Psychoneuroendocrinology 29 (2004) 174184

www.elsevier.com/locate/psyneuen
Free cortisol awakening responses are
inuenced by awakening time
Ilona Federenko
a,
, Stefan Wust
a
, Dirk H. Hellhammer
a
,
Ralph Dechoux
b
, Robert Kumsta
a
, Clemens Kirschbaum
c
a
Department of Clinical and Theoretical Psychobiology, University of Trier, Karl-Marx-Str. 9496,
54290 Trier, Germany
b
Department of Neurology and Sleep Research, Caritas Clinic, Werkstr. 3, 66763 Dillingen, Germany
c
Department of Experimental Psychology II, University of Dusseldorf, Universitatsstr. 1,
40225 Dusseldorf, Germany
Received 6 May 2002; received in revised form 19 November 2002; accepted 4 December 2002
Abstract
Psychobiological investigations on the hypothalamuspituitaryadrenal (HPA) axis depend
on markers that adequately describe the activity of this system. There is evidence that the free
cortisol response to awakening, proposed as a marker for the HPA axis, can be inuenced by
time of awakening. To further investigate this possible confounder, 24 shift working nurses
and 31 female students on a regular sleepwake cycle collected saliva samples 0, 30, 45 and
60 minutes after awakening. Nurses were investigated on the rst and second day of their
early (awakening: 04000530 h), late (awakening: 06000900 h), and night shift (awakening:
11001400 h), respectively. Students were studied after taking a short nap on two consecutive
weekdays (awakening: 18452030 h). Mean cortisol levels after awakening increased signi-
cantly under all three shift conditions (p0.01), but decreased in the student sample (p.05).
Within the three shift conditions, cortisol responses following waking in the early shift were
more pronounced than in late (p.01) and night shift (p.05). The present study shows that
in a sample with a large range of awakening times, an impact of this variable on the cortisol
awakening response can be observed. The data furthermore strongly suggest that waking up
per se is insufcient for adrenocortical stimulation.
2003 Elsevier Ltd. All rights reserved.
Keywords: Salivary cortisol; Cortisol awakening response; Awakening time; Circadian rhythm; Shift work

Corresponding author. Tel.: +49-651-2013684; fax: +49-651-2013690.

E-mail address: ilona@federenko.de (I. Federenko).
0306-4530/$ - see front matter 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0306-4530(03)00021-0
175 I. Federenko et al. / Psychoneuroendocrinology 29 (2004) 174184
1. Introduction
The hypothalamuspituitaryadrenal (HPA) axis is of major importance with
regard to an organisms response to physical or psychosocial stimulation. The activity
of this endocrine system is characterized by a robust circadian rhythm with cortisol
levels peaking in the early morning hours around the time of awakening and being
lowest around midnight in subjects with a normal sleep-wake cycle (Weitzman et
al., 1971; Gallagher et al., 1973; Van Cauter et al., 1994). Stimulation of the HPA
axis results in higher HPA hormone responses in the morning compared to a compa-
rable stimulation in the evening hours (e.g. Brandenberger et al., 1982; Kanaley et
al., 2001). The circadian rhythm can be modied by alterations in the sleepwake
cycle. Some studies have reported lower early morning cortisol levels after night
work (e.g. Motohashi, 1992; Leese et al., 1996), with a clear reversal of secretion
after the fth night (Hennig et al., 1998). However, in a sample of nurses working
in a rapidly rotating shift system, a signicant alteration of the normal cortisol circad-
ian rhythm could not be observed (Costa et al., 1994).
Research on the HPA axis depends on the existence of adequate markers for this
endocrine system. The cortisol awakening response (CAR) has been proposed as a
new marker for HPA activity. Within 30 minutes after awakening, mean cortisol
increases by 50100% and remains elevated for at least one hour (Pruessner et al.,
1997). The CAR is rather consistent (responder rate of 76.8%), intraindividually
relatively stable (r=0.63 for the area under the curve), and normal values are now
available (Wust et al., 2000b).
Across several studies, no considerable impact of age, smoking habits, use of oral
contraceptives, menstrual cycle phase, sleep quality and physical activity on the CAR
could be found (e.g. Pruessner et al., 1997; Wust et al., 2000b; Kudielka and Kirsch-
baum, 2003). Morning routines were also reported not to affect the CAR, although
Born and colleagues (1999) could demonstrate a higher ACTH response in subjects
who were woken unexpectedly compared to subjects who expected to wake up at
the same time.
Despite its relatively high stability, response magnitude and time course of the
CAR were reported to be signicantly inuenced by gender, persisting pain, burnout,
and chronic stress, each of these variables accounting for 1% to 14% of the variance
explained (Gei et al., 1997; Pruessner et al., 1997; Schulz et al., 1998; Kudielka
and Kirschbaum, 2003). Furthermore, evidence for a signicant genetic impact on
the CAR could be provided (Wust et al., 2000a).
Some studies did not observe a signicant impact of time of awakening on the
subsequent cortisol response (Pruessner et al., 1997; Wust et al., 2000b). In contrast,
Hucklebridge and coworkers (2000) found lower levels in the total area under the
curve (AUC) measure for subjects waking up early, but comparable responses for
the AUC with reference to baseline in early and late awakeners. Other studies could
show a clear impact of awakening time with subjects waking up early having a
higher CAR than subjects waking up late (Edwards et al., 2001; Kudielka and Kirsch-
baum, 2003).
With regard to the growing number of studies using the CAR as a marker for
176 I. Federenko et al. / Psychoneuroendocrinology 29 (2004) 174184
unstimulated adrenocortical activity, the present study aimed to further clarify the
impact of awakening time on this measure. In order to broaden the range of awaken-
ing time, nurses working in a shift system and students who volunteered to sleep in
the early evening hours were investigated. Furthermore, it was intended to elucidate
whether waking up is a sufcient stimulus by itself to trigger a cortisol response, as
has been suggested before (Hucklebridge et al., 2000).
2. Methods
2.1. Sample
Participants in this study were 24 female nurses working in a two or three shift
system at the Caritas Clinic in Dillingen, Germany, and 31 female students with a
regular sleepwake cycle currently enrolled at the University of Trier, Germany.
Working time in the shift conditions was from 0600 h to 1400 h in early shift (ES),
from 1330 h to 2000 h in late shift (LS) and from 2000 h to 0600 h in night shift
(NS). Nurses had a mean age of 40.3 years (2952 yrs, SD=6.97) and students a
mean age of 25.0 years (2031 yrs, SD=2.94). Six nurses and three students were
habitual smokers. All participants reported to be in good health and were medication
free except for the use of oral contraceptives in 19 students. Informed consent was
obtained from all participants. As an incentive, participants were provided with
detailed information on their results in the study. Participating students additionally
received a small monetary incentive after returning all study material.
2.2. Procedure
Nurses were investigated on the rst and second day of their ES, LS and NS,
respectively, to minimize the effect of an altered circadian cortisol rhythm due to
the change of shift condition (Hennig et al., 1998). Students were asked to take a
nap in the early evening hours on two consecutive days. On each day, subjects
obtained saliva samples 0, 30, 45, and 60 minutes after awakening, using the Salivette
sampling device (Sarstedt, Numbrecht, Germany). Subjects agreed not to smoke, eat
or drink other than water in the rst hour after awakening. Furthermore, they were
instructed not to brush their teeth until completion of data sampling, to avoid micro-
injuries in the oral cavity leading to contamination of the sample with blood. Apart
from these restrictions, subjects were free to follow their normal morning routines
(e.g. physical activities, use of alarm clock). Students kept the saliva samples in their
freezer and returned them to the laboratory after the second day, the nurses samples
were frozen in the clinic on each study day and were stored at 20 C until they
were assayed.
All participants lled out a state questionnaire on days of saliva sampling to assess
bed time, total time slept, time of waking up, smoking, medication and general
health status.
177 I. Federenko et al. / Psychoneuroendocrinology 29 (2004) 174184
2.3. Cortisol analysis
Free salivary cortisol was analyzed with a time-resolved immunoassay with uor-
ometric detection as described in detail elsewhere (Dressendorfer et al., 1992). Intra-
and interassay variability of this assay was less than 10 and 12%, respectively.
2.4. Statistical analysis
After salivary free cortisol levels at +0, +30, +45 and +60 minutes were averaged
across two days, two-way ANOVAs (shift by time) with repeated measures on both
factors (three shifts, four saliva samples) were computed to test for differences within
the shift groups. For a comparison between students and nurses, nurses CAR were
rst averaged across shifts. Two-way ANOVAs (group by time) with repeated meas-
ures on one factor (four saliva samples) were then computed. For all ANOVAs,
GreenhouseGeisser corrections were applied when appropriate and effect sizes (w
2
)
as a measure of explained variance were calculated when results were signicant.
Only nurses who provided complete material for all three shifts were included in
these comparisons. Single cortisol measures after awakening were compared with t-
tests for paired (within nurses) or independent (nurses vs. students) samples and an
a-adjustment according to Bonferroni was performed. In order to obtain indices for
the cortisol response, the area under the curve (AUC) with reference to zero as well
as a mean increase (MnInc) were computed
1
(Wust et al., 2000b).
To facilitate comparison of the number of responders in each condition, a secretory
episode was dened as an increase in cortisol level of at least 2.5 nmol/l above
individual baseline (adapted from Weitzman et al., 1971). Pearson correlations of
the AUCs and MnIncs were calculated between the two sampling days to test the
stability of the cortisol response. Furthermore, Pearson correlations were computed
to test for possible associations between the AUCs as well as MnIncs and the sleep
duration within the four study groups. t-tests for paired samples were performed to
test for possible differences of the AUCs and MnIncs within the shift working sam-
ple.
3. Results
Out of the 24 participating nurses, saliva samples for both days were returned by
16 nurses in the ES, by 18 nurses in the LS and by 18 nurses in the NS. 12 nurses
provided complete material in both ES and LS, 12 nurses in both ES and NS and
13 nurses in both LS and NS. Nine nurses returned complete material from all three
shifts. Detailed information on awakening time and sleep duration in the three nurse
1
AUC=(C
0
x30+((C
30
C
0
)x30)/2)+ (C
30
x15+((C
45
C
30
)x15)/2)+(C
45
x15+((C
60
C
45
)x15)/2)
MnInc=(C
30
+C
45
+C
60
)/3C
0
.
178 I. Federenko et al. / Psychoneuroendocrinology 29 (2004) 174184
conditions and the student sample is given in Table 1. Minimum sleep duration in
the student group was 60 minutes.
Cortisol levels for the four single measures as well as the AUC and MnInc showed
a higher stability across days than previously observed in a normative study (Wust
et al., 2000b). Pearson correlations for the single measures ranged between r=0.38
and r=0.74 (r
AUC
=0.65, r
MnInc
=0.56) in the nurse sample and between r=0.53 and
r=0.81 (r
AUC
=0.84, r
MnInc
=0.50; all p0.05) in the student sample. For further analy-
ses, cortisol levels were therefore averaged across days. As previously reported (Wust
et al., 2000b), negative correlations were found between the awakening sample
(+0 min) and the Mean Increase (ES: r=0.46, p=0.07; LS: r=0.64, p0.01; NS:
r=0.38, p=0.12; nap: r=0.79, p0.01), indicating smaller responses after high
baseline levels.
Consistent with earlier ndings, a signicant increase of salivary free cortisol lev-
els was observed in the ES (F
2.0,29.5
=27.38, p0.0001, w
2
=0.39), LS (F
1.6,27.6
=10.66,
p0.01, w
2
=0.28), and NS conditions (F
1.9,32.3
=8.53, p0.01, w
2
=0.25). No signi-
cant interaction of time with smoking behavior was found for any of the three shifts.
A typical response pattern was found for the LS and NS conditions with cortisol
levels peaking at about 45 minutes after awakening and decreasing thereafter. How-
ever, in the ES condition cortisol levels continued to increase over the entire
60 minutes sampling period after awakening. On the other hand, in the student group,
cortisol levels decreased as would be expected with regard to the circadian rhythm
(F
1.7,52.6
=3.93, p0.05, w
2
=0.10). Neither smoking (time by smoking interaction:
F
1.7,50.8
=0.36, n.s.) nor the use of oral contraceptives (F
1.6,44.2
=1.06, n.s.) had a sig-
nicant impact on the cortisol responses (Fig. 1).
Furthermore, responder rates were assessed on both days. In the ES, a secretory
episode occurred in 93.8%, in LS and NS in 77.8% and in the student group in
18.8% of the observations.
A repeated measures ANOVA (three conditions, four samples) including only the
nine nurses providing complete material across sessions, revealed a signicant inter-
action effect condition by sample (F
2.4,19.4
=4.38, p0.05, w
2
=0.35). A main effect
condition was not observed (p=0.12). Post hoc analyses pointed out that the sig-
nicant interaction effect is attributable to the difference between ES and LS
Table 1
Number of observations, range of reported awakening time and reported sleep duration in nurses (n=24)
and students (n=31). Due to an incomplete within subject design in the sample of nurses, data is available
from 16 nurses in the early shift, and from 18 nurses in the late and night shifts, respectively
Sample Condition n Range of reported Mean reported sleep duration
awakening time (h) (min, SD)
Nurses(n=24) Early shift 16 04000530 337 (67)
Late shift 18 06000900 417 (69)
Night shift 18 11001400 318 (67)
Students 31 18452030 105 (24)
179 I. Federenko et al. / Psychoneuroendocrinology 29 (2004) 174184
Fig. 1. Mean cortisol awakening responses (and SEM) in 24 nurses and 31 students. Due to an incom-
plete within subject design in the sample of nurses, data is available from 16 nurses in the early shift,
and from 18 nurses in the late and night shifts, respectively.
(F
1.2,9.3
=5.47, p0.05, w
2
=0.41) and ES and NS conditions (F
1.4,11.3
=7.04, p0.05,
w
2
=0.47), only. A difference between the LS and NS conditions was not detected.
t-tests for independent samples revealed a signicant difference only for the compari-
son ES versus NS 60 minutes after waking up (t
8
=2.53, p=0.035), although the sig-
nicance level is missed after a-correction for four comparisons (p
5%
=0.0125). In
accordance with these results, t-tests for paired samples revealed signicant differ-
ences between the mean increases of the ES versus the LS (t
8
=4.94, p0.01) and a
non-signicant trend for the respective comparison between ES and NS conditions
(t
8
=2.20, p=0.059). The comparison of the AUCs between these two conditions did
not show any differences.
Next, cortisol responses of all three nurse conditions were collapsed into a single
response curve and compared with the student sample. A two-way ANOVA (two
groups, four samples) revealed a pronounced group difference (F
1.0,39.0
=50.93,
p0.0001, w
2
=0.57) as well as a considerable interaction effect (F
1.7,66.6
=28.89,
p0.0001, w
2
=0.43). t-tests revealed a clear difference between the cortisol levels
of students and nurses at 0 (t
39
=2.72, p0.05), 30 (t
39
=9.28, p0.001), 45 (t
39
=7.99,
p0.001) and 60 (t
39
=8.26, p0.001; all after Bonferroni a-correction) minutes after
awakening (Fig. 2).
Finally, possible associations between the CAR and sleep duration were investi-
gated. When sleep duration was averaged across both days, none of the correlations
between sleep duration and the cortisol response after awakening in ES (AUC:
r=0.002; MnInc: r=0.02, n.s.), LS (AUC: r=0.10; MnInc: r=0.27, n.s.), NS
(AUC: r=0.17; MnInc: r=0.24, n.s.) or in the student group (AUC: r=0.22;
MnInc: r=0.03, n.s.) were signicant. Similar results were obtained when correlations
were computed for each day separately.
180 I. Federenko et al. / Psychoneuroendocrinology 29 (2004) 174184
Fig. 2. Mean cortisol awakening responses (and SEM). Left: CAR in the three shift conditions (n=9;
nurses with complete material across all three shifts); right: CAR in the nurse conditions collapsed across
sessions (n=9) compared to the student group (n=31).
4. Discussion
The current study was designed primarily to add knowledge to the growing evi-
dence that time of awakening can have an impact on the subsequent CAR, and
secondly to explore whether waking up is a sufcient stimulus per se to elicit a CAR.
Summarizing the present data, a signicant increase of cortisol occurred in the
rst hour after awakening in all three shift conditions, but not in the student group
after a short nap in the early evening hours. In nurses, the CAR was most pronounced
when awakening occurred very early in the morning (ES group). This response was
signicantly different from cortisol increases in the LS and NS condition, where
awakening occurred later in the morning or around noon time, respectively. Sleep
duration within the shift conditions was not associated with the CAR.
The observed differences within the shift conditions for the nurses support the
hypothesis that an impact of awakening time on the subsequent CAR becomes detect-
able in samples with a distinct variance of awakening time. Our result is in correspon-
dence with two recent reports. Edwards and coworkers (2001) assigned 31 female
and 9 male subjects to a group of early and late awakeners by a median split at
0721 h (range of awakening time: 0500 h1130 h). Early awakeners secreted more
cortisol in the rst 45 minutes after awakening as well as throughout the rest of the
day. These results are in line with a study by Kudielka and Kirschbaum (2003) who
investigated 53 female and 49 male participants, assigned to a group of early (mean
awakening time: 0649 h) versus late (mean awakening time: 0943 h) awakeners
(range of awakening time: 0455 h1203 h). Larger cortisol increases were found in
early awakeners compared to late awakeners. Although both studies covered a wide
range of awakening times (6.5 and 7 h, respectively), it seems reasonable to assume
that a relatively small proportion of participants woke up as early as the nurses in
the ES of the present study. This might explain why cortisol levels in our ES con-
dition continue to increase over the 60 minutes sampling period, while in both other
181 I. Federenko et al. / Psychoneuroendocrinology 29 (2004) 174184
studies cortisol levels start to decrease after the peak at 30 (Edwards et al., 2001)
or 45 (Kudielka and Kirschbaum, 2003) minutes after awakening.
Two earlier studies do not contradict our result, although an impact of awakening
time on the cortisol response could not be observed. Wust and coworkers (2000b)
did not nd a signicant correlation between awakening time (mean: 0745 h) and
the CAR (AUC with reference to zero, mean increase) in 509 subjects. An unpub-
lished reanalysis revealed that awakening in this sample occurred between 0430 h and
1100 h on day 1 and between 0420 and 1245 on day 2. On both days, most subjects
awoke within the time range of this studys late shift condition (day 1: 89.1%,
day 2: 85.9%). Similarly, Pruessner and colleagues (1997) failed to uncover a sig-
nicant correlation between awakening time and the CAR (AUC with reference to
the rst awakening sample, n=152). Since this study did not report awakening time,
one may assume that most subjects awoke within a rather narrow, normal range in
the morning hours, in which the present study failed to uncover an impact of awaken-
ing time as well.
Finally, Hucklebridge and coworkers (2000) found a smaller AUC (with reference
to zero) but similar cortisol increases after nocturnal (mean awakening time: 0403 h)
compared to normal awakening (mean awakening time: 0803 h) in a group of eight
male and three female subjects. One can speculate that these discrepancies are due
to the small sample size in this study, since this same group found a higher AUC
(with reference to zero) in early awakeners in a larger sample (Edwards et al., 2001).
Furthermore, the CAR was monitored only until 30 minutes after awakening, while
the present data indicate that group differences in increases become more prominent
when the observation period is extended.
A rather striking result was the complete lack of mean cortisol responses after
awakening in the group of students who took a nap in the early evening hours.
Obviously, there are many differences between our students and our shift working
nurses as well as subjects in other recent studies on the CAR (e.g. preceding sleep
duration, time of waking up). Nevertheless, this result strongly supports the idea that
waking up by itself is insufcient for a subsequent cortisol increase. The lack of
signicant correlations between sleep duration and the cortisol response within each
study group suggests that awakening time is an important modulating factor for the
subsequent cortisol response. This interpretation is supported by reports that docu-
mented an impact of time of day on cortisol responses to other stimuli. For instance,
a clear cortisol response was observed after a meal at noon time but not in the
evening (Brandenberger et al., 1982). Likewise, Kanaley and coworkers (2001) found
most pronounced cortisol responses to a 30 minute treadmill exercise at 0700 h and
signicantly lower increases at 2400 h and 1900 h. As mentioned above, an alterna-
tive interpretation for the lack of cortisol responsivity in the student group is the
short sleep duration. As was summarized in two recent reviews on the regulation of
the HPA system during sleep (Born and Fehm, 1998, 2000), there is evidence for
inhibited pituitary adrenal (re)activity during the rst few hours of regular nocturnal
sleep. For instance, a blunted cortisol response to intravenous administration of
corticotropin releasing hormone (CRH) was found only in the rst half, but not in
the second half of the night (Spath-Schwalbe et al., 1993). Likewise, suppressed
182 I. Federenko et al. / Psychoneuroendocrinology 29 (2004) 174184
cortisol and ACTH responses were observed after simultaneous administration of
vasopressin and CRH during the rst nocturnal epoch of slow wave sleep (Bierwolf
et al., 1997). In studies investigating sleep at daytime, an inhibitory effect of early
sleep on cortisol was observed in some (e.g. Weitzman et al., 1974, 1983; Van Cauter
et al., 1991) but not in all studies (e.g. Pietrowsky et al., 1994; Weibel et al., 1995).
Although students in the present study were examined after awakening and not dur-
ing early sleep, the above mentioned studies support the idea that the absence of a
cortisol response is attributable to the preceding short period of sleep. Since the
group of students in the present paper is clearly distinguishable by both awakening
time and sleep duration from the group of nurses, it is not possible to conclude which
of these two variables is responsible for the absent cortisol response. Most likely,
both factors contribute to the lack of cortisol responsivity.
One could argue that nurses are not the best sample for the present analysis due
to a possible alteration in their circadian rhythm. However, when nurses woke up
within a normal time range (between 0600 h and 0900 h), they showed a mean
cortisol response that was almost identical with the respective normal values for
women (Wust et al., 2000b). Furthermore, existing differences between nurses and
students (e.g. socioeconomic status, anticipation of a work period vs. a relaxing
evening, physical activity after awakening) and within the nurse conditions (e.g. level
of stress in the shifts, time between waking up and beginning of the shift) could be
regarded as possible confounders. Recently, for example, higher increases in the
CAR were observed on work days compared to weekend days (Schlotz et al., 2002).
However, this and other identied confounders of the CAR explained about 1% to
14% of the variance of this measure (e.g. Wust et al., 2000b; Pruessner et al., 1997;
Kudielka and Kirschbaum, 2003). These effects are smaller than the one found in
the present study (35%). It therefore appears unlikely that the inuence of any of
the known CAR confounders could have signicantly altered the result of the
present study.
It has been demonstrated that light, a factor not controlled in the present study,
has a stimulatory effect on the HPA-axis that might be modulated by melatonin.
This effect seems to be related to time of day, since some studies could show this
effect in the morning, but not in the evening hours (e.g. Scheer and Buijs, 1999;
Leproult et al., 2001; Caufriez et al., 2002). Since we did not control for electric
light, we can only speculate that nurses waking up very early were exposed to less
(naturalistic) light than nurses waking up later. But contrary to our nding, this
should result in lower cortisol increases early in the morning. It should be noted that
an additional stimulatory effect of light might have occurred in nurses only, but not
in students waking up in the evening hours. As a methodological remark, Kudielka,
Broderick and Kirschbaum (in press) point out that subjects uncompliant with regard
to the exact timing of saliva sampling can show atter cortisol responses after awak-
ening. However, this compliance problem should have affected all four study groups
in a comparable manner. An obvious weakness of the present study is, however, that
sleep was not monitored in a sleep laboratory so that it cannot be veried that parti-
cipants slept the complete period in question. In fact, about half of the students
reported that they woke up once during the respective sleep period. A separate analy-
183 I. Federenko et al. / Psychoneuroendocrinology 29 (2004) 174184
sis of students waking up once and students not waking up produced comparable
results so that all subjects were kept in the analysis.
In conclusion, the present study suggests an impact of time of day on the free
CAR in a sample of shift working nurses. Furthermore, it was clearly demonstrated
that waking up per se is insufcient for a subsequent CAR. The question whether
this lack of response is due to the time of day and/or to the short sleep duration
remains to be elucidated. Future studies on the CAR should either dene a minimum
sleep duration and limit awakening to a dened time range or control for both factors
in the statistical analysis.
Acknowledgements
The authors wish to thank Aileen Wiglesworth, Ph.D., for her very helpful com-
ments on the present manuscript. This study was supported by the INTEREG II pro-
gram.
References
Bierwolf, C., Struve, K., Marshall, L., Born, J., Fehm, H.L., 1997. Slow wave sleep drives inhibition of
pituitaryadrenal secretion in humans. J. Neuroendocrinol. 9, 479484.
Born, J., Fehm, H.L., 1998. Hypothalamuspituitaryadrenal activity during human sleep: a coordinating
role for the limbic hippocampal system. Exp. Clin. Endocrinol. Diabetes 106, 153163.
Born, J., Fehm, H.L., 2000. The neuroendocrine recovery function of sleep. Noise and Health 7, 2537.
Born, J., Hansen, K., Marshall, L., Molle, M., Fehm, H.L., 1999. Timing the end of nocturnal sleep.
Nature 397, 2930.
Brandenberger, G., Follenius, M., Hietter, B., 1982. Feedback from meal-related peaks determines diurnal
changes in cortisol response to exercise. J. Clin. Endocrinol. Metab. 54, 592596.
Caufriez, A., Moreno-Reyes, R., Leproult, R., Vertongen, F., Van Cauter, E., Copinschi, G., 2002.
Immediate effects of an 8-h advance shift of the restactivity cycle on 24-h proles of cortisol. Am.
J. Physiol. Endocrinol. Metab. 282, E1147E1153.
Costa, G., Ghirlanda, G., Tarondi, G., Minors, D., Waterhouse, J., 1994. Evaluation of a rapidly rotating
shift system for tolerance of nurses to nightwork. Int. Arch. Occup. Environ. Health 65, 305311.
Dressendorfer, R.A., Kirschbaum, C., Rohde, W., Stahl, F., Strasburger, C.J., 1992. Synthesis of a cor-
tisolbiotin conjugate and evaluation as a tracer in an immunoassay for salivary cortisol measurement.
J. Steroid. Biochem. Mol. Biol. 43, 683692.
Edwards, S., Evans, P., Hucklebridge, F., Clow, A., 2001. Association between time of awakening and
diurnal cortisol secretory activity. Psychoneuroendocrinology 26, 613622.
Gallagher, T.F., Yoshida, K., Roffwarg, H.D., Fukushima, D.K., Weitzman, E.D., Hellman, L., 1973.
ACTH and cortisol secretory patterns in man. J. Clin. Endocrinol. Metab. 36, 10581068.
Gei, A., Varadi, E., Steinbach, K., Bauer, H.W., Anton, F., 1997. Psychoneuroimmunological correlates
of persisting sciatic pain in patients who underwent discectomy. Neurosci. Lett 237, 6568.
Hennig, J., Kieferdorf, P., Moritz, C., Huwe, S., Netter, P., 1998. Changes in cortisol secretion during
shiftwork: implications for tolerance to shiftwork? Ergonomics 41, 610621.
Hucklebridge, F., Clow, A., Rahman, H., Evans, P., 2000. The cortisol response to normal and nocturnal
awakening. Psychophysiology 14, 2428.
Kanaley, J.A., Weltman, J.Y., Pieper, K.S., Weltman, A., Hartman, M.L., 2001. Cortisol and growth
hormone responses to exercise at different times of day. J. Clin. Endocrinol. Metab. 86, 28812889.
Kudielka, B.M., Broderick, J.E., Kirschbaum, C., in press. Compliance with saliva sampling protocols:
184 I. Federenko et al. / Psychoneuroendocrinology 29 (2004) 174184
electronic monitoring reveals invalid cortisol daytime proles in noncompliant subjects. Psychosom.
Med.
Kudielka, B.M., Kirschbaum, C., 2003. Awakening cortisol responses are inuenced by health status and
awakening time but not by menstrual cycle phase. Psychoneuroendocrinology 28, 3547.
Leese, G., Chattington, P., Fraser, W., Vora, J., Edwards, R., Williams, G., 1996. Short-term night-
shift working mimics the pituitaryadrenocortical dysfunction in chronic fatigue syndrome. J. Clin.
Endocrinol. Metab. 81, 18671870.
Leproult, R., Colecchia, E.F., LHermite-Baleriaux, M., Van Cauter, E., 2001. Transition from dim to
bright light in the morning induces an immediate elevation of cortisol levels. J. Clin. Endocrinol.
Metab. 86, 151157.
Motohashi, Y., 1992. Alteration of circadian rhythm in shift-working ambulance personnel. Monitoring
of salivary cortisol rhythm. Ergonomics 35, 13311340.
Pietrowsky, R., Meyrer, R., Kern, W., Born, J., Fehm, H.L., 1994. Effects of diurnal sleep on secretion
of cortisol, luteinizing hormone, and growth hormone in man. J. Clin. Endocrinol. Metab. 78, 683687.
Pruessner, J.C., Wolf, O.T., Hellhammer, D.H., Buske-Kirschbaum, A., Von Auer, K., Jobst, S., Kaspers,
F., Kirschbaum, C., 1997. Free cortisol levels after awakening: a reliable biological marker for the
assessment of adrenocortical activity. Life Sci 61, 25392549.
Scheer, A.J.L., Buijs, R.M., 1999. Light affects salivary cortisol in humans. J. Clin. Endocrinol. Metab.
84, 33953398.
Schlotz, W., Hellhammer, J., Schulz, P., Stone, A.A., Hellhammer, D.H., 2002. Perceived work-overload
and chronic worrying predict weekendweekday differences in the cortisol awakening response. Stress
5 (Suppl.), 68.
Schulz, P., Kirschbaum, C., Pruessner, J., Hellhammer, D.H., 1998. Increased free cortisol secretion after
awakening in chronically stressed individuals due to work overload. Stress Med 14, 9197.
Spath-Schwalbe, E., Uthgenannt, D., Voget, G., Kern, W., Born, J., Fehm, H.L., 1993. Corticotropin-
releasing hormone-induced adrenocorticotropin and cortisol secretion depends on sleep and wakeful-
ness. J. Clin. Endocrinol. Metab. 77, 11701173.
Van Cauter, E., Blackman, J.D., Roland, D., Spire, J.P., Refetoff, S., Polonsky, K.S., 1991. Modulation
of glucose regulation and insulin secretion by circadian rhythmicity and sleep. J. Clin. Invest. 88,
934942.
Van Cauter, E., Sturis, J., Byrne, M.M., Blackman, J.D., Leproult, R., Ofek, G., LHermite Baleriaux,
M., Refetoff, S., Turek, F.W., Van Reeth, O., 1994. Demonstration of rapid light-induced advances and
delays of the human circadian clock using hormonal phase markers. Am. J. Physiol. 266, E953E963.
Weibel, L., Follenius, M., Spiegel, K., Ehrhart, J., Brandenberger, G., 1995. Comparative effect of night
and daytime sleep on the 24-hour cortisol secretory prole. Sleep 18, 549556.
Weitzman, E.D., Fukushima, D., Nogeire, C., Roffwarg, H., Gallagher, T.F., Hellman, L., 1971. Twenty-
four hour pattern of the episodic secretion of cortisol in normal subjects. J. Clin. Endocrinol. Metab.
33, 1422.
Weitzman, E.D., Nogeire, C., Perlow, M., Fukushima, D., Sassin, J., McGregor, P., Hellman, L., 1974.
Effects of a prolonged 3-hour sleepwake cycle on sleep stages, plasma cortisol, growth hormone and
body temperature in man. J. Clin. Endocrinol. Metab. 38, 10181030.
Weitzman, E.D., Zimmerman, J.C., Czeisler, C.A., Ronda, J., 1983. Cortisol secretion is inhibited during
sleep in normal man. J. Clin. Endocrinol. Metab. 56, 352358.
Wust, S., Federenko, I., Hellhammer, D.H., Kirschbaum, C., 2000a. Genetic factors, perceived chronic
stress, and the free cortisol response. Psychoneuroendocrinology 25, 707720.
Wust, S., Wolf, J., Hellhammer, D.H., Federenko, I., Schommer, N., Kirschbaum, C., 2000b. The cortisol
awakening response normal values and confounds. Noise and Health 7, 7988.