The cortisol awakening response in patients remitted from depression

The cortisol awakening response in patients remitted from depression

Journal of Psychiatric Research 44 (2010) 1199e1204 Contents lists available at ScienceDirect Journal of Psychiatric Research journal homepage: www...

189KB Sizes 3 Downloads 95 Views

Journal of Psychiatric Research 44 (2010) 1199e1204

Contents lists available at ScienceDirect

Journal of Psychiatric Research journal homepage: www.elsevier.com/locate/psychires

The cortisol awakening response in patients remitted from depression Jean-Michel Aubry a, *, Françoise Jermann b, Marianne Gex-Fabry c, Liliane Bockhorn d, e, Martial Van der Linden f, Nicola Gervasoni g, Gilles Bertschy c, Michel F. Rossier d, e, Guido Bondolfi b a

Geneva University Hospital, Department of Psychiatry, Bipolar Program, 6-8 rue du 31 Décembre, 1207 Geneva, Switzerland Geneva University Hospital, Department of Psychiatry, Depression Program, 6-8 rue du 31 Décembre, 1207 Geneva, Switzerland c Geneva University Hospital, Department of Psychiatry, Division of Adult Psychiatry, 2 ch. Du Petit-Bel-Air, 1225 Chêne-Bourg, Switzerland d Geneva University Hospital, Department of Genetics and Laboratory Medicine, Service of Laboratory Medicine, Switzerland e Geneva University Hospital, Department of Internal Medicine, Service of Endocrinology and Diabetology, Switzerland f Cognitive Psychopathology and Neuropsychology Unit, University of Geneva, Switzerland g Academic Department of Psychiatry, University of Geneva and Clinique la Métairie, Nyon, Switzerland b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 16 December 2009 Received in revised form 14 April 2010 Accepted 14 April 2010

An impressive number of data has been accumulated on dysfunctions of the hypothalamo-pituitaryeadrenal (HPA) axis and cortisol hypersecretion in depression. To assess the dynamic HPA functioning, the cortisol awakening response (CAR) is an easily accessible and reliable approach. Some data suggest that elevated CAR in depressed patients has trait-like characteristics. Therefore we investigated whether patients in remission from a depressive episode have elevated CAR compared to control subjects. CAR of thirty-eight patients in remission from depression (11 men, 27 women, age range 24e66) and 52 control participants were analyzed (18 men, 34 women, age range 24e63). All patients had experienced 3 previous depressive episodes and were off psychotropic medication since at least 3 months. Saliva samples were collected only once, at home, either on weekend or weekday at 0, 15, 30, 45 and 60 min post-awakening. The area under the curve (AUC) above minimum cortisol concentration displayed large interindividual variability (6.4-fold in remitted patients and 8.1-fold in controls, based on 80% range). Investigation of possible variability factors showed that percent explained variance increased from 3.9% when group was considered alone to 8.8%, 12.3% and 19.2% after adjusting for significant effects of weekday vs. weekend, wake-up time and sleep duration, respectively. According to the latter model, AUC was estimated to be 51% higher in remitted patients than in controls (p ¼ 0.007), while a 21% AUC decrease was associated with a 1-h longer sleep duration (p < 0.001). In future studies, detection of between-group differences might benefit from adjusting for sleep duration and other possible confounders. Ó 2010 Elsevier Ltd. All rights reserved.

Keywords: Cortisol awakening response Salivary cortisol Depression Remission HPA system Sleep duration

1. Introduction There is growing evidence that dysfunctions of the hypothalamo-pituitaryeadrenal (HPA) axis in depression are not just epiphenomena of the disorder, but instead, endophenotypes playing a key role in the pathophysiology of depression (Hasler et al., 2004; Oswald et al., 2006). Using the combined dexamethasone/ corticotrophin-releasing-hormone test (DEX/CRH test), it has been demonstrated that in the long-term course of depression, the HPA axis dysfunction increases in parallel with the number of previous episodes (Gervasoni et al., 2004; Hatzinger et al., 2002). Moreover, two studies using the DEX/CRH test with inpatients suggested that persistence of an exaggerated cortisol response, despite clinical

* Corresponding author. Tel.: þ41 22 718 45 11; fax: þ41 22 718 45 99. E-mail address: [email protected] (J.-M. Aubry). 0022-3956/$ e see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.jpsychires.2010.04.015

remission, predicted depression relapse (Zobel et al., 1999, 2001). In outpatients, DEX/CRH data also showed that abnormalities in the HPA axis persisted in patients at high risk of depressive relapse (Appelhof et al., 2006; Aubry et al., 2007; Ising et al., 2007). To investigate HPA axis activity, salivary cortisol measurement has generated a tremendous interest and has the advantage of being less invasive than the DEX/CRH test (Cowen, 2009). Although the assessment of salivary cortisol has been well validated (Kirschbaum and Hellhammer, 2000) and reflects the free hormone fraction that is physiologically relevant, single measures of salivary cortisol are not adequate (Strickland et al., 2002). In contrast, repeated measures immediately after awakening, i.e. the cortisol awakening response (CAR), have attracted considerable interest in recent years. The CAR may be more sensitive to moderate degrees of depression than the DEX/CRH test (Cowen, 2009). Therefore, it could be a more suitable neuroendocrine exploratory tool for depressed outpatients and patients in remission.

1200

J.-M. Aubry et al. / Journal of Psychiatric Research 44 (2010) 1199e1204

The CAR, a discrete part of the cortisol circadian cycle easily measured through saliva samples taken at home, represents the change in cortisol concentration that typically occurs within 20e45 min after waking up in the morning (Hucklebridge et al., 2005; Pruessner et al., 1997b; Wust et al., 2000a). Wilhelm et al. (2007) reported that home assessment of the CAR in saliva provided results similar to those obtained in highly controlled laboratory conditions. They concluded that although the CAR is modulated by circadian influences, it reflects phasic psychophysiological processes specific to the sleep-wake transition. It has been suggested that one role of the CAR is to mobilize energy necessary for the shift from sleeping to being awake (Pruessner et al., 1997b). Another physiological function may be to switch the immune system from nighttime to daytime (Hucklebridge et al., 1998). More recently, it has been demonstrated that the CAR is distinct from the diurnal variations in HPA system and seems to be an additional phenomenon related to some processes specific to sleep-wake transition (Wilhelm et al., 2007). Finally, even though the exact function of the CAR is still unknown, Fries et al. (2009) made the interesting hypothesis that the cortisol rise after awakening may accompany an activation of prospective memory representations at awakening as well as anticipation of demands of the upcoming day. A number of factors likely to influence cortisol levels have been identified over the years. Regarding time of awakening, it has been shown that early awakeners had higher CAR than late awakeners (Edwards et al., 2001; Kudielka and Kirschbaum, 2003). Results about total sleep duration are controversial. According to some studies, the CAR is rather independent of sleep duration (Pruessner et al., 1997b), whereas other studies suggest larger cortisol increase after shorter sleep (Schlotz et al., 2004; Wust et al., 2000a), flattened CAR with longer sleep duration (Vreeburg et al., 2009a) or CAR increase with longer sleep (Griefahn and Robens, 2008; SpathSchwalbe et al., 1992). It has also been reported that the CAR is increased on workdays relative to non-workdays (Kunz-Ebrecht et al., 2004; Mannie et al., 2007) or in case of stress early in the day (Williams et al., 2005), suggesting a modulation of the CAR in dependence of demands of the upcoming day. Using the CAR approach, it has been reported that the morning cortisol increase was elevated in depressed patients (Bhagwagar et al., 2005). Interestingly, the CAR was also elevated in patients who had recovered from depression (Bhagwagar et al., 2003) as well as in young non-depressed people at familial risk of depression (Mannie et al., 2007). In a large cohort study, Vreeburg et al. (2009b) reported that both subjects with current major depressive disorder and in remission from a depressive episode had higher CAR compared to control subjects. The aims of the present study were twofold. First, we investigated whether we could show an increased CAR in patients with recurrent depressive episodes who were in remission and were not taking antidepressant medication. Second, we analyzed the association of sleep parameters with CAR and explored whether detection of between-group differences might benefit from adjustment for such factors since sleep duration had not been taken into account in Bhagwagar et al.’s study (2003). 2. Methods 2.1. Participants Thirty-nine patients in remission from a major depressive episode collected saliva for cortisol measurement. One subject who reported a delay longer than 15 min between wake-up time and first cortisol sample was excluded (see Procedure section). Therefore, the remission group was composed of 38 patients (11 men, 27

women; median age 49, range 24e66). Patients were recruited by press advertisements to participate in a clinical trial investigating depression relapse prophylaxis with mindfulness-based cognitive therapy (MBCT) (Bondolfi et al., 2010). Patients were screened for past depressive episodes by an experienced clinician with the Structured Clinical Interview for DSM-IV (SCID e First et al., 1996). Depressive symptoms were assessed with the MontgomeryeÅsberg Depression Rating Scale (MADRS e Montgomery and Asberg, 1979). In keeping with inclusion criteria of the MBCT study, all patients had experienced at least three previous depressive episodes (median 4, range 3e14), were in remission for at least three months (MADRS scores  13 at screening and three months later, when saliva was collected) and off antidepressant medication since at least three months. Patients received a small financial compensation for their participation. Fifty-two control participants were recruited by word of mouth (control group; 18 males, 34 females; median age 43, range 24e63). Participation was anonymous. Subjects only provided information about age, gender and whether they had experienced a depressive episode in the past (either minor or major). DSM-IV criteria for major and minor depression were listed on the answer form. Subjects were excluded from the control group if a depressive episode was reported. Other psychiatric disorders were not investigated. The study protocol received approval from the ethics committee of the Geneva University Hospitals. Each participant provided written informed consent before being enrolled. 2.2. Procedure Saliva samples were collected using Salivettes (Sarstedt, Nümbrecht, Germany), a device containing a small cotton swab to be placed in the mouth for 45 s. Participants were instructed to collect saliva once, at home, either on weekend or weekday at specific times (awakening, 15, 30, 45 and 60 min post-awakening, 3 PM and 8 PM). In order to avoid sample contamination by food fragments or blood, they were instructed not to eat and brush their teeth during the first hour after they woke up. Participants reported actual sampling times on a timetable, as well as the time they went to bed and the time they woke up. The first sample was collected within 5 min after awakening in 85.6% of the participants. As a delay longer than 15 min between wake-up time and first cortisol sample was shown to significantly affect the CAR (Okun et al., 2010), we excluded one subject who reported a delay longer than 15 min. Samples were kept in the refrigerator until transport to the laboratory by priority mail (delivery on the next day). Prior to the study, we tested the stability of cortisol sampling and found that it was stable for at least one week at 4  C (data not shown). 2.3. Analytical method Upon arrival of the Salivettes at the laboratory, samples were immediately centrifuged (10 min at 3000 rpm), the swab was discarded and saliva was frozen at 20  C until cortisol assay. Upon thawing, samples were centrifuged again and 100 ml saliva aliquots were collected. Cortisol concentration was measured directly (without extraction) using a modified commercial solid-phase radioimmunoassay (Coat a CountÒ, Diagnostic Product Corporation, Los Angeles, CA). Calibrators of the assay were adapted to the lower salivary cortisol concentrations (as compared to serum concentrations) by successive dilution in water. Under these conditions, the analytical detection limit was 0.7 nmol/l and inter-assay imprecision at 7.5 nmol/l was below 10% (coefficient of variation). The whole analytical procedure for cortisol assessment is accredited according to ISO15189 laboratory norms. For data analysis,

J.-M. Aubry et al. / Journal of Psychiatric Research 44 (2010) 1199e1204

1201

on a weekday (median 8.3 h vs. 7.3 h, U ¼ 173, p ¼ 0.004), no difference was observed in remitted patients.

concentrations below detection limit were arbitrarily set to half this value (0.35 nmol/l), an approach often used with single-subject pharmacokinetic data (Beal, 2001).

3.2. Variability of cortisol awakening response 2.4. Statistical analyses

Median cortisol level as a function of time is presented in Fig. 1 for remitted patients and controls. In the remission group, median AUC was 368 min nmol/l and interindividual variability was 6.4-fold when considering 80% range. In the control group, median AUC was 202 min nmol/l, with 8.1-fold interindividual variability. Median delta values were 9.3 (4.4-fold variability) and 6.0 (6.7-fold variability), respectively. Patients in remission displayed significantly elevated AUC (U ¼ 740, p ¼ 0.043) and delta values (U ¼ 701, p ¼ 0.019) when compared with controls (Table 1). Cortisol levels at 3 PM and 8 PM did not differ between remitted patients and controls (median 4.8 vs. 4.5 nmol/l, U ¼ 944, p ¼ 0.84 at 3 PM; 2.2 vs. 2.2 nmol/l, U ¼ 939, p ¼ 0.81 at 8 PM).

The first five cortisol measures (0, 15, 30, 45 and 60 min postawakening) were used to compute two composite parameters. The delta value was calculated as the difference between maximum and minimum concentration. It reflects the maximum cortisol increase over the 1-h post-awakening period. The area under the curve (AUC) above the minimum concentration in the 1-h observation period was calculated according to the trapezoidal rule and taking into account actual sampling times. It provides an estimate of exposure to free cortisol in response to awakening. Comparison between independent groups proceeded with the Fisher’s exact test for categorical variables and the ManneWhitney U-test for continuous variables. AUC and delta values were first investigated for interindividual variability factors using univariate statistics (ManneWhitney U-tests and Spearman’s rank-order correlation coefficients). In a second step, analysis of covariance models (ANCOVA) considered AUC and delta values (log-transformed because of their positively skewed distributions) as dependent variables. Possible predictors included age, bedtime, wake-up time, sleep duration (continuous covariates), gender, weekday vs. weekend and remission vs. control (categorical factors). Statistical significance was set at 0.05 (two-tailed tests). Data analysis was performed using SPSS 17 (SPSS Inc., Chicago, IL).

3.3. Cortisol awakening response: screening for influent variables Screening for possible determinants of interindividual variability proceeded by analyzing AUC and delta values with respect to gender, age, weekday vs. weekend saliva collection, bedtime, wakeup time and sleep duration. AUC (U ¼ 799, p ¼ 0.47) and delta (U ¼ 758, p ¼ 0.28) did not significantly differ according to gender and were not significantly associated with age (Spearman rS ¼ 0.009, p ¼ 0.93 and rS ¼ 0.048, p ¼ 0.65, respectively). As expected, median AUC and delta values were significantly lower on weekends than on weekdays when considering patients and controls together (U ¼ 634, p ¼ 0.013 and U ¼ 605, p ¼ 0.006, respectively). Values in each group are provided in Table 1. Whereas bedtime showed no significant association with CAR (rS ¼ 0.05, p ¼ 0.63 for delta, rS ¼ 0.04, p ¼ 0.71 for AUC), later awakening was significantly associated with lower response (rS ¼ 0.26, p ¼ 0.015 for delta, rS ¼ 0.27, p ¼ 0.010 for AUC). Longer sleep duration, which correlated with later awakening (rS ¼ 0.55, p < 0.001), was also significantly associated with lower cortisol response (rS ¼ 0.35, p ¼ 0.001 for both delta and AUC).

3. Results 3.1. Sample characteristics Remitted patients and controls did not differ with respect to gender (Fisher’s exact test, p ¼ 0.65) but patients were significantly older than controls (median age 49 vs. 43, ManneWhitney U ¼ 744 p ¼ 0.046). As indicated in Table 1, no significant difference between-groups was observed for bedtime (U ¼ 789, p ¼ 0.10), wake-up time (U ¼ 944, p ¼ 0.72), sleep duration (U ¼ 786, p ¼ 0.10) and whether saliva was collected on a weekend or a weekday (Fisher’s exact test, p ¼ 0.37). Whereas controls assessed on a weekend reported significantly later wake-up time and longer sleep duration than those assessed

3.4. Cortisol awakening response as a function of group and sleep-related variables Analysis of covariance models (Table 2) considered AUC and delta as dependent variables and gender, age, bedtime, wake-up

Table 1 Sleep parameters and cortisol secretion (AUC and delta) of patients in remission from a depressive episode and control participants. Remission group (n ¼ 38)

Control group (n ¼ 52) a

a

Weekday (n ¼ 27)

Weekend (n ¼ 11)

U

p-value

Weekday (n ¼ 31)

Weekend (n ¼ 21)

U

Ub

p-value

p-value

Bedtime

Median Range

23:00 21:30e01:25

23:00 22:00e02:30

112

0.24

23:30 21:30e01:00

23:45 21:30e04:30

235

0.089

789

0.10

Wake-up time

Median Range

7:05 3:25e9:45

7:55 5:15e9:45

108

0.19

6:50 4:00e9:00

8:00 6:45e11:22

118

<0.001

944

0.72

Sleep duration (h)

Median Range

8.3 4.7e11.0

7.7 6.0e11.3

138

0.75

7.3 5.0e9.3

8.3 4.5e10.1

173

0.004

786

0.10

AUC (min nmol/l)

Median Range

385 53e956

242 45 - 651

110

0.23

287 31 - 964

166 67 - 747

223

0.057

740

0.043

delta (nmol/l)

Median Range

9.9 1.1e30.7

7.9 1.3e16.2

102

0.13

7.7 1.2e50.8

4.5 2.1e16.8

216

0.040

701

0.019

Abbreviations: AUC, area under the curve above minimum cortisol concentration, over the 1-h post-awakening interval; delta, difference between maximum and minimum cortisol concentration. a Weekday vs. weekend comparison for each group (ManneWhitney U-test). b Remission vs. control comparison (ManneWhitney U-test).

1202

J.-M. Aubry et al. / Journal of Psychiatric Research 44 (2010) 1199e1204

Fig. 1. Median salivary cortisol level as a function of time post-awakening in patients in remission from a depressive episode (n ¼ 38) and control subjects (n ¼ 52). Error bars depict inter-quartile range. Saliva samples were taken in both groups at awakening (time 0) and every 15 min over the next hour, with additional sample at 3 PM and 8 PM. For clarity, symbols are slightly shifted to the right in the control group.

difference (42% higher values on weekdays, p ¼ 0.033; model 2). Whereas bedtime did not significantly contribute to variability (model 3), adjusting for wake-up time (14% lower AUC per 1-h later awakening, p ¼ 0.005; model 4) allowed detecting a significant group effect (p ¼ 0.049). Furthermore, adjusting for sleep duration (21% lower AUC per 1-h longer sleep, p < 0.001; model 5) led to predict a highly significant 51% higher AUC value in remitted patients than in controls (p ¼ 0.007). Percent explained variance increased from 3.9% for a model including group alone to 8.8%, 12.3% and 19.2% when adjusting for weekday vs. weekend, wake-up time and sleep duration, respectively. Thus, sleep duration might be the most relevant factor, even though a major part of interindividual variability remained unexplained. After taking sleep duration into account, weekday vs. weekend and wake-up time did not provide significant additional contribution to the model (model 6). No significant group by sleep duration interaction was detected. When added to the model adjusted for sleep duration (model 5), age and gender did not significantly contribute to explained variability. Similar models were estimated for delta values, with convergent results (Table 2). 4. Discussion

time, sleep duration, weekday vs. weekend and remission vs. control as independent variables, in order to determine whether detection of between-group differences might benefit from adjusting for the other influent variables. When group was considered alone (model 1), the AUC difference between remitted patients and controls was estimated at 35% and did not reach statistical significance (p ¼ 0.063). The group effect remained nonsignificant after taking into account the weekday vs. weekend Table 2 Variability factors influencing cortisol awakening response (analysis of covariance models on log-transformed AUC and delta values, n ¼ 90). Model Variability factor

AUC (min nmol/l)

1

Remission vs. control

1.35

2

Remission vs. control Weekday vs. weekend

3

4

5

6

delta (nmol/l) 2

Multiplying factor

0.063

3.9

1.33

0.060

4.0

1.29

0.102

8.8

1.28

0.099

9.6

1.42

0.033

1.43

0.022

Remission vs. control Bedtime (1-h later)

1.37

0.055

1.36

0.051

1.04

0.566

1.04

0.532

Remission vs. control Wake-up time (1-h later)

1.36

0.049

1.34

0.046

0.86

0.005

0.87

0.004

Remission vs. control Sleep duration (1-h longer)

1.51

0.007

1.49

0.006

0.79

<0.001

0.80

<0.001

Remission vs. control Sleep duration (1-h longer) Weekday vs. weekend Wake-up time (1-h later)

1.45

0.015

1.43

0.013

0.82

0.005

0.82

0.003

1.24

0.206

1.26

0.152

0.98

0.795

0.99

0.851

pvalue

4.2

12.3

19.2

21.3

pvalue

R2 (%)

R (%)

Multiplying factor

4.4

12.7

20.3

22.7

Abbreviations: AUC, area under the curve above minimum cortisol concentration, over the 1-h post-awakening interval; delta, difference between maximum and minimum cortisol concentration; R2 (%), percent explained variance for each model.

The present study confirmed the hypothesis that patients in remission from depression displayed significantly elevated CAR when compared with controls. Furthermore, the difference was emphasized (þ51% in remitted patients vs. controls) after adjusting for sleep duration, which was identified as a highly significant determinant of interindividual variability. Some years ago, Bhagwagar et al. (2003) reported that cortisol secretion after waking was greater in recovered depressed subjects than in controls. They found no correlation between time of awakening and cortisol AUC, but did not take into account bedtime, total sleep duration and whether sampling was performed on a weekday or a weekend. Recently, Vreeburg et al. (2009b) also reported higher CAR in a large cohort study of recovered depressed patients, either without or with adjustment for a number of variables, including time of awakening and average sleep duration during the last 4 weeks (dichotomized as 6 vs >6 h per night). In contrast with the present study, however, 23% of patients were treated with antidepressant medication, which can influence the HPA axis. About half the patients did not have recurrent depressive episodes. Therefore, the present study provides confirmation for the difference between recovered patients (with at least three depressive episodes in our sample) and controls. It also shows that weekday vs. weekend sampling, time of awakening and sleep duration significantly contribute to variability, whereas bedtime does not. The association between longer sleep duration and lower cortisol response is in accordance with some previous studies (Schlotz et al., 2004; Vreeburg et al., 2009a; Wust et al., 2000a). In contrast, other studies found no correlation between CAR and sleep duration (Federenko et al., 2004; Pruessner et al., 1997b) or even a positive correlation when using polysomnography to measure sleep duration precisely (Spath-Schwalbe et al., 1992). More recently, in a pilot study on the effects of shift work, morningness and sleep duration, Griefahn and Robens (2008) similarly reported that CAR increased significantly with total sleep duration (ascertained by means of the polysomnogram) in evening-oriented healthy young men. The association between cortisol secretion and sleep patterns is mediated at least partially by an interaction between the suprachiasmatic nucleus and the HPA system (Van Cauter and Turek, 1995) but the mechanisms by which sleep duration might influence CAR remain partially unclear (review in Buckley and Schatzberg, 2005). The CAR seems to be independent of the mode

J.-M. Aubry et al. / Journal of Psychiatric Research 44 (2010) 1199e1204

of awakening, with an alarm clock or naturally (Pruessner et al., 1997a) and is not linked to a sleep stage (Spath-Schwalbe et al., 1992). Sleep-related parameters are obviously associated with each other and with stress-related factors such as anticipation of wake-up (Born et al., 1999; Fries et al., 2009). An increased CAR has been reported for workdays versus work free weekend days (KunzEbrecht et al., 2004; Mannie et al., 2007) or in anticipation of upcoming demand (Rohleder et al., 2007). In our study, we did not find a difference between weekend and workdays in the remission group. This could be due to the fact that many remitted patients were not as active professionally as the control group and therefore had a lower level of upcoming stress in the morning. Additional studies are needed to further determine the roles of chronic stress, sleep quality (e.g. undisturbed vs. disturbed nights) and individual morning vs. evening orientation, among other factors. Beyond the relationship between CAR and sleep duration, the present study suggested that the detection of differences between remitted patients and control participants might benefit from adjusting for relevant interindividual variability factors, such as sleep duration. Indeed, the estimated between-group AUC difference increased from 35% to 51%, while the percent explained variability increased from 3.9% in the unadjusted model to 19.2% after taking sleep duration into account. The large fraction of unexplained variability nevertheless suggested that a variety of other determinants might be involved. Wust et al. (2000b) provided evidence for a genetic influence on the morning cortisol response to awakening in healthy individuals. Moreover, salivary cortisol response to acute challenge is influenced by genetic factors, as shown in several studies (reviewed in Kudielka et al., 2009). Some limitations of the present study have to be addressed. First, saliva samples were collected on a single day. A single case study with 50 measurement days at 3-day intervals reported considerable day-to-day variability in the CAR, particularly in the dynamic increase (Stalder et al., 2009). Second, the present study relied on self-reported sleep duration that might be biased, in contrast with objective measures obtained from polysomnography. With a sample of adolescents, DeSantis et al. (2010) nevertheless showed that self-reported waketimes were reasonably accurate when compared to objective estimates. Moreover, we did not check for sleep disorders in patients and control subjects. It should be highlighted that sleep duration might be a latent variable for a broad variety of other sleep-related and psychological dimensions. A third limitation relates to uncontrolled variability factors and confounders, such as cigarette smoking (reported by some but not all studies) and alcohol consumption. As an example, a negative association was reported between the amount of alcohol consumed the evening before sampling and the dynamics of the next morning CAR (Stalder et al., 2009). This finding suggests that acute effects of alcohol consumption might differ from chronic effects, for which no or positive associations were reported (Badrick et al., 2008; KunzEbrecht et al., 2004). Regarding socio-demographic factors, our patient group was significantly older than controls. However, several studies found no association between CAR and age (Edwards et al., 2001; Pruessner et al., 1997b; Vreeburg et al., 2009a; Wust et al., 2000a), whereas others reported older age to be associated with lower CAR (Ice, 2005; Kudielka and Kirschbaum, 2003; Lasikiewicz et al., 2008). Thus, older age in patients would be expected to contribute to a decreased difference between remitted patients and controls. The present study indicated that sleep duration might be a significant parameter to control for when documenting CAR differences between patients and comparison subjects. Facing the unmet need for valid and specific biomarkers of the risk of developing affective disorders and experimenting depressive relapse,

1203

large-scale studies are awaited to identify and quantify the multiple factors influencing the HPA axis in general, and the CAR in particular. Role of the funding source This study was supported by a grant of the Swiss National Science Foundation (Grant No 3200BO-108432 to Guido Bondolfi, Gilles Bertschy, Jean-Michel Aubry and Martial Van der Linden). Contributors Jean-Michel Aubry: designed the study and wrote the protocol, managed the literature searches, wrote the manuscript. Françoise Jermann: contributed to the protocol, collected saliva samples and contributed to the manuscript. Marianne Gex-Fabry: undertook the statistical analysis, contributed to the final manuscript. Lilianne Bockhorn: did all the cortisol assays. Martial van der Linden: participated to the design of the study and contributed to the manuscript. Nicola Gervasoni: contributed to the inclusion of patients and control subjects. Gilles Bertschy: contributed to the protocol and inclusion of patients as well as control subjects. Michel Rossier: contributed to the protocol, supervised the salivary cortisol analysis and contributed to the paper. Guido Bondolfi: contributed to the study protocol, contributed to the manuscript. Conflict of interest There is no conflict of interest in this research. Acknowledgment We thank Sandra Ter Pelle for her expert technical support in the preparation of this manuscript and Christiane Gonzalez for technical help with the salivette collection. References Appelhof BC, Huyser J, Verweij M, Brouwer JP, van Dyck R, Fliers E, et al. Glucocorticoids and relapse of major depression (dexamethasone/corticotropinreleasing hormone test in relation to relapse of major depression). Biological Psychiatry 2006;59:696e701. Aubry JM, Gervasoni N, Osiek C, Perret G, Rossier MF, Bertschy G, et al. The DEX/CRH neuroendocrine test and the prediction of depressive relapse in remitted depressed outpatients. Journal of Psychiatric Research 2007;41:290e4. Badrick E, Bobak M, Britton A, Kirschbaum C, Marmot M, Kumari M. The relationship between alcohol consumption and cortisol secretion in an aging cohort. The Journal of Clinical Endocrinology and Metabolism 2008;93:750e7. Beal SL. Ways to fit a PK model with some data below the quantification limit. Journal of Pharmacokinetics and Pharmacodynamics 2001;28:481e504. Bhagwagar Z, Hafizi S, Cowen PJ. Increase in concentration of waking salivary cortisol in recovered patients with depression. The American Journal of Psychiatry 2003;160:1890e1. Bhagwagar Z, Hafizi S, Cowen PJ. Increased salivary cortisol after waking in depression. Psychopharmacology 2005;182:54e7. Bondolfi G, Jermann F, Van der Linden M, Gex-Fabry M, Bizzini L, Rouget BW, et al. Depression relapse prophylaxis with mindfulness-based cognitive therapy: replication and extension in the Swiss health care system. Journal of Affective Disorders 2010;122:224e31. Born J, Hansen K, Marshall L, Molle M, Fehm HL. Timing the end of nocturnal sleep. Nature 1999;397:29e30. Buckley TM, Schatzberg AF. On the interactions of the hypothalamicepituitaryeadrenal (HPA) axis and sleep: normal HPA axis activity and circadian rhythm, exemplary sleep disorders. The Journal of Clinical Endocrinology and Metabolism 2005;90:3106e14. Cowen PJ. Not fade away: the HPA axis and depression. Psychological Medicine; 2009:1e4.

1204

J.-M. Aubry et al. / Journal of Psychiatric Research 44 (2010) 1199e1204

DeSantis AS, Adam EK, Mendelsohn KA, Doane LD. Concordance between selfreported and objective wakeup times in ambulatory salivary cortisol research. International Journal of Behavioral Medicine 2010;17:74e8. Edwards S, Evans P, Hucklebridge F, Clow A. Association between time of awakening and diurnal cortisol secretory activity. Psychoneuroendocrinology 2001;26:613e22. Federenko I, Wust S, Hellhammer DH, Dechoux R, Kumsta R, Kirschbaum C. Free cortisol awakening responses are influenced by awakening time. Psychoneuroendocrinology 2004;29:174e84. First MB, Spitzer RL, Gibbon M, Williams JBW. Structured clinical interview for DSM-IV axis I disorders e patient edition (SCID-I/P, Version 2.0). New York: New York State Psychiatric Institute, Biometrics Research Department; 1996. Fries E, Dettenborn L, Kirschbaum C. The cortisol awakening response (CAR): facts and future directions. International Journal of Psychophysiology 2009;72:67e73. Gervasoni N, Bertschy G, Osiek C, Perret G, Denis R, Golaz J, et al. Cortisol responses to combined dexamethasone/CRH test in outpatients with a major depressive episode. Journal of Psychiatric Research 2004;38:553e7. Griefahn B, Robens S. The cortisol awakening response: a pilot study on the effects of shift work, morningness and sleep duration. Psychoneuroendocrinology 2008;33:981e8. Hasler G, Drevets WC, Manji HK, Charney DS. Discovering endophenotypes for major depression. Neuropsychopharmacology 2004;29:1765e81. Hatzinger M, Hemmeter UM, Baumann K, Brand S, Holsboer-Trachsler E. The combined DEX-CRH test in treatment course and long-term outcome of major depression. Journal of Psychiatric Research 2002;36:287e97. Hucklebridge F, Sen S, Evans PD, Clow A. The relationship between circadian patterns of salivary cortisol and endogenous inhibitor of monoamine oxidase A. Life Sciences 1998;62:2321e8. Hucklebridge F, Hussain T, Evans P, Clow A. The diurnal patterns of the adrenal steroids cortisol and dehydroepiandrosterone (DHEA) in relation to awakening. Psychoneuroendocrinology 2005;30:51e7. Ice GH. Factors influencing cortisol level and slope among community dwelling older adults in Minnesota. Journal of Cross-Cultural Gerontology 2005;20:91e108. Ising M, Horstmann S, Kloiber S, Lucae S, Binder EB, Kern N, et al. Combined dexamethasone/corticotropin releasing hormone test predicts treatment response in major depression e a potential biomarker? Biological Psychiatry 2007;62:47e54. Kirschbaum C, Hellhammer DH. Salivary free cortisol. In: Fink G, editor. Encyclopedia of stress. San Diego: Academic Press; 2000. p. 379e83. Kudielka BM, Kirschbaum C. Awakening cortisol responses are influenced by health status and awakening time but not by menstrual cycle phase. Psychoneuroendocrinology 2003;28:35e47. Kudielka BM, Hellhammer DH, Wust S. Why do we respond so differently? Reviewing determinants of human salivary cortisol responses to challenge. Psychoneuroendocrinology 2009;34:2e18. Kunz-Ebrecht SR, Kirschbaum C, Marmot M, Steptoe A. Differences in cortisol awakening response on work days and weekends in women and men from the Whitehall II cohort. Psychoneuroendocrinology 2004;29:516e28. Lasikiewicz N, Hendrickx H, Talbot D, Dye L. Exploration of basal diurnal salivary cortisol profiles in middle-aged adults: associations with sleep quality and metabolic parameters. Psychoneuroendocrinology 2008;33:143e51. Mannie ZN, Harmer CJ, Cowen PJ. Increased waking salivary cortisol levels in young people at familial risk of depression. The American Journal of Psychiatry 2007;164:617e21. Montgomery SA, Asberg M. A new depression scale designed to be sensitive to change. The British Journal of Psychiatry 1979;134:382e9.

Okun ML, Krafty RT, Buysse DJ, Monk TH, Reynolds 3rd CF, Begley A, et al. What constitutes too long of a delay? Determining the cortisol awakening response (CAR) using self-report and PSG-assessed wake time. Psychoneuroendocrinology 2010;35:460e8. Oswald LM, Zandi P, Nestadt G, Potash JB, Kalaydjian AE, Wand GS. Relationship between cortisol responses to stress and personality. Neuropsychopharmacology 2006;31:1583e91. Pruessner JC, Wolf OT, Hellhammer DH, Buske-Kirschbaum A, von Auer K, Jobst S, et al. Free cortisol levels after awakening: a reliable biological marker for the assessment of adrenocortical activity. Life Sciences 1997a;61:2539e49. Pruessner JC, Gaab J, Hellhammer DH, Lintz D, Schommer N, Kirschbaum C. Increasing correlations between personality traits and cortisol stress responses obtained by data aggregation. Psychoneuroendocrinology 1997b;22:615e25. Rohleder N, Beulen SE, Chen E, Wolf JM, Kirschbaum C. Stress on the dance floor: the cortisol stress response to social-evaluative threat in competitive ballroom dancers. Personality and Social Psychology Bulletin 2007;33:69e84. Schlotz W, Hellhammer J, Schulz P, Stone AA. Perceived work overload and chronic worrying predict weekendeweekday differences in the cortisol awakening response. Psychosomatic Medicine 2004;66:207e14. Spath-Schwalbe E, Scholler T, Kern W, Fehm HL, Born J. 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 1992;75:1431e5. Stalder T, Hucklebridge F, Evans P, Clow A. Use of a single case study design to examine state variation in the cortisol awakening response: relationship with time of awakening. Psychoneuroendocrinology 2009;34:607e14. Strickland PL, Deakin JF, Percival C, Dixon J, Gater RA, Goldberg DP. Bio-social origins of depression in the community. Interactions between social adversity, cortisol and serotonin neurotransmission. The British Journal of Psychiatry 2002;180:168e73. Van Cauter E, Turek FW. Endocrine and other biological rhythms. In: DeGroot LJ, editor. Endocrinology. Philadelphia: W.B. Saunders; 1995. p. 2487e548. Vreeburg SA, Kruijtzer BP, van Pelt J, van Dyck R, DeRijk RH, Hoogendijk WJ, et al. Associations between sociodemographic, sampling and health factors and various salivary cortisol indicators in a large sample without psychopathology. Psychoneuroendocrinology 2009a;34:1109e20. Vreeburg SA, Hoogendijk WJ, van Pelt J, Derijk RH, Verhagen JC, van Dyck R, et al. Major depressive disorder and hypothalamicepituitaryeadrenal axis activity: results from a large cohort study. Archives of General Psychiatry 2009b;66:617e26. Wilhelm I, Born J, Kudielka BM, Schlotz W, Wust S. Is the cortisol awakening rise a response to awakening? Psychoneuroendocrinology 2007;32:358e66. Williams E, Magid K, Steptoe A. The impact of time of waking and concurrent subjective stress on the cortisol response to awakening. Psychoneuroendocrinology 2005;30:139e48. Wust S, Wolf J, Hellhammer DH, Federenko I, Schommer N, Kirschbaum C. The cortisol awakening response e normal values and confounds. Noise & Health 2000a;2:79e88. Wust S, Federenko I, Hellhammer DH, Kirschbaum C. Genetic factors, perceived chronic stress, and the free cortisol response to awakening. Psychoneuroendocrinology 2000b;25:707e20. Zobel AW, Yassouridis A, Frieboes RM, Holsboer F. Prediction of medium-term outcome by cortisol response to the combined dexamethasone-CRH test in patients with remitted depression. The American Journal of Psychiatry 1999;156:949e51. Zobel AW, Nickel T, Sonntag A, Uhr M, Holsboer F, Ising M. Cortisol response in the combined dexamethasone/CRH test as predictor of relapse in patients with remitted depression. A prospective study. Journal of Psychiatric Research 2001;35:83e94.