catheter connected to a long thin tube. To prevent clotting about 500 ml saline (0.9%) were infused throughout the study period.
Recording
and data analysis
Sleep stages were determined from the electroencephalographic, electrooculographic, and electromyographic recordings, which were scored off line according to the criteria described by Rechtschaffen and Kales (13). Blood samples were centrifuged immediately after collection, and plasma was stored at -20 C until assay. Plasma ACTH was measured by immunoradiometric assay (Euro-Diagnostics BV, Apeldoorn, The Netherlands), using two specific antibodies directed against ACTH (216) and ACTH (34-39), respectively (14). Cross-reaction of the corticotropin fragments l-24 and CLIP in the two-site immunoradiometric assay were not observed. The assay detects only intact ACTH. Plasma cortisol was measured by RIA (Hermann Biermann GmbH, Bad Nauheim, Germany). The sensitivities of ACTH and cortisol assays were 1.7 pmol/L and 5.5 nmol/L, respectively. The intraassay precision ranged from 10% at 3.3 pmol/L to 4% at 11 pmol/L for ACTH, and 5% at 27.6 nmol/L to 3% at 415 nmol/L for cortisol. The intraassay coefficients of variation were below 10%. All samples from a subject were analyzed in duplicate in the same assay. For each night, the total sleep time as well as the percentage of time spent in the different sleep stages (wakefulness, W; sleep stage l-4, Sl54; and REM) were calculated. The total sleep time lasted from sleep onset (SO), defined as the onset of the first Sl epoch followed by S2 sleep until the definitive awakening in the morning, when no sleep (defined as any sleep stage Sl-S4 or REM) followed. SO latencies were computed with reference to 2300 h. The total subjects sample was divided into two groups according to their sleep time. Subjects of the short sleep group (SSG) displayed sleep times less than 540 min whereas the subject’s sleep time of the long sleep group (LSG) exceeded 540 min. Both groups consisted of seven subjects. For comparisons between SSG and LSG the individual data were referenced to SO. This presentation served to illustrate the differences in mean hormone levels between the SSG and LSG after SO. The significance of spontaneous awakening in the morning on subsequent hormone secretion was evaluated by comparing the mean levels during a 2-h period preceding and after spontaneous awakening. To assess nocturnal ACTH and cortisol release, the mean plasma concentrations were calculated from the concentrations measured every 15 min. Then, integrated ACTH and cortisol release was determined by calculating the area under the hormone curves. To assess the effect of awakening on plasma hormone levels, for each subject the average slope of the individual hormone curves was calculated for a time interval between awakening and 1100 h, and a time interval of equal duration immediately before awakening. Statistical evaluation was performed with nonparametric tests (Kruskal-Wallis with subsequent Mann-Whitney’s U test and Wilcoxon’s t test).
ET AL.
JCE & M. 1992 .vol75.No6 .
in REM sleep tended to be longer in the LSG without reaching statistical significance (130.1 + 9.4 US. 106.3 + 16.8 min; Fig. 1). No significant differences in the percentage of time spent in the different sleep stages were observed between both groups (Table 1). The mean of the ACTH (Fig. 2) levels between SO and 476.9 min (average sleep time of the SSG) tended to be higher in the LSG compared with the SSG (2.8 + 0.4 VS. 1.9 + 0.4 pmol/L), but this difference remained nonsignificant. Plasma cortisol appeared to rise earlier in the SSG than the LSG. The mean of plasma cortisol (Fig. 3) levels between SO and 476.9 min was significantly higher in the SSG compared with the LSG (210 + 15 US. 155 + 9 nmol/L; P < 0.01). During this time interval integrated cortisol release also was significantly higher (P < 0.01) in the SSG. In contrast to ACTH, all mean cortisol levels of the samples taken between SO and 5 12 min were higher in the SSG than the corresponding cortisol levels in the LSG. However, no significant differences were found, when mean ACTH and cortisol levels were calculated for the total sleep time (LSG IIS. SSG; 3.0 + 0.3 US. 1.9 f 0.4 pmol/L for ACTH, and 201 + 9 US. 210 +15 nmol/L for cortisol). Spontaneous awakening in the morning was not clearly linked to any specific sleep stage or to rising plasma ACTH and cortisol levels. When the average slopes of ACTH and cortisol profiles were compared in all subjects for a time interval between awakening and 1100 h with average slopes during a time interval of identical length immediately before awakening (ending with awakening) positive slopes of ACTH (0.011 +
W
1
REM
sws
2
FIG. 1. Sleep stage amounts (mean + SEM) for LSG (dotted column, n = 7), and SSG (hatched column, n = 7). W, Wake; S 1, 2, stages 1, 2; SWS, slow wave sleep. Asterisk indicates significant (P < 0.05) differences between both groups. TABLE 1. Percent (mean -t SEM)
of total
sleep time
spent
in each stage of sleep
Results Parameter
SO was similar in both groups (18.4 f 5.6 min in the SSG; 15.0 + 4.8 min in the LSG). Mean total sleep time was 476.9 + 15.0 min in the SSG (range 432-518 min), and 596.9 + 14.4 min in the LSG (range 563-642 min). The time spent in S 2 sleep was significantly longer in the LSG than in the SSG (289.7 + 12.0 US. 221.8 + 8.2 min; P < 0.01). Time spent
W Sl s2 sws REM W, Wake;
All (n = 14) 5.5 13.3 47.5 11.7 22.0
(1.3) (0.9) (1.1) (1.3) (1.6)
S 1, 2, stage 1, 2; SWS,
LSG (n = 7) 5.1 13.7 46.5 12.5 22.3
SSG (n = 71
(1.8) (1.4) (1.5) (1.8) (3.1)
slow wave
6.0 12.8 48.5 10.9 21.8 sleep.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 16 February 2015. at 03:56 For personal use only. No other uses without permission. . All rights reserved.
(1.9) (1.2) (1.9) (2.0) (1.1)
SLEEP,
AWAKENING,
AND
ACTH,
CORTISOL
.I20
.I80
-240
-120
-180
0240
SECRETION
ACTH [pmol/Ll
FIG. 2. Plasma ACTH levels between SO and 660 min. Hatched area corresponds to mean + SEM, SSG (n = 7); dotted urea, LSG (n = 7).
,
so
-60
.300
-360
.420
-480
440
-600
Cminl
-420
-480
.540
.600
Cminl
Cods01 Cn mo 500
400
3. Plasma cortisol levels between SO and 660 min. Hatched area, SSG; dotted area, LSG. FIG.
300
200
100
C
0
-60
0.003 pmol/L X min) and cortisol (0.64 + 0.27 nmol/L X min) were observed before awakening, but negative slopes of ACTH (-0.014 f 0.003 pmol/L X min) and cortisol(-1.10 + 0.13 nmol/L X min) were observed after awakening (Fig. 4). Comparison of the slopes before and after awakening showed significant differences for ACTH (P < 0.001) and cortisol (P < 0.002), indicating that the early morning activation of the HPA system is terminated in close association with awakening.
Discussion In the present study nocturnal hormone patterns were investigated during undisturbed sleep in subjects who, in
-300
*360
contrast to most comparable studies, were allowed to sleep “ad libitum.” This has the advantage that sleep is not interrupted artificially, since as shown previously, sleep manipulations have profound effects on hormone secretion (3). As in earlier studies (15, 16), our groups of short and long sleepers differed significantly in the duration of stage 2 sleep. There was also a tendency towards a longer time spent in REM sleep in long sleepers. Mean cortisol levels in the total sample of 14 subjects during sleep and the latency of the first cortisol rise were comparable to the data from a sample of 25 male subjects in a recent study (17). When the subjects in the present study were divided into two groups according to the duration of sleep we observed significantly higher plasma cortisol levels in short sleepers
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 16 February 2015. at 03:56 For personal use only. No other uses without permission. . All rights reserved.
SPATH-SCHWALBE
1434
JCE & M .1992 Vol75.No6
ET AL. ACTH Cpmol/Ll
ACTH [pmol/Ll 1
7I
01, -120
-60
0
-60
420
[mid
Corlisol Cnmol/Ll
C0h.d
Cnmol/Ll 600 7
1 ssc
1 LSG
0 1, -120
4. Mean of spontaneous
FIG.
40
0
4.0
plasma ACTH and cortisol levels referenced awakening (denoted 0 min) for all subjects
420
[mid
to the time (n = 14).
during their sleepperiod compared with long sleepersduring the same time period. Remarkably, already at the time of sleep onset mean plasma cortisol levels were higher and remained higher during the complete sleepperiod in the SSG compared with the corresponding levels in the LSG. In contrast to cortisol plasma ACTH levels were not reduced but even tended to be higher in the LSG. This reversed relationship between plasma ACTH and cortisol levels may be explained by a negative feedback effect of the elevated cortisol levels in the SSG (18, 19). However, there is also evidence from recent studies for diurnal changes in the adrenocortical responsiveness to ACTH (20, 21), and for substantial interindividual differences in the adrenocortical sensitivity to ACTH (22). Differences in the adrenocortical sensitivity to ACTH, hence, could also account for the reversed relationship between plasma ACTH and cortisol, with a higher sensitivity in the SSG than in the LSG. We failed to observe that spontaneous awakenings predominantly occurred during REM sleep. There was also no close association between rising ACTH and cortisol levels and spontaneousawakening. However, this doesnot exclude that the activation of the HPA axis in the early morning
-120
-60
0
40
420
480
Cminl
plasma ACTH and cortisol levels referenced to the time of awakening (indicated by arrows) for SSG, and LSG. Curves are presented according to the mean difference between the time of awakening in both groups, which was 120 min. FIG.
5. Mean
hours lowers the threshold for awakenings. Since most short spontaneous awakenings during the night were observed after the beginning of the endogenous rise in ACTH and cortisol. Also, in previous studieselevated plasmaACTH and cortisol levels during the night were accompanied by an increase of light sleep and wakefulness (9, 10, 17). Spontaneous awakenings in the morning, followed by wakefulness, mostly induced brief risesin plasma ACTH and cortisol. Similar increasesin ACTH and cortisol releasehave been observed after experimentally induced awakenings in recent studies (3, 7). Therefore we conclude that awakening, i. e. the transition from sleep to wakefulness in the morning, irrespective whether spontaneous or induced, stimulates ACTH and cortisol release.Also the switch from darknessto lights on after awakening may have contributed to the awakening-dependent ACTH and cortisol peak, as has been sug-
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 16 February 2015. at 03:56 For personal use only. No other uses without permission. . All rights reserved.
SLEEP,
AWAKENING,
AND
gested recently (23). Comparing both groups of the present study, the increase of ACTH and cortisol due to awakening appeared to be more pronounced in the LSG than in the SSG. Possibly, this resulted from the lower cortisol levels prior to awakening in the LSG (Fig. 5). This interpretation derives from the finding that prestimulus levels of ACTH and cortisol have a modulatory effect on stimulated ACTH and cortisol release (24). Shortly after awakening, while the subjects kept lying still in bed, the elevated plasma ACTH and cortisol concentrations fell abruptly. This fall during undisturbed wakefulness suggests an important role of the wake phase of the sleepwake cycle in terminating the endogenous early morning activation of the HPA system. Correspondingly, prolonged sleep in the morning was accompanied by sustained endogenous activation of the HPA system. The decline of plasma ACTH and cortisol during wakefulness supports the view that quietly lying awake inhibits adrenocortical activity in the morning, as already suggested by results from previous studies (3, 11). In sum, our data suggest differences in the nocturnal ACTH and cortisol secretion between short and long sleepers. The transition from sleep to wakefulness in the morning is associated with a transition from enhanced to inhibited endogenous HPA secretory activity under undisturbed conditions.
ACTH,
8.
9.
10.
11.
12.
13.
14.
15. 16. 17.
18.
References 1. Kupfer DJ, Bulik CM, Jarret DB. Night time plasma cortisol secretion and EEG sleep-are they associated? Psychiatry Res. 1983;10:191-199, 2. Weitzman ED, Zimmerman JC, Czeisler CA, Ronda J. Cortisol secretion is inhibited during sleep in normal man. J Clin Endocrinol Metab. 1983;56:353-358. 3. Spath-Schwalbe E, Gofferje M, Kern W, Born J, Fehm HL. Sleep disruption alters nocturnal ACTH and cortisol secretory patterns. Biol Psychiatry. 1991;29:575-584. 4. Born J, Kern W, Bieber K, Fehm-Wolfsdorf G, Schiebe M, Fehm HL. Night-time plasma cortisol secretion is associated with specific sleep stages. Biol Psychiatry. 1986;12:1415-1424. 5. Follenius M, Brandenberger G, Simon C, Schlienger JL. REM sleep in humans begins during decreased secretory activity of the anterior pituitary. Sleep. 1988;11:546-555. 6. Orth DN, Island DP, Liddle GW. Experimental alterations of the circadian rhythm in plasma cortisol concentrations in man. J Clin Endocrinol. 1967;27:549-555. 7. Weitzman ED, Nogeire C, Perlow M et al. Effects of a prolonged 3-hour sleep-wake cycle on sleep stages, plasma cortisol, growth
19.
20.
21.
22.
23.
24.
CORTISOL
SECRETION
1435
hormone and body temperature in man. J Clin Endocrinol Metab. 1974;38:1018-1030. Born J, Muth S, Fehm HL. The significance of sleep onset and slow wave sleep for nocturnal release of growth hormone (GH) and cortisol. Psychoneuroendocrinology. 1988;13:233-243. Born J, Spath-Schwalbe E, Schwakenhofer H, Kern W, Fehm HL. Influences of corticotropin-releasing hormone, adrenocorticotropin, and cortisol on sleep in normal man. J Clin Endocrinol Metab. 1989;68:904-911. Fehm HL, Bieber K, Benkowitsch R, Fehm-Wolfsdorf G, Voigt KH, Born J. Relationships between sleep stages and plasma cortisol: a single case study. Acta Endocrinol (Copenh). 1986;111:264-270. Born J, Schenk U, Spath-Schwalbe E, Fehm HL. Influence of partial REM sleep deprivation and awakenings on nocturnal cortisol release. Biol Psychiatry. 1988;24:801-811. Fehm HL, Born J. Evidence for entrainment of nocturnal cortisol secretion to sleep processes in human beings. Neuroendocrinology. 1991;53:171-176. Rechtschaffen A, Kales A. A manual of standardized terminology, techniques, and scoring system for sleep stages of human subjects. Washington DC: National Institute of Neurological Diseases and Blindness, Neurological Information Network: 1968. Hodgkinson SC, Allolio B, Landon J, Lowry PJ. Development of a non-extracted “two-site” immunoradiometric assay for corticotropin utilizing extreme aminoand carboxy-terminally directed antibodies. Biochem J. 1984;218:703-711. Webb WB, Agnew HW. Sleep stage characteristics of long and short sleepers. Science. 1970;168:146-147. Webb WB, Friel J. Sleep stage and personality characteristics of “natural” long and short sleepers. Science. 1971;171:587-588. Steiger A, Herth T, Holsboer F. Sleep-electroencephalography and the secretion of cortisol and growth hormone in normal controls. Acta Endocrinol (Copenh). 1978;116:36-42. Reader SCJ, Alaghbad-Zadeh J, Daly JR, Robertson WR. Negative, rate-sensitive feedback effects on adrenocorticotrophin secretion by cortisol in normal subjects. J Endocr. 1982;92:443-448. Won JGS, Tap TS, Chang SC, Ching KN, Chiang BN. Evidence for a delayed, .integral, and proportional phase of giucocorticoid feedback on ACTH secretion in normal human volunteers. Metabolism. 1986;35:254-259. Dallman MF, Engeland WC, Rose JC, Wilkinson CW. Shinsako J, Siedenburg F. Nycthemeral rhythm in adrenal responsiveness to ACTH. Am J Physiol. 1978;235:R210-R218. Ferrari E, Boss010 PA, Vailate A, Baldi F, Carnevale Schianca GP, Fioravanti M, Romona S. Chrono-sensitivity of the adrenal cortex to exogenous corticotropin. Int J Chronobiol. 1980;7:47-54. Roberts NA, Barton RN, Horan MA. Ageing and the sensitivity of the adrenal gland to physiological doses of ACTH in man. J Endocrinol. 1990;126:507-513. Orth DN, Island DP. Light synchronization of the circadian rhythm in plasma cortisol(17-OHCS) concentration in man. J Clin Endocrinol. 1969;29:479-486. Hermus ARMM, Pieters GFFM, Smals AGH, Benraad TJ, Kloppenborg PWC. Plasma adrenocorticotropin, cortisol, and aldosterone responses to corticotropin-releasing factor: modulatory effects of basal cortisol levels. J Clin Endocrinol Metab. 1984;58:187-191.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 16 February 2015. at 03:56 For personal use only. No other uses without permission. . All rights reserved.
