Biological Psychology 74 (2007) 92–103

www.elsevier.com/locate/biopsycho

Circadian cortisol profiles and psychological self-reports in shift workers

with and without recent change in the shift rotation system
Brigitte M. Kudielka *, Jörg Buchtal, Alexander Uhde, Stefan Wüst
Department of Theoretical and Clinical Psychobiology, University of Trier, Johanniterufer 15, D-54290 Trier, Germany
Received 24 February 2006; accepted 17 August 2006
Available online 13 November 2006

Abstract
Cortisol profiles including the cortisol rise in the first hour after awakening (CAR) were assessed during shift work and days off (eight saliva
samples per shift). Participants were 102 healthy permanent day and night shift workers (comparison groups) and former permanent day and night
shift workers after implementation of a new fast-forward rota including morning, evening, and night shifts. Results show that the CAR is detectable in
day as well as night shifts. In permanent night workers cortisol profiles appear to be blunted during night work and days off. However, circadian cortisol
profiles are not disturbed in former night workers who recently switched to the fast rotating shift schedule. In contrast, implementation of night work in
former day workers seems to lead to initially blunted cortisol profiles that normalize after a short adjustment period. Results of a psychological
assessment including exhaustion, chronic stress, effort–reward imbalance, and ratings of sleep quality and sleep length are also presented.
# 2006 Elsevier B.V. All rights reserved.

Keywords: Cortisol awakening rise (CAR); Cortisol circadian rhythm; Shift work; Night shift; Salivary cortisol; Psychological assessment; Sleep

1. Introduction
In industrialized countries, shift work is rather common with
approximately 20% of the workforce being engaged in shift
work and one third of shift workers working permanently at
night (see Wedderburn, 1992; Knutsson et al., 1999; see Costa,
2003). Shift work interferes, at least, with basic biological
functions, work ability (Folkard and Tucker, 2003), social
relationships and psycho-physical health conditions (Costa,
2003). Night work causes a mismatch between the endogenous,
internal circadian timing system and the external environmental
synchronizers, with consequent disturbances of the normal
circadian rhythms of psycho-physiological functions. Conse-
quently, the body repeatedly attempts to adjust as quickly as
possible to varying working hours. A complete inversion of the
circadian rhythm is seldom reached, even across successive
Abbreviations: AUC, area-under-the-curve; CAR, cortisol awakening rise;
ERI, effort–reward imbalance; HPA axis, hypothalamus–pituitary–adrenal axis;
SCN, suprachiasmatic nucleus; S.D., standard deviation; S.E.M., standard error
of mean; TICS, Trier Inventory for the Assessment of Chronic Stress; VE, vital
exhaustion
* Corresponding author. Tel.: +49 651 201 2981; fax: +49 651 201 3690.
E-mail address: kudielka@uni-trier.de (B.M. Kudielka).
night shifts and after years of night work experience (Roden
et al., 1993; Weibel et al., 1996; Weibel and Brandenberger,
1998; Costa, 2003). Shift work is recognized to be a serious risk
factor for workers’ health. In particular, night work is
associated with cardiovascular and gastrointestinal diseases,
sleep disorders, and may promote metabolic disturbances or
depression (Tuchsen et al., 1994; Knutsson et al., 1999;
Holmbäck et al., 2003; for reviews see Boggild and Knutsson,
1999; Costa, 2003; Knutsson, 1989, 2003). In sum, possible
mechanisms linking shift work with health problems are the
mismatch of circadian rhythms (internal desynchronization),
sleep problems (Regestein and Monk, 1991; Spiegel et al.,
1999; Vgontzas et al., 1999, 2004), behavioral changes (life
style factors, stress) and disturbed socio-temporal patterns (for
overviews see Knutsson, 1989, 2003).
The adrenal hormone cortisol shows a marked circadian
rhythm with typically lowest secretion during the first half of
night time sleep (quiescent period), an abrupt elevation during
the second half of sleep, peak levels shortly after awakening
and continuously decreasing levels over the remainder of the
day, except for stress-related cortisol surges that superimpose
on the normal circadian rhythm (Van Cauter, 1990; Weibel
et al., 1995, 1996; Born et al., 1999; Czeisler and Klerman,
1999; Kirschbaum and Hellhammer, 2000). The two major

0301-0511/$ – see front matter # 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.biopsycho.2006.08.008
 B.M. Kudielka et al. / Biological Psychology 74 (2007) 92–103
processes which control the overall diurnal cortisol variation
are the circadian signal generated by the circadian pacemaker
located in the suprachiasmatic nucleus (SCN) of the
hypothalamus, and the alteration of wakefulness and sleep
(Van Cauter, 1990; Czeisler and Klerman, 1999).
In the last years, data has accumulated showing that this
typical circadian cortisol rhythm is altered during shift work,
especially in night duties. Previous studies in experimental
short-term shift work (Fujiwara et al., 1992; Goichot et al.,
1998) or under naturalistic conditions (Fujiwara et al., 1992,
2004; Costa et al., 1994; Kobayashi et al., 1997; Munakata
et al., 2001) concluded that early morning free and total cortisol
levels are reduced after only one night of shift work. This is
mainly caused by the absence of awakening-related activation
(Van Cauter, 1990) and by a phase-advancement of the
circadian peak time (Motohashi, 1992). In a controlled
laboratory sleep study, Weibel and coworkers (1996, 1998)
monitored total plasma cortisol levels in 10-min intervals in
permanent night workers during their regular sleep. Despite an
adapted sleep structure, cortisol levels were abnormally
enhanced during usual day sleep while alterations persisted
during work time. Furthermore, the acrophase was almost
adequately shifted by 6.5 h whereas the quiescent period
showed only a partial shift of 3 h and was thus uncoupled in
regard to the day sleep period.
To date, there are only few studies (including single case
observations) that monitored circadian cortisol profiles during
shift work (Touitou et al., 1990; Motohashi, 1991, 1992;
Fujiwara et al., 1992; Roden et al., 1993; Shinkai et al., 1993;
Costa et al., 1994; Lac and Chamoux, 2003). Under controlled
laboratory conditions, Hakola et al. (1996) collected saliva
samples in 2-h intervals in 20 shift workers during one
laboratory day and three consecutive night shifts. Although the
pattern of cortisol changed significantly between the work
shifts, there was a very slow and only partial adjustment. Goh
et al. (2000) investigated how night duties affected the circadian
cortisol rhythm in military personnel. On a 3-day sortie of a
navy ship, 60 male military servants collected saliva samples
every 2 h within two 24-h periods. Beside a wide variation in
individual endocrine responses to night shift work, the cortisol
cycle seemed to be disrupted in the total group. Lac and
Chamoux (2004) compared free cortisol profiles in 16-day
workers and two different shift work groups each comprised of
eight subjects performing different morning, evening, and night
shift rotas. Night duties induced a consistent change in the
cortisol circadian profile with both a displacement of the
acrophase and a global flattening of the cortisol profile.
Although, there was a tendency for the cortisol circadian
pattern to synchronize with altered awakening times, synchro-
nization was not completed. In a sample of 24 nurses, Hennig
et al. (1998) observed higher evening than morning cortisol
values after the fifth out of seven consecutive night duties in 18
subjects. Accordingly, reviewing laboratory studies Van Cauter
(1990) summarized that the cortisol rhythm takes 5–10 days to
fully adapt to a shift in light–dark and sleep–wake cycles.
The aim of the present study was to compare circadian
cortisol profiles between different groups of shift workers with
93
a main focus on the cortisol awakening rise (CAR). Evidence
has accumulated that the CAR can serve as a useful index of
adrenocortical activity when measured with strict reference to
time of awakening (Pruessner et al., 1997; Wüst et al., 2000b;
for review see Clow et al., 2004). Sample collection was
undertaken in 118 industrial workers of the same plant shortly
after implementation of a new rapid-rotating shift system in two
pilot test divisions. This organizational change offered the
chance to compare cortisol levels across shifts and days off in
permanent night and day shift workers with former permanent
night and day shift workers who recently switched to a new rota
including fast rotating early, late, and night shifts.
2. Materials and methods
2.1. Study design and recruitment of participants
Shift workers of a globally operating company in the electronic manufac-
turing sector were recruited from a plant located in West Germany. Four
different shift groups volunteered for study participation rendering a total study
sample of N = 118. Groups were comprised according to the workers current
and former shift schedule. Regularly, workers in this company belong to a night
shift group (permanent night shift workers) or to a day shift group with weekly
forward rotating morning and evening shifts. Night workers work 3 weeks of six
consecutive night shifts followed by a week off with each night shift lasting
from 10 p.m. to 6 a.m. (N shift = 8 h). For day workers, morning shifts last from
6 a.m. to 2 p.m. (M shift = 8 h) and evening shifts are from 2 p.m. to 10 p.m. (E
shifts = 8 h). Three weeks before study onset the company introduced a rapidly
rotating shift schedule in two pilot test divisions consisting of one former night
and one former day shift group, in the following referred to as P1 and P2. Two
groups of shift workers of the regular shift rota served as comparison groups
consisting of permanent night shift workers or day shift workers, referred to as
C1 and C2. The choice of the particular test divisions was solely based on
organizational aspects; in general the assignment of workers to divisions with
the same shift schedule is accidental.
The new shifts followed the above mentioned time points for start and end of
working shifts and the schedule was similar to the so-called ‘Metropolitan rota’
(Costa et al., 1994) with two morning shifts, two evening shifts, and up to three
night shifts, followed by 1–4 days off. Measurements in all four groups were
undertaken about 3 weeks after implementation of the new rapid-rotating shift
system. P groups were instructed to collect saliva samples at the second day of
each shift during one shift cycle and the second of three consecutive free days
resulting in the following sampling pattern MMEENN(N)XX(X)(X) (sampling
days underlined, X = day off). The C1 group (permanent night workers)
collected samples throughout their second day of night shift (night 1) and
additionally during the fifth out of six night shifts in a row and the second of
consecutive free days resulting in the following sampling pattern
NNNNNNXXX, etc. The C2 group (permanent day workers) selected the
second day of a morning shift week and an evening shift week, respectively, and
the second of consecutive free days resulting in the following sampling pattern
MMMMMMXEEEEEEXXX, etc. A summary of the four study groups includ-
ing their rotating shifts and sampling patterns are given in Table 1. A second
measurement could be realized 2 months later in the P2 group (former
permanent day workers) following the same sampling pattern (referred to as
T2). The other groups did not participate at T2 due to financial and organiza-
tional restrictions. Informed consent was obtained from the company as well as
all study participants. As incentive, each participant was offered a voucher for
two spa entries.
2.2. Sample collection and biochemical analysis
On each sampling day (see above), subjects gained a total of eight saliva
samples. Samples were obtained 0, +30, +45, and +60 min after awakening to
assess CAR as well as +4, +8, +12, and +16 h after awakening to assess cortisol
levels during working time. Saliva samples were collected using polypropylene
94 B.M. Kudielka et al. / Biological Psychology 74 (2007) 92–103
Table 1
Scheme of different shift groups and rotating shifts in the four study groups C1, C2, P1 and P2
N indicates subjects entering final analyses; M: morning shift, E: evening shift, N: night shift, X: day off, sampling days underlined.
tubes (IBL, Hamburg, Germany). Subjects were instructed not to brush their
teeth before completing the first four saliva samples (awakening samples) to
avoid contamination of saliva with blood caused by micro-injuries in the oral
cavity. Food intake and smoking was not allowed until the first four saliva
samples were collected and 60 min before each of the remainder samples,
respectively. Besides these restrictions, subjects were free to follow their normal
daily routines on all sampling days. Subjects obtained detailed oral and written
instructions handed over personally by one of two coauthors (JB, AU).
Participants stored the saliva samples in their freezers until completing the
experimental protocol and then returned the samples to the experimenters on
site.
Salivary cortisol was assayed with a time-resolved immunoassay with
fluorescence detection (DELFIA) to obtain levels of free unbound cortisol,
the biologically active moiety of cortisol. Intra- and interassay variability of the
assay were between 4.0–6.7% and 7.1–9.0%, respectively.
2.3. Psychological assessment
Before saliva sampling, a questionnaire package was administered for the
assessment of demographics and health status, sleep, vital exhaustion, perceived
chronic stress, effort–reward imbalance and overcommitment.
2.3.1. Demographics and health status
Information on sex, age, shift work history, number of children, engagement
in child care and other extra (non-paid) duties, smoking habits, and acute or
chronic illness was gained.
2.3.2. Sleep
Firstly, the average sleep quality was assessed for each shift separately (M,
E, N, day off). This rating consisted of a one-item question on average perceived
sleep quality over the last couple of weeks ranging from 1 = very bad to 5 = very
good, mean sleep duration during a 24 h cycle, and secondly the Jenkins sleep-
quality-index (Jenkins et al., 1988). This scale comprises four items with a six-
point answering format focusing on the most common sleep problems (‘‘diffi-
culties falling asleep’’, ‘‘waking up during sleep’’, ‘‘waking up having diffi-
culties falling asleep again’’, ‘‘waking up tired). The internal consistency
coefficient was reported to be 0.79. Thirdly, a one-item rating on perceived
sleep quality was given for each specific sampling day.
2.3.3. Maastricht-vital-exhaustion-questionnaire
Vital exhaustion (VE) is a psychological (not a physical) state, which is
defined by mental fatigue, increased irritability, and demoralization. We applied
the nine-item short form (Appels et al., 1987; Kopp et al., 1998) assessing undue
tiredness, troubles falling asleep, repeated waking up at night, general malaise,
listlessness, irritability, loss of energy, demoralization, and waking up
exhausted. The nine-point answering format gives rise to scores from 0 to
18. Cronbach’s alpha was 0.86, reflecting good reliability.
2.3.4. Trier inventory for the assessment of chronic stress (TICS)
Perceived stress was assessed by the ‘Trier Inventory for the Assessment of
Chronic Stress (TICS)’, which was developed in this laboratory (Schulz et al.,
2004). Subjects have to indicate how often the described stressful situations
were experienced during the past 3 months. The 57-items TICS comprises nine
subscales, namely ‘work overload’, ‘work discontent’, ‘social overload, ‘lack of
social recognition’, ‘pressure to succeed’, ‘excessive demands’, ‘social stress’,
‘social isolation’, and ‘worrying’. Reported internal consistency (Cronbach’s
alpha) ranges between 0.84 and 0.91.
2.3.5. Effort–reward imbalance
Effort and reward were assessed by six and eleven items, respectively (five-
point rating scales) (all items are provided in Siegrist et al., 2004). Effort refers
to psychologically and physically demanding aspects of the work environment.
The reward scale focuses on rewards received at work which are offered to the
worker as part of a social exchange process in terms of monetary remuneration,
social approval/esteem, job security or career opportunities. The ratio of effort
and reward expresses the amount of perceived effort–reward imbalance at work
(ERI). Overcommitment (OC) was assessed by the six-item short form of the
intrinsic effort scale (four-point rating scale). OC stands for a specific individual
pattern of coping with job demands and eliciting rewards at work designating a
personality trait. Items assessing overcommitment focus on the ‘‘inability to
withdraw from work’’ and ‘‘disproportionate irritability’’. Internal consistency
(Cronbach’s alpha) was tested in up to ten different studies (Siegrist et al.,
2004). For the scale effort, alpha varies from 0.61 to 0.88, for reward from 0.70
to 0.91 and for overcommitment from 0.64 to 0.81.
2.4. Statistical analyses
Crosstabulation, chi-square tests, and UNIANOVAs were applied to com-
pare groups in respect to demographic variables and health status. One-way
 B.M. Kudielka et al. / Biological Psychology 74 (2007) 92–103
ANOVAs were used to compare awakening times across groups. General linear
model procedures (GLMs) with the repeated factor ‘‘time’’ (eight saliva
samples) and shift group as grouping variable (C1, C2, P1, and P2) were
applied to investigate cortisol profiles in each shift (morning, evening, night
shift, day off). To control for sex, smoking, and individual time of awakening
these variables were entered as covariates (Wüst et al., 2000b; Edwards et al.,
2001; Kudielka and Kirschbaum, 2003; Federenko et al., 2004). Where
appropriate, results were corrected by Greenhouse–Geisser procedure, indi-
cated by uneven degrees of freedom. Effect sizes were calculated by partial eta-
squared (h2), expressing the amount of variance explained in the dependent
variable cortisol by the respective effect. Subsequent GLMs focused separately
on the four CAR and four working time samples. Areas-under-the-curve
(AUCs) were computed to compare overall cortisol secretion for each shift
by Student’s t-tests following the trapezoid formula with reference to zero and
accounting for different time spans between samples. GLMs with the repeated
factor ‘‘time’’ and shift as grouping variable were additionally applied to
compare cortisol profiles within each group. For the C1 (permanent night
workers), both sampling nights could be entered here. For the P2 group, this
analysis was repeated for the second measurement 2 months later (T2).
Additionally, for comparison of the first and second measurement in the P2,
AUCs were compared for each shift by Student’s t-tests. Finally, UNIANOVAs
and Student’s t-tests, respectively, were applied to compare questionnaire data
between shift groups and between shifts within a group. Bonferroni correction
of the nominal alpha error was determined to adjust for multiple comparisons.
Statistical analyses were performed using the SPSS statistical software
package (version 12.0.2; Chicago, Illinois, USA). Results in the text are given as
mean  standard deviation (S.D.); data in figures are presented as mean  stan-
standard error of mean (S.E.M.).
3. Results
3.1. Description of sample
Out of the 118 subjects, 16 reported an acute illness on
sampling days, a major chronic disease or regular medication
intake and were therefore excluded from data analysis. This
rendered a final sample of 102 healthy subjects with N = 26 in the
P1 (2 excluded), N = 28 in the P2 (5 excluded), N = 26 in the C1
(2 excluded), and N = 22 in the C2 (7 excluded). Interestingly in
the original sample, the C1 and P1 (former/night workers)
reported less health problems compared to the C2 and P2
(former/day workers) (x2[1] = 0.053). The remaining sample
comprised 63% men, had a mean age of 40.3  7.3(S.D.) years
and had on average 10.4  4.7(S.D.) years of shift work
experience with 0–100% night shift duties. Age, smoking, self-
reported engagement in child care, and other extra duties were
not differently distributed across shift groups (age: F[1] = 4.6,
p = 0.71; smoking: x2[3] = 0.63, child care: F[3] = 1.2, p = 0.30,
extra duties: F[1] = 0.58, p = 0.63). However, significantly more
women belonged to the C groups while more men were in the P2
(x2[3] = 0.023). The C2 has worked marginally shorter in shift
systems (P1: 11.2  5.5 years, P2: 11.7  4.9 years, C1
10.5  4.1 years, and C2 8.1  3.6 years; F[3] = 2.6,
p = 0.058). Finally, participants in the P1 and C1 had more
children than P2 and C2 (F[3] = 4.1, p = 0.01).
3.2. Awakening times
Within a respective shift, awakening times did not differ
significantly between groups showing the following mean
awakening times: morning shift 4:38  2:03 h (F[3] = 0.17,
p = 0.84), evening shift 8:27  1:29 h (F[3] = 1.6, p = 0.21),
95
night shift 13:46  1:34 h (F[3] = 0.39, p = 0.68), and day off
8:31  2:01 h (F[3] = 0.741, p = 0.53).
3.3. Comparison of cortisol profiles between shift groups
across shifts
Within the morning shift, cortisol profiles followed the typical
circadian course showing the cortisol awakening rise (CAR)
during the first hour after awakening and decreasing levels
thereafter (main effect time: F[3.3, 175.5] = 3.28, p = 0.02,
h2 = 0.06). No differences emerged between the P1, P2, and C2
(main effect group: F[2, 55] = 0.81, p = 0.45; interaction time by
group: F[6.8, 55] = 0.36, p = 0.91; see Fig. 1A). Accordingly,
separate analyses of CAR and working time samples did not
reveal any group differences (both main group effects and time by
group interactions non-significant). The same pattern of results
emerged for the evening shift (main effect time: F[2.4,
141.5] = 3.20, p = 0.034, h2 = 0.052; main effect group: F[2,
58] = 1.83, p = 0.17; interaction time by group: F[4.9,
58] = 1.06, p = 0.39; see Fig. 1B). Again, subsequent analyses
of CAR and working time samples did not reveal any group
differences (all group effects non-significant). In the night shift,
the overall GLM revealed a significant main effect group (F[2,
64] = 3.77, p = 0.028, h2 = 0.11) as well as interaction time by
group (F[5.9, 64] = 2.9, p = 0.01, h2 = 0.08), showing a
significantly flattened cortisol awakening rise in the P2 (main
effect group: p = 0.009; interaction time by group: p = 0.046) but
no group differences during the remainder of the day (main effect
group: p = 0.83; interaction time by group: p = 0.56; see Fig. 1C).
At the day off, result showed a marginally significant main effect
time (F[2.9, 181.3] = 2.48, p = 0.064, h2 = 0.039), a significant
main effect group (F[2.8, 62] = 2.83, p = 0.045, h2 = 0.12) as
well as interaction time by group (F[8.8, 62] = 2.98, p = 0.003,
h2 = 0.126; see Fig. 1D). Post hoc analysis excluding the C2 or
C1, respectively, revealed that the C1 profile was marginally
flatter compared to the P1 and P2 (main effect group: p = 0.11;
interaction time by group: p = 0.07) and the C2 profile was
significantly higher compared to P1 and P2 (main effect group:
p = 0.016; interaction time by group: p = 0.03). In the separate
analyses of CAR and working time samples group differences
remained significant (CAR main effect group: p = 0.046;
interaction time by group: p = 0.033; working time samples
main effect p = 0.82; interaction time by group: p = 0.008).
The following AUC analysis presents crude estimates of
overall cortisol secretory activity during measurement periods.
Comparisons of AUCs between shifts revealed that on average
AUCnight shift was significantly smaller than the AUCmorning shift
(t[40] = 4.75, p = 0.0001) as well as AUCevening shift (t[44] =
2.99, p = 0.005), AUCmorning shift was significantly larger than
AUCevening shift (t[55] = 6.13, p = 0.0001) as well as AUCday off
(t[55] = 6.97, p = 0.0001).
3.4. Comparison of cortisol profiles between shifts across
shift groups
Within the P1, the main effect time was significant (F[1.8,
23.4] = 3.52, p = 0.051, h2 = 0.21) while the main effect shift
96 B.M. Kudielka et al. / Biological Psychology 74 (2007) 92–103
Fig. 1. Mean salivary cortisol profiles (S.E.M.) in shift groups P1, P2, C1, and C2 for each shift: (A) morning shift, (B) evening shift, (C) night shift, and (D) day off.
(F[2.4, 30.8] = 2.01, p = 0.14) and interaction time by shift
(F[5.4, 70.8] = 1.79, p = 0.12) did not reach the significance
level (see Fig. 2A). Accordingly, the separate analyses of CAR
and working time samples did not yield significant effects of
shift, though for CAR the main effect of shift approached the
level of significance ( p = 0.073). Within the P2, the main effect
time (F[2.6, 44.8] = 3.33, p = 0.033, h2 = 0.16), main effect
shift (F[2.6, 43.6] = 6.91, p = 0.001, h2 = 0.29) as well as
interaction time by shift (F[6.0, 101.9] = 3.07, p = 0.008,
h2 = 0.15) was significant in the overall analysis, documenting
a reduced cortisol profile during the night shift (see Fig. 2B).
Also, separate analysis for CAR and working time samples
yielded significant shift effects (CAR main effect shift:
p = 0.024; interaction time by shift p = 0.033; working time
samples main effect shift: p = 0.022; interaction time by group:
p = 0.015). Comparison of the day off and first night shift in the
C1 resulted only in a significant main effect time (F[2.3,
29.6] = 7.08, p = 0.002, h2 = 0.35; main effect shift: F[1,
13] = 0.40, p = 0.54; interaction time by shift: F[2.7,
34.5] = 0.08, p = 0.96). Also inclusion of the second night
profile as shown in Fig. 2C rendered the same results (main
effect time: p = 0.017; main effect shift: p = 0.84; interaction
time by shift p = 0.75). Again, separate analyses of CAR and
working time samples did not result in significant shift effects.
For the C2, only the main effect time (F[2.0, 21.8] = 6.04,
p = 0.008, h2 = 0.354) reached significance in the overall
analysis (main effect shift: F[2, 18.2] = 0.14, p = 0.84;
interaction time by shift: F[4.5, 48.5] = 1.03, p = 0.41; see
Fig. 2D). Analysis of CAR did not reveal significant shift
effects, while analysis of working time samples yielded a
significant main effect of shift ( p = 0.038) as well as interaction
time by shift ( p = 0.036).
3.5. Comparison of cortisol profiles in the P2 at first and
second measurement
Two months later, the differences in profiles across shifts as
observed at T1 (see above) has considerably decreased at T2 as
indicated by the finding that this time the main effect of shift
only approached the level of significance (F[2.3, 27.9] = 2.86,
p = 0.067) while the interaction missed significance (F[6.0,
71.9] = 1.16, p = 0.34; see Fig. 3). Also, this time the CAR
analysis resulted in non-significant shift effects. The separate
analysis of the working time samples revealed (only) a
 B.M. Kudielka et al. / Biological Psychology 74 (2007) 92–103
Fig. 2. Mean salivary cortisol profiles (S.E.M.) in morning, evening, and night shifts as well as day off for each shift group: (A) P1, (B) P2, (C) C1, and (D) C2.
Fig. 3. Mean salivary cortisol profiles (S.E.M.) in morning, evening, and
night shifts as well as day off for group P2 at T2 (2 months after T1).
97
significant main effect of shift ( p = 0.01), an effect that should
be primarily attributed to morning shift cortisol levels and not
to cortisol levels during night shifts.
Comparisons of AUCs between first versus second measure-
ment within each shift revealed that only AUCsnight shift changed
significantly across time (AUCsmorning shift t[19] = 0.35, p =
0.73; AUCsevening shifts t[19] = 1.28, p = 0.22; AUCsnight shifts
t[20] = 3.29, p = 0.004; AUCsday offs t[16] = 1.49, p = 0.16).
3.6. Comparison of self-reported questionnaire data
between shift groups
Vital exhaustion differed significantly between shift groups
(F[3] = 3.07, p = 0.032) with the highest score in the P1 and
lowest in the C1 (mean VE scores P1: 10.1  5.8, P2: 7.6  4.8,
C1: 5.3  5.5, C2: 7.5  5.0). In respect to perceived chronic
stress, mean scores on TICS subscales ‘work discontent’
(F[3] = 3.63, p = 0.016) and ‘lack of social recognition’
(F[3] = 2.69, p = 0.051) were significantly different between
groups and showed the same pattern: highest scores were
reported by the P1 (1.60  0.8 and 1.73  0.9) and lowest by the
C1 (1.04  0.4 and 1.15  0.6) while P2 (1.55  0.7 and
98 B.M. Kudielka et al. / Biological Psychology 74 (2007) 92–103

1.57  0.9) and C2 (1.23  0.6 and 1.30  0.5) were in between. Table 2
Self-reported average sleep hours in a 24 h cycle for all shift groups in each shift
The group effect for the TICS subscale ‘social isolation’
approached significance (F[3] = 2.52, p = 0.064) with higher Shift Shift group Mean  S.D. F p
scores in the P2 (1.29  0.9) and C2 (1.28  0.9) than the P1 Morning P1 5.43 h  1.2 3.54 0.035*
(0.77  0.6) and C1 (0.94  0.6). All other mean scores of TICS P2 5.42 h  0.9
subscales did not differ between groups (all F below 1.2, all p C2 6.16 h  1.0
above 0.30). Mean scores of the ERI scale ‘reward’ was highest Evening P1 6.57 h  1.2 5.87 0.005*
in the C1 (47.6  5.4), lowest in the P1 (42.3  7.0) and lay in P2 7.35 h  0.8
between for P2 and C2 (44.0  6.4 and 44.3  9.1), however the C2 7.45 h  0.6
group effect was only marginally significant (ERI reward: Night P1 6.72 h  0.9 6.45 0.003*
F[3] = 2.29, p = 0.084). Mean ‘effort’ scores did not differ P2 5.98 h  1.0
between groups (F[3] = 1.2, p = 0.31). Therefore, a group C1 6.96 h  1.1
difference for ERI (effort–reward imbalance) approached Day off P1 6.98 h  1.2 3.97 0.011*
significance (F[3] = 2.04, p = 0.11). ERI ‘overcommitment’ P2 7.81 h  0.8
was significantly different between groups (ERI OC: F[3] = C1 7.09 h  1.3
C2 7.82 h  0.9
3.34, p = 0.023) showing the following scores P1: 12.14  3.3,
*
P2: 12.24  3.8, C1: 10.17  2.7, and C2: 13.19  3.2. p < 0.05.

3.7. Comparison of self-reported sleep data between shifts
across shift groups
Concerning shift specific average sleep ratings, only two
significant results emerged out of the eight ratings. In the
evening shift, the Jenkins score was highest (indicating lowest
sleep quality) in the C2 (6.61  2.0) and lowest (indicating
highest sleep quality) in the P2 (3.58  3.2) while the P1 scored
in between (5.90  4.6) (F[2] = 4.70, p = 0.013). Also, the one-
item sleep rating during the night shift differed between groups
(F[2] = 3.02, p = 0.055) with better sleep in the C1 (3.87  0.9)
in contrast to the P1 (3.36  1.1) and P2 (3.20  1.0). However,
if these comparisons are considered as multiple comparisons,
the results should only be interpreted as trends (adjusted alpha
for 4 or 8 comparisons is 0.013 and 0.0064, respectively).
Average hours of sleep in a 24 h cycle differed significantly
between shift groups across all shifts as summarized in Table 2.
Reported day specific sleep quality, as measured at sampling
days, differed significantly between shift groups in all shifts but
not at the day off and not between shifts within the P1. Results
are summarized in Table 3 presenting F and p statistics for
comparisons between shift groups within each shift (columns
4–5) as well as for comparisons between shifts within each shift
group separately (columns 6–8). GLM procedures comparing
shift groups during each shift were rerun to test for a significant
impact of day specific sleep quality on cortisol profiles. In none
of the comparisons significant differences could be observed for
the variable day specific sleep quality.
3.8. Comparison of self-reported questionnaire data in the
P2 at T1 and T2
In regard to the psychological scales, the P2 reported a non-
significant increase in effort–reward imbalance (T1: 0.60  0.04
and T2: 0.65  0.04; t[22] = 1.56, p = 0.13) due to a marginal
increase in ‘effort’ (T1: 14.0  3.6, T2: 14.7  2.7; t[22] =
1.50, p = 0.16). For the TICS, only mean scores of two
subscales showed marginally significant differences between T1

Table 3
Self-reported day specific sleep quality
Shift Shift group Mean  S.D. Comparison between Comparison between shifts within each group
groups within each shift
F p Group F p
*
Morning P1 2.57  0.98 4.53 0.015 P1 4.28 0.014*
P2 3.25  0.70
C2 3.22  0.88
Evening P1 3.15  0.99 3.820 0.027* P2 2.50 0.080
P2 3.81  0.63
C2 3.50  0.79
Night P1 3.50  0.58 6.381 0.003* C1 1.50 0.23
P2 3.23  0.91
C1 3.96  0.68
Day off P1 3.23  0.87 1.565 0.21 C2 3.40 0.09
P2 3.59  0.80
C1 3.67  1.09
C2 3.85  0.69
*
p < 0.05.
 B.M. Kudielka et al. / Biological Psychology 74 (2007) 92–103
and T2 with a slight decrease in ‘social overload’ (t[22] = 1.7,
p = 0.10) and ‘excessive demands’ (t[22] = 2.0, p = 0.06) from
T1 to T2. As to the average sleep variables (sleep quality, hours
sleep/24 h) only a marginally significant improvement of average
sleep quality in the morning shift (T1: 2.6  1.0 and T2:
2.8  0.8; t[22] = 2.0, p = 0.06) and a significant worsening of
sleep quality (Jenkins score) in the night shift over time was
reported (T1: 4.1  3.2 and T2: 5.1  4.0; t[20] = 2.7,
p = 0.014). All other comparisons remained non-significant
(all t below 1.2, all p above 0.20). While the day specific sleep
ratings within the P2 differed marginally at T1 (F[2.5,
48.2] = 2.50, p = 0.08, see Table 3), this trend disappeared at
T2 (F[2.8, 49.8] = 0.18, p = 0.89). Also, reported day specific
sleep quality in the night shift differed only marginally between
T1 and T2 in the morning (T1: 3.2  0.7 and T2: 3.6  0.7;
t[22] = 1.7, p = 0.10) and night shift (T1: 3.9  3.2 and T2:
3.6  4.0; t[19] = 1.7, p = 0.11) and did not change over time for
the night shift and day off (both t below 0.2, both p above 0.85)
4. Discussion
The present study investigated free cortisol profiles across
shifts in four different groups of shift workers namely
permanent night and day shift workers (with weekly rotation)
and former permanent night and day shift workers across fast-
rotating morning, evening, night shifts, and days off. Sampling
started directly after awakening in order to account for the
cortisol awakening rise (CAR) which typically takes place
during the first hour after awakening (Pruessner et al., 1997;
Wüst et al., 2000b; for review see Clow et al., 2004). While
there is still an ongoing debate on the psychoneuroendocrine
significance of the CAR, recent studies have indicated
significantly altered CARs in subjects with health problems
(Kudielka and Kirschbaum, 2003) as well as various conditions
related to psychopathology. For example, the CAR was shown
to be positively associated with perceived chronic stress
(Schulz et al., 1998; Pruessner et al., 1999; Wüst et al., 2000a;
Schlotz et al., 2004) and depressive symptomatology (Bhag-
wagar et al., 2003, 2005; Pruessner et al., 2003). On the other
hand, a reduced CAR was, for example, found to be related to
low sleep quality in primary insomnia patients (Backhaus et al.,
2004), burnout symptomatology (Pruessner et al., 1999),
chronic fatigue syndrome (Roberts et al., 2004), posttraumatic
stress disorder (Rohleder et al., 2004; Wessa et al., 2006) and
early loss experience (Meinlschmidt and Heim, 2005).
Furthermore, in a recent sleep laboratory study, we could
demonstrate that the CAR is a genuine response to morning
awakening that is distinct from the circadian rise in HPA
activity in the morning hours (Wilhelm et al., unpublished
data).
A further strength of the present data is that cortisol has been
sampled during each (existing) shift, and that profiles can be
compared between shift workers still engaged in the old rota
(groups C1 and C2: permanent night and day work) and those
who recently switched to the new rota (groups P1 and P2: fast
rotating morning, evening, and night shifts). This comparison
was made possible due to a recent implementation of the new
99
work schedule in pilot test divisions. Additionally, we had the
chance to repeat the cortisol sampling in the P2 (former
permanent day workers) about 3 months after reorganization in
order to investigate the impact of duration of exposure.
Before resuming the study results, it is important to consider
major limitations of the present study. Analyses are mainly
based on cross-sectional data (except P2) not allowing for any
causal conclusions. Unfortunately, due to operational reasons
we could not start data collection before onset of the pilot
period. Furthermore, ambulatory saliva sampling always relies
on the subjects’ motivation and accuracy. Adherence to
sampling procedures and given sampling times is crucial
(Kudielka et al., 2003; Broderick et al., 2004). Unfortunately,
we did not have the chance to control for sampling compliance
as by electronic monitoring. Although 102 subjects entered the
analyses, group size should at best be considered as medium-
sized. Lastly, we wonder whether the psychological assessment
was possibly subjected to bias due to motivational processes.
The implementation of the new shift rota in the test divisions (P
groups) seemed to be a major issue for management and
workforce and subjects were certainly aware of their status as
study participants. This might have impacted significantly on
the questionnaire data (see below).
The main findings regarding the cortisol data are the
following: Firstly, during the morning and evening shift, all
groups showed the typical cortisol increase after awakening and
decreasing levels thereafter. The circadian cortisol profiles did
not differ between fast-rotating former permanent night and day
workers and weekly rotating day workers. It might be
speculated that both former weekly rotating day workers as
well as former permanent night workers adapted relatively
smoothly to the new fast rotating work schedule in respect to
normal cortisol rhythmicity. Secondly, the typical awakening
effect could also be observed during the night shift although
former permanent day workers showed a reduced CAR
compared to both groups of workers which had worked night
duties already in the past. It seems to be striking here that the
P2, the group with no long-lasting night work experience,
showed the most unusual CAR and that these group differences
could not be detected during working time (samples 5–8; see
Fig. 1C). This finding underlines the value of the awakening
samples in contrast to the remainder samples. Within the P2, a
comparison of the cortisol profiles across shifts also certified
that the night shift profile was significantly attenuated (see
Fig. 2B). This effect was not found within the P1 who had
worked night shifts before; their cortisol profiles did not differ
significantly between shifts (see Fig. 2A). It might be
speculated that adaptation to the new rota was most challenging
for the P2 since it included for the first time night duties.
Interestingly, at T2 the attenuation of the night shift profile was
much less pronounced and cortisol profiles across shifts
differed only marginally (see Fig. 3). It appears that the cortisol
profile during night shifts had almost normalized in the P2.
Accordingly, in the above described study by Lac and Chamoux
(2004) greatest disturbance in salivary cortisol was also
observed during night shifts, in particular for the slower rotating
group, resulting in blunted cortisol patterns during night shift
100 B.M. Kudielka et al. / Biological Psychology 74 (2007) 92–103
work. Thirdly, at the day off the C1 (permanent night workers)
showed a similar cortisol profile than during night work;
compared to the other shift groups this profile was slightly
reduced (see Figs. 1D and 2C). The group difference cannot be
ascribed to different times of awakening. This finding might
reflect reduced cortisol synchronization during days off in
permanent night workers. Since this effect did not emerge in
former permanent night workers currently performing fast
rotating shift work (P1), it might be speculated that such altered
circadian adaptation can be readjusted in relatively short time.
However, the present data is cross-sectional and does not allow
for causal inferences. The observation that the C2 had relatively
high morning cortisol levels are a little bit puzzling and difficult
to explain. At least, the typical morning increase emerged and
the profile was not flattened as in the C1. Within the C2 the
overall shift profiles did not differ significantly (see Fig. 2D).
Finally, the present data nicely reflects both major regulatory
inputs on cortisol curves. As outlined above, the two major
processes which control the overall diurnal cortisol variation
are the signal generated by the circadian pacemaker (SCN) and
the alteration of wakefulness and sleep (Van Cauter, 1990;
Goichot et al., 1998; Czeisler and Klerman, 1999). In this study,
timing of the major cortisol rise was significantly determined by
the sleep-wake cycle, namely time of awakening. Nevertheless,
the cortisol secretion differed significantly between shifts
documenting circadian regulation. As expected, overall cortisol
secretory activity was lowest during night shifts compared to
morning and evening shifts and higher during morning
compared to evening shifts and days off.
The questionnaire data is difficult to interpret since
responses might have been biased by motivational factors.
Although the management never claimed that introduction of
the new rota in the entire company would be based on the study
results, on site we gained the impression that at least some
participants might have assumed that their responses could
impact on the company’s decision. Beside others, Smith et al.
(1998) have emphasized the importance of shiftworker
participation in shift rotation redesign. However, the present
questionnaire data is not necessarily biased. First of all, vital
exhaustion, chronic stress, and the effort–reward scales appear
to indicate that the P1 (former night workers now engaged in
fast rotating shift work) had in general the most unfavorable
scores especially in contrast to the group that still works on
permanent night shifts (C1) who reported the most favorable
scores. In accordance with Wilkinson (1992) one might
conclude that permanent night work is preferable to rotating
shifts. However, as in other cross-sectional studies comparisons
between permanent night workers (here C1) and the former day
shift groups might also, at least in part, reflect a process of self-
selection, called the healthy worker effect (Costa, 2003). The
effect describes the tendency that the fitter and healthier
workers stick to night shifts while others who were unable to
cope had already switched to other schedules. Accordingly, in
the Lac and Chamoux (2004) study, shift workers in the slow
rotating rota had the best health and fitness scores among the
three study groups. This reasoning might be also consistent
with our observation that significantly more day shift workers
(P2 and C2) reported health problems. This could (partially)
explain more favorable scores in the C1. Furthermore, former
permanent night workers (P1) reported lowest reward (and
therefore yielded marginally higher effort–reward imbalance)
especially compared to current permanent night workers (C1).
Maybe, our former night workers felt more stressed-out by
reorganization of their working, family, and social life, maybe
because they lost an ‘‘elite’’ status among peers, and probably
disapprove the loss of the night shift premium and other
advantages (e.g., incentives, less control by management during
night, etc.). Costa (2003) argues that ‘‘shift workers sometimes
do not report their health troubles completely, or even deny
them, because they are more afraid of losing the economic
benefits associated with shift and night work’’ and that ‘‘shift
work may be well accepted by people who are able/willing to
use daytime hours or off-duty periods to comply with other
personal/family needs’’. Interestingly, workers in the P1 and C1
had significantly more children than the P2 and C2.
While self-reported average sleep quality did not differ
between groups for most comparisons, average hours of sleep in
a 24 h cycle differed significantly between groups for each shift
(see Table 2). Results of sleep length are relatively plausible
(see Akerstedt and Torsvall, 1981). For the morning shift,
shorter sleep was reported for the two pilot groups who for the
first time were engaged in fast rotating shifts all around the
clock. During evening shifts, the day and former day shift
groups (P2, C2) reported longer sleep times compared to the P1
who just started working day shifts. During night shifts, longer
sleep was reported by the groups with night shift experience
(P1, C1), while during days off this pattern was reversed. Day
specific sleep quality differed significantly between groups in
all shifts (except the day off) and seems to parallel sleep length.
In the morning and evening shift, reported sleep quality was
lowest in the P1 (no day shift experience) and was higher in the
P2 and C2 (experience with rotating morning and evening
shifts). In contrast, during night shifts permanent night workers
had the highest sleep quality, followed by former night workers
(P1). Interestingly, within this group the highest sleep quality
was reported during night shifts. In contrast to awakening time
(Wüst et al., 2000b; Edwards et al., 2001; Kudielka and
Kirschbaum, 2003; Federenko et al., 2004), the CAR was not
shown to be affected by sleep duration in healthy subjects with
regular sleep in earlier studies (for review see Clow et al.,
2004). However, HPA axis regulation can be affected by
experimentally induced total sleep deprivation, sleep restriction
or sleep disturbances (Vgontzas et al., 1999, 2003, 2004).
Based on such findings, researchers should bear in mind that
systematic and pronounced differences in sleep duration across
different shifts might contribute to the cortisol circadian rhythm
in shift workers.
For the P2, we can also report data at T2. The (non-
significant) increase of effort–reward imbalance (due to
marginally higher effort) over time might be attributed to the
additional burden of ongoing night shifts. Furthermore, average
sleep quality for night shifts worsened significantly. Interest-
ingly, even the day specific sleep quality for the night shift (no
change over time) did not correspond with the cortisol results
 B.M. Kudielka et al. / Biological Psychology 74 (2007) 92–103
which indicated normalization at T2. Others also reported on
dissociations between cortisol levels and subjective reports in
shift workers. For example, Axelsson et al. (2003) observed
differences in morning testosterone levels but not total plasma
cortisol in satisfied versus dissatisfied shift workers.
In sum, endocrine results show that the CAR is observable in
day and night shift work. Furthermore, this data appears to
indicate that circadian cortisol profiles are not disturbed in
former night workers who recently switched to a fast rotating
shift rota. Implementation of night work in former day workers
seems to cause initially blunted cortisol profiles that normalize
after a short adjustment period. In permanent night workers,
cortisol profiles are blunted during night shift and day off.
Finally, overall cortisol secretion is generally lower during
night shifts compared to day shifts. Focusing on alterations in
the circadian rhythm of cortisol as one biological approach in
the evaluation of shift work-induced disturbances, we gain the
impression that fast rotating shift work is more favorable
compared to permanent night shift work. This is consistent with
earlier reviews (Knauth, 1997, 1998; Knauth and Hornberger,
2003). However, our subjective sleep data does not corroborate
this conclusion unequivocally. This is somewhat surprising
since convincing evidence exists showing that objective sleep
data measured by polysomnography is interrelated with the
circadian cortisol secretion (Van Cauter, 1990; Späth-Schwalbe
et al., 1992; Vgontzas et al., 2003, 2004). Also, others have
repeatedly found that shift work, and in particular night work, is
characterized by increased subjective and objective sleepiness
(for reviews see Akerstedt, 1990, 1998, 2003). Comparing
slowly and fast-rotating shift work schedules, Fischer et al.
(1997) concluded that reducing the number of consecutive
nights of shift work is favorable for sleep duration and sleep
quality and Smith et al. (1998) found increased day sleep
quality and a general improvement of quality of life. However,
contradictory reports exist (Hakola et al., 1996). To date, there
is still an ongoing debate on an ‘‘optimal’’ shift system and the
question ‘How fast should the night shift rotate?’ remains still
open (Smith et al., 1998; for discussion see Folkard, 1992;
Wedderburn, 1992; Wilkinson, 1992).
Acknowledgements
This study was supported by the participants’ company and
grants of the German Research Foundation (DFG) for BMK
and SW, grant nos. KU1401/4-1 and WU 324/3-3.
References
Akerstedt, T., 1990. Psychological and psychophysiological effects of shift
work. Scandinavian Journal of Work, Environment & Health 1 (16 Suppl.),
67–73.
Akerstedt, T., 1998. Shift work and disturbed sleep/wakefulness. Sleep Med-
icine Reviews 2 (2), 117–128.
Akerstedt, T., 2003. Shift work and disturbed sleep/wakefulness. Occupational
Medicine (London) 53 (2), 89–94.
Akerstedt, T., Torsvall, L., 1981. Shift work. Shift-dependent well-being and
individual differences. Ergonomics 24 (4), 265–273.
101
Appels, A., Höppener, P., Mulder, P., 1987. A questionnaire to assess pre-
monitory symptoms of myocardial infarction. International Journal of
Cardiology 17 (1), 15–24.
Axelsson, J., Akerstedt, T., Kecklund, G., Lindqvist, A., Attefors, R., 2003.
Hormonal changes in satisfied and dissatisfied shift workers across a shift
cycle. Journal of Applied Physiology 95 (5), 2099–2105.
Backhaus, J., Junghanns, K., Hohagen, F., 2004. Sleep disturbances are
correlated with decreased morning awakening salivary cortisol. Psycho-
neuroendocrinology 29 (9), 1184–1191.
Bhagwagar, Z., Hafizi, S., Cowen, P.J., 2003. Increase in concentration of
waking salivary cortisol in recovered patients with depression. The Amer-
ican Journal of Psychiatry 160 (10), 1890–1891.
Bhagwagar, Z., Hafizi, S., Cowen, P.J., 2005. Increased salivary cortisol after
waking in depression. Psychopharmacology (Berlin) 182 (1), 54–57.
Boggild, H., Knutsson, A., 1999. Shift work, risk factors and cardiovascular
disease. Scandinavian Journal of Work, Environment & Health 25 (2),
85–99.
Born, J., Hansen, K., Marshall, L., Molle, M., Fehm, H.L., 1999. Timing the end
of nocturnal sleep. Nature 397 (6714), 29–30.
Broderick, J.E., Arnold, D., Kudielka, B.M., Kirschbaum, C., 2004. Salivary
cortisol sampling compliance: comparison of patients and healthy volun-
teers. Psychoneuroendocrinology 29 (5), 636–650.
Clow, A., Thorn, L., Evans, P., Hucklebridge, F., 2004. The awakening
cortisol response: methodological issues and significance. Stress 7 (1),
29–37.
Costa, G., 2003. Shift work and occupational medicine: an overview. Occupa-
tional Medicine (London) 53 (2), 83–88.
Costa, G., Ghirlanda, G., Tarondi, G., Minors, D., Waterhouse, J., 1994.
Evaluation of a rapidly rotating shift system for tolerance of nurses to
nightwork. International Archives of Occupational and Environmental
Health 65 (5), 305–311.
Czeisler, C.A., Klerman, E.B., 1999. Circadian and sleep-dependent regulation
of hormone release in humans. Recent Progress in Hormone Research 54,
97–130 (Discussion 130–132).
Edwards, S., Evans, P., Hucklebridge, F., Clow, A., 2001. Association between
time of awakening and diurnal cortisol secretory activity. Psychoneuroen-
docrinology 26 (6), 613–622.
Federenko, I., Wüst, S., Hellhammer, D.H., Dechoux, R., Kumsta, R., Kirsch-
baum, C., 2004. Free cortisol awakening responses are influenced by
awakening time. Psychoneuroendocrinology 29 (2), 174–184.
Fischer, F.M., Bruni Ade, C., Berwerth, A., Moreno, C.R., Fernandez Rde, L.,
Riviello, C., 1997. Do weekly and fast-rotating shiftwork schedules differ-
entially affect duration and quality of sleep? International Archives of
Occupational and Environmental Health 69 (5), 354–360.
Folkard, S., 1992. Is there a ‘best compromise’ shift system? Ergonomics 35
(12), 1453–1463 (Discussion 1456–1465).
Folkard, S., Tucker, P., 2003. Shift work, safety and productivity. Occupational
Medicine (London) 53 (2), 95–101.
Fujiwara, K., Tsukishima, E., Kasai, S., Masuchi, A., Tsutsumi, A., Kawakami,
N., Miyake, H., Kishi, R., 2004. Urinary catecholamines and salivary
cortisol on workdays and days off in relation to job strain among female
health care providers. Scandinavian Journal of Work, Environment &
Health 30 (2), 129–138.
Fujiwara, S., Shinkai, S., Kurokawa, Y., Watanabe, T., 1992. The acute effects
of experimental short-term evening and night shifts on human circadian
rhythm: the oral temperature, heart rate, serum cortisol and urinary cate-
cholamines levels. International Archives of Occupational and Environ-
mental Health 63 (6), 409–418.
Goh, V.H., Tong, T.Y., Lim, C.L., Low, E.C., Lee, L.K., 2000. Circadian
disturbances after night-shift work onboard a naval ship. Military Medicine
165 (2), 101–105.
Goichot, B., Weibel, L., Chapotot, F., Gronfier, C., Piquard, F., Brandenberger,
G., 1998. Effect of the shift of the sleep–wake cycle on three robust
endocrine markers of the circadian clock. The American Journal of Phy-
siology 275 (2 Pt 1), E243–E248.
Hakola, T., Harma, M.I., Laitinen, J.T., 1996. Circadian adjustment of men and
women to night work. Scandinavian Journal of Work, Environment &
Health 22 (2), 133–138.
102 B.M. Kudielka et al. / Biological Psychology 74 (2007) 92–103
Hennig, J., Kieferdorf, P., Moritz, C., Huwe, S., Netter, P., 1998. Changes in
cortisol secretion during shiftwork: implications for tolerance to shiftwork?
Ergonomics 41 (5), 610–621.
Holmbäck, U., Forslund, A., Lowden, A., Forslund, J., Akerstedt, T., Lennernäs,
M., Hambraeus, L., Stridsberg, M., 2003. Endocrine responses to nocturnal
eating—possible implications for night work. European Journal of Nutrition
42 (2), 75–83.
Jenkins, C.D., Stanton, B.A., Niemcryk, S.J., Rose, R.M., 1988. A scale for the
estimation of sleep problems in clinical research. Journal of Clinical
Epidemiology 41 (4), 313–321.
Kirschbaum, C., Hellhammer, D.H., 2000. Salivary cortisol. In: Fink, G. (Ed.),
Encyclopedia of Stress, vol. 3. Academic Press, San Diego, pp. 379–383.
Knauth, P., 1997. Changing schedules: shiftwork. Chronobiology International
14 (2), 159–171.
Knauth, P., 1998. Innovative worktime arrangements. Scandinavian Journal of
Work, Environment & Health 3 (24 Suppl.), 13–17.
Knauth, P., Hornberger, S., 2003. Preventive and compensatory measures for
shift workers. Occupational Medicine (London) 53 (2), 109–116.
Knutsson, A., 1989. Shift work and coronary heart disease. Scandinavian
Journal of Social Medicine 44 (Suppl.), 1–36.
Knutsson, A., 2003. Health disorders of shift workers. Occupational Medicine
(London) 53 (2), 103–108.
Knutsson, A., Hallquist, J., Reuterwall, C., Theorell, T., Akerstedt, T., 1999.
Shiftwork and myocardial infarction: a case–control study. Occupational
and Environmental Medicine 56 (1), 46–50.
Kobayashi, F., Furui, H., Akamatsu, Y., Watanabe, T., Horibe, H., 1997.
Changes in psychophysiological functions during night shift in nurses.
Influence of changing from a full-day to a half-day work shift before night
duty. International Archives of Occupational and Environmental Health 69
(2), 83–90.
Kopp, M.S., Falger, P.R., Appels, A., Szedmak, S., 1998. Depressive sympto-
matology and vital exhaustion are differentially related to behavioral risk
factors for coronary artery disease. Psychosomatic Medicine 60 (6), 752–
758.
Kudielka, B.M., Broderick, J.E., Kirschbaum, C., 2003. Compliance with saliva
sampling protocols: electronic monitoring reveals invalid cortisol daytime
profiles in noncompliant subjects. Psychosomatic Medicine 65 (2), 313–
319.
Kudielka, B.M., Kirschbaum, C., 2003. Awakening cortisol responses are
influenced by health status and awakening time but not by menstrual cycle
phase. Psychoneuroendocrinology 28 (1), 35–47.
Lac, G., Chamoux, A., 2003. Elevated salivary cortisol levels as a result of sleep
deprivation in a shift worker. Occupational Medicine (London) 53 (2), 143–
145.
Lac, G., Chamoux, A., 2004. Biological and psychological responses to two
rapid shiftwork schedules. Ergonomics 47 (12), 1339–1349.
Meinlschmidt, G., Heim, C., 2005. Decreased cortisol awakening response
after early loss experience. Psychoneuroendocrinology 30 (6), 568–
576.
Motohashi, Y., 1991. Alteration of circadian rhythm in shift-working ambu-
lance personnel. Japanese Journal of Psychiatry and Neurology 45 (1), 147–
148.
Motohashi, Y., 1992. Alteration of circadian rhythm in shift-working ambu-
lance personnel. Monitoring of salivary cortisol rhythm. Ergonomics 35
(11), 1331–1340.
Munakata, M., Ichi, S., Nunokawa, T., Saito, Y., Ito, N., Fukudo, S., Yoshinaga,
K., 2001. Influence of night shift work on psychologic state and cardio-
vascular and neuroendocrine responses in healthy nurses. Hypertension
Research 24 (1), 25–31.
Pruessner, J.C., Hellhammer, D.H., Kirschbaum, C., 1999. Burnout, perceived
stress, and cortisol responses to awakening. Psychosomatic Medicine 61 (2),
197–204.
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 adre-
nocortical activity. Life Sciences 61 (26), 2539–2549.
Pruessner, M., Hellhammer, D.H., Pruessner, J.C., Lupien, S.J., 2003. Self-
reported depressive symptoms and stress levels in healthy young men:
associations with the cortisol response to awakening. Psychosomatic Med-
icine 65 (1), 92–99.
Regestein, Q.R., Monk, T.H., 1991. Is the poor sleep of shift workers a disorder?
The American Journal of Psychiatry 148 (11), 1487–1493.
Roberts, A.D., Wessely, S., Chalder, T., Papadopoulos, A., Cleare, A.J., 2004.
Salivary cortisol response to awakening in chronic fatigue syndrome.
British Journal of Psychiatry 184, 136–141.
Roden, M., Koller, M., Pirich, K., Vierhapper, H., Waldhauser, F., 1993. The
circadian melatonin and cortisol secretion pattern in permanent night
shift workers. The American Journal of Physiology 265 (1 Pt 2), R261–
R267.
Rohleder, N., Joksimovic, L., Wolf, J.M., Kirschbaum, C., 2004. Hypocorti-
solism and increased glucocorticoid sensitivity of pro-Inflammatory cyto-
kine production in Bosnian war refugees with posttraumatic stress disorder.
Biological Psychiatry 55 (7), 745–751.
Schlotz, W., Hellhammer, J., Schulz, P., Stone, A.A., 2004. Perceived work
overload and chronic worrying predict weekend–weekday differences in
the cortisol awakening response. Psychosomatic Medicine 66 (2), 207–
214.
Schulz, P., Kirschbaum, C., Pruessner, J.C., Hellhammer, D.H., 1998. Increased
free cortisol secretion after awakening in chronically stressed individuals
due to work overload. Stress Medicine 14, 91–97.
Schulz, P., Schlotz, W., Becker, P., 2004. Das Trierer Inventar zum chronischen
Stress (TICS): Manual. Hogrefe, Göttingen.
Shinkai, S., Watanabe, S., Kurokawa, Y., Torii, J., 1993. Salivary cortisol for
monitoring circadian rhythm variation in adrenal activity during shiftwork.
International Archives of Occupational and Environmental Health 64 (7),
499–502.
Siegrist, J., Starke, D., Chandola, T., Godin, I., Marmot, M., Niedhammer, I.,
Peter, R., 2004. The measurement of effort–reward imbalance at
work: European comparisons. Social Science & Medicine 58 (8),
1483–1499.
Smith, P.A., Wright, B.M., Mackey, R.W., Milsop, H.W., Yates, S.C., 1998.
Change from slowly rotating 8-h shifts to rapidly rotating 8-h and 12-h shifts
using participative shift roster design. Scandinavian Journal of Work,
Environment & Health 3 (24 Suppl.), 55–61.
Späth-Schwalbe, E., Scholler, T., Kern, W., Fehm, H.L., Born, J., 1992.
Nocturnal adrenocorticotropin and cortisol secretion depends on sleep
duration and decreases in association with spontaneous awakening in the
morning. The Journal of Clinical Endocrinology and Metabolism 75 (6),
1431–1435.
Spiegel, K., Leproult, R., Van Cauter, E., 1999. Impact of sleep debt
on metabolic and endocrine function. Lancet 354 (9188), 1435–
1439.
Touitou, Y., Motohashi, Y., Reinberg, A., Touitou, C., Bourdeleau, P.,
Bogdan, A., Auzeby, A., 1990. Effect of shift work on the night-time
secretory patterns of melatonin, prolactin, cortisol and testosterone.
European Journal of Applied Physiology and Occupational Physiology
60 (4), 288–292.
Tuchsen, F., Jeppesen, H.J., Bach, E., 1994. Employment status, non-daytime
work and gastric ulcer in men. International Journal of Epidemiology 23 (2),
365–370.
Van Cauter, E., 1990. Diurnal and ultradian rhythms in human endocrine
function: a minireview. Hormone Research 34 (2), 45–53.
Vgontzas, A.N., Mastorakos, G., Bixler, E.O., Kales, A., Gold, P.W., Chrousos,
G.P., 1999. Sleep deprivation effects on the activity of the hypothalamic–
pituitary–adrenal and growth axes: potential clinical implications. Clinical
Endocrinology (Oxford) 51 (2), 205–215.
Vgontzas, A.N., Zoumakis, E., Bixler, E.O., Lin, H.M., Follett, H., Kales, A.,
Chrousos, G.P., 2004. Adverse effects of modest sleep restriction on
sleepiness, performance, and inflammatory cytokines. The Journal of
Clinical Endocrinology and Metabolism 89 (5), 2119–2126.
Vgontzas, A.N., Zoumakis, M., Bixler, E.O., Lin, H.M., Prolo, P., Vela-
Bueno, A., Kales, A., Chrousos, G.P., 2003. Impaired nighttime sleep in
healthy old versus young adults is associated with elevated plasma
interleukin-6 and cortisol levels: physiologic and therapeutic implica-
tions. The Journal of Clinical Endocrinology and Metabolism 88 (5),
2087–2095.
 B.M. Kudielka et al. / Biological Psychology 74 (2007) 92–103
Wedderburn, A.A., 1992. How fast should the night shift rotate? A rejoinder.
Ergonomics 35 (12), 1447–1451.
Weibel, L., Brandenberger, G., 1998. Disturbances in hormonal profiles of night
workers during their usual sleep and work times. Journal of Biological
Rhythms 13 (3), 202–208.
Weibel, L., Follenius, M., Spiegel, K., Ehrhart, J., Brandenberger, G., 1995.
Comparative effect of night and daytime sleep on the 24-h cortisol secretory
profile. Sleep 18 (7), 549–556.
Weibel, L., Spiegel, K., Follenius, M., Ehrhart, J., Brandenberger, G., 1996.
Internal dissociation of the circadian markers of the cortisol rhythm in night
workers. The American Journal of Physiology 270 (4 Pt 1), E608–E613.
103
Wessa, M., Rohleder, N., Kirschbaum, C., Flor, H., 2006. Altered cortisol
awakening response in posttraumatic stress disorder. Psychoneuroendocri-
nology 31 (2), 209–215.
Wilkinson, R.T., 1992. How fast should the night shift rotate? Ergonomics 35
(12), 1425–1446.
Wüst, S., Federenko, I., Hellhammer, D.H., Kirschbaum, C., 2000a. Genetic
factors, perceived chronic stress, and the free cortisol response to awaken-
ing. Psychoneuroendocrinology 25 (7), 707–720.
Wüst, S., Wolf, J., Hellhammer, D.H., Federenko, I., Schommer, N., Kirsch-
baum, C., 2000b. The cortisol awakening response—normal values and
confounds. Noise & Health 2 (7), 79–88.