Digitized Bodies - Virtual Spectacles ProjectBy Jennifer Leonard"We don't know what to make of ourselves," starts Mark Dery in his EscapeVelocity: Cyberculture at the End of the Century, "precisely because we are, morethan ever before, able to remake ourselves."Exploring notions, like Dery's, of the human body in this digital era is curator NinaCzegledy, with the interdisciplinar
Scand J Work Environ Health 2006;32(5):339–348
Do burned-out and work-engaged employees differ in the functioning of thehypothalamic-pituitary-adrenal axis? by Saar Langelaan, MSc,1, 2 Arnold B Bakker, PhD,2, 3 Wilmar B Schaufeli, PhD,1, 2 Willem van Rhenen,MSc,4 Lorenz JP van Doornen, PhD 2, 5 Langelaan S, Bakker AB, Schaufeli WS, van Rhenen W, van Doornen LJP. Do burned-out and work-engagedemployees differ in the functioning of the hypothalamic-pituitary-adrenal axis? Scand J Work Environ Health2006;32(5):339–348.
Objectives The central aim of the present study was to examine differences in the functioning of the
hypothalamic-pituitary-adrenal (HPA) axis between 29 burned-out, 33 work-engaged, and 26 healthy reference
managers, as identified with the Maslach Burnout Inventory-General Survey and the Utrecht Work Engagement
Methods All of the managers were employed in a large Dutch telecommunications company. Salivary cortisol
was sampled on three consecutive workdays and one nonworkday to determine the cortisol awakening response.
Salivary dehydroepiandrosterone-sulfate (DHEAS), a cortisol counterbalancing product of the HPA axis, was
measured on these days 1 hour after managers awakened. The dexamethasone suppression test was used to
investigate the feedback sensitivity of the HPA axis.
Results The morning cortisol levels were higher on the workdays than on the nonworkday, but this effect did
not differ between the three groups. The burned-out, work-engaged, and reference groups did not differ in the
cortisol and DHEAS levels, the slope of the cortisol awakening response, and the cortisol : DHEAS ratio. The
work-engaged group showed a stronger cortisol suppression in response to the dexamethasone suppression test
than the other two groups, the finding suggesting higher feedback sensitivity among work-engaged managers.
Conclusions Burned-out and work-engaged managers only differ marginally in HPA-axis functioning.
Key terms burnout; cortisol awakening response; dehydroepiandrosterone-sulfate; dexamethasone suppression
test; work engagement.
Recent studies have expanded burnout research by fo- opposites of exhaustion and cynicism, respectively (2, cusing on its presumed opposite, work engagement (1– 3). Burnout is a reaction to chronic occupational stress- Burnout is positively associated with illness and dis- ors and insufficient recovery (4, 5) and is characterized ease, including myocardial infarction (7), common cold by exhaustion, cynicism, and reduced professional effi- (8), and type 2 diabetes (9). To the extent that work en- cacy (5). In contrast, work engagement is strongly as- gagement is the opposite of burnout, one may assume sociated with job resources and is defined as a positive, that work engagement is associated with positive health fulfilling state of mind. Work engagement is character- outcomes. Indeed, it has been reported that work-en- ized by vigor (high levels of energy while working, will- gaged employees seem to enjoy good mental and psy- ingness to invest effort in work, and persistence in the chosomatic health (1). Any association between well- face of difficulties), dedication (sense of enthusiasm, being (burnout and work engagement) and physical inspiration, pride, and challenge), and absorption (be- health status is presumably mediated by (stress) physi- ing fully concentrated and happily engrossed in one’s ological mechanisms, the main candidate being the hy- work) (3). Vigor and dedication are the direct positive pothalamic-pituitary-adrenal axis (HPA axis), which Department of Social and Organizational Psychology, Utrecht University, Utrecht, Netherlands.
Research School Psychology and Health, Utrecht University, Utrecht, Netherlands.
Institute of Psychology, Erasmus University Rotterdam, Rotterdam, Netherlands.
ArboNed Utrecht, Utrecht, Netherlands.
Department of Health Psychology, Utrecht University, Utrecht, Netherlands.
Reprint requests to: S Langelaan, Utrecht University, Department of Social & Organizational Psychology, PO Box 80.140,3508 TC Utrecht, Netherlands. [E-mail: email@example.com] Scand J Work Environ Health 2006, vol 32, no 5 regulates the long-term adaptation of organisms to that more attention should be paid to the hypoactivity stress. Consistent with this view, changes in HPA-axis of the HPA axis in stress-related bodily disorders (12, functioning have been observed in many stress-related 26). In short, the overall picture is confusing, and it is disorders, for example, (vital) exhaustion (10), the hard to predict what is to be found in stress-related dis- chronic fatigue syndrome (11), post-traumatic stress dis- orders, either hypo- or hyperfunction of the HPA axis.
Adding to this confusion are the observations that posi- Disruptions in HPA-axis functioning, due to chronic tive well-being (27) and positive affect (28, 29) have stress, are usually examined by investigating the level been associated with lower cortisol levels as well. This and course of cortisol during the day and the cortisol finding would imply that work engagement may be as- awakening response. The cortisol awakening response sociated with lower cortisol levels.
is the acute increase of the cortisol level in the 30 min- However, cortisol levels give only a partial picture utes after awakening. It has shown to be independent of of HPA-axis functioning. To complete the picture, we the general cortisol level and thus provides independent used two other indicators in our present study. First, we information on HPA-axis functioning (13). Because the included the dexamethasone suppression test, which cortisol awakening response reflects the capacity of the provides information about the feedback sensitivity of adrenal cortex to produce cortisol, the measure is con- the HPA axis (22, 30). Cortisol regulates its own level sidered to be an indicator of HPA axis (re)activity (13).
by exerting a negative feedback function on the hypo- A strong cortisol awakening response is generally asso- thalamus and hypophysis and thereby inhibits the syn- ciated with chronic work stress (14–16), the main pre- thesis of corticotropin-releasing hormone (CRH) and adrenocorticotropic hormone (ACTH) and thus also the Unfortunately, so far, researchers have paid little at- synthesis of cortisol by the adrenal cortex. This phenom- tention to the psychophysiological correlates of burn- enon is called the negative feedback loop of the HPA out and work engagement. The studies that did exam- axis. Dexamethasone is a synthetic cortisol and mimics ine HPA-axis functioning in burnout reported inconsis- the negative feedback effect of cortisol. Its application tent results (17–23), and, to date, there are no studies inhibits own-body cortisol synthesis. The extent to that have examined HPA-axis functioning in work en- which cortisol release is inhibited after dexamethasone gagement. The central aim of our study was to examine intake is a measure of the feedback sensitivity of the psychophysiological differences, as measured by HPA- HPA axis. Both studies that have used the dexametha- axis functioning, between employees high in burnout sone suppression test with burned-out persons yielded and their counterparts, high in work engagement.
conflicting results. One reported lowered morning cor- In chronically stressed or burned-out employees who tisol levels in burnout after dexamethasone intake (23), are not on sick leave, elevated cortisol levels during the and the other found no differences between a clinical first hour after awakening (16) and during the workday burnout group and a healthy reference group after dex- (20) have been reported, as well as lowered morning cortisol levels (23). Other studies have shown, however, Second, it is potentially promising to examine an- no deviations in cortisol in burned-out persons (18) and other product of the HPA axis, dehydroepiandrosterone- employees who are confronted with high job strain (24).
sulfate (DHEAS), which is considered to antagonize the In clinical burned-out groups (eg, people on sick leave effects of cortisol. DHEAS is the sulfated form of who received a clinical burnout diagnosis and who at- DHEA and circulates in the blood in relatively large tended psychotherapy for their complaints) the same in- quantities (about 10 times that of cortisol). It has a long consistencies have been found. Both lowered (21, 25) biological half-time (7–10 hours) and displays a weak and elevated cortisol levels during the first hour after circadian rhythm (31). Higher levels of DHEA(S) have awakening (17, 19) have been reported. Furthermore, been shown to be positively associated with positive af- there seemed to be no differences in cortisol levels be- fect (32), positive well-being, and better physical and tween burnout patients and a healthy reference group mental health (33), whereas lower DHEAS levels have during the day (17). The most recent and extensive study been found in depression (34, 35) and in chronic fatigue showed that the cortisol awakening response and the di- syndrome (35–37). It can thus be expected that work urnal cortisol course did not differ between a clinical engagement is positively associated with DHEAS lev- burn-out group and a healthy, matched reference group els and negatively associated with burnout. However, the only study that investigated DHEAS levels in rela- Remarkably, in studies on other stress-related dis- tion to different levels of burnout reported no differ- orders that have the exhaustion component in common ences between a high- and a low-burnout group (18).
with burnout, like vital exhaustion (10) and chronic fa- Due to their antagonistic relationship, the cortisol- tigue syndrome (11), lower cortisol levels have often to-DHEAS ratio is considered a parameter of interest.
been reported. This finding led some scholars to argue A metabolic shift of DHEAS production (androgen) to Scand J Work Environ Health 2006, vol 32, no 5 cortisol production (glucocorticoid) has been found to (MBI-GS) met the following inclusion criteria (41): (i) be associated with illnesses and chronic stress (38, 39) exhaustion ≥2.2 and (ii) either cynicism ≥2.0 or personal and might thus be associated with burnout as well. On accomplishment ≤3.66. Twenty-nine managers fulfilled the other hand, work engagement may be associated these criteria. Recently, the validity of these cut-off with a shift towards androgen production.
scores has been demonstrated (42). The managers who The aim of our study was to combine in one and the scored higher than 4.67 (ie, mean sum of all 17 items) same study the opposite poles of work-related well-be- on the engagement questionnaire (UWES) were as- ing, burnout and engagement, with the opposite roles signed to the work-engaged group (N=33) (40). The ref- of cortisol and DHEAS in the stress physiological realm.
erence group (N=26) consisted of managers with an in-dividual score according to the following inclusion cri-teria: burnout questionnaire (MBI-GS): exhaustion ≤1.5,cynicism ≤1.0 and personal accomplishment >3.66; the Study population and methods
engagement questionnaire (UWES): mean score ≤4.67.
The managers who were willing to participate signed an informed consent at the feedback meeting and re-ceived a package with an instruction letter, the A total of 88 male managers participated in this study salivettes, and a short diary to keep a record of their sa- (table 1). These men were selected from a larger sample liva sampling. They were instructed to conduct the sam- of managers (N=338), employed at a Dutch telecommu- pling the following week and were asked to return the nications company. An extensive periodic employee saliva samples and the diary by surface mail to the re- health and well-being survey was carried out in this searchers. The university ethics committee approved of company by an occupational health service. The survey was sent by surface mail, along with a cover letter, tothe home addresses of 450 managers, of which 338 re-turned the completed survey in a prestamped envelope (response rate 75%). In the cover letter, the managerswere asked to contact the occupational health service Burnout was measured with the Dutch version (41) of to make an appointment with the occupational health the MBI-GS (43). The MBI-GS consists of 15 items and physician, who provided personal feedback on the sur- taps three subscales, namely exhaustion (5 items, for vey. The participants did not receive any monetary re- example, “I feel mentally exhausted because of my ward for participation but were freely offered a general work”; α=0.93), cynicism (four items, for example, “I medical health check about which they also received doubt the significance of my work”; α=0.85), and pro- feedback. During their feedback meeting, the managers fessional efficacy (6 items, for example, “I can effec- were invited to participate in the study if their scores tively solve the problems that arise in my work”; on the Maslach Burnout Inventory-General Survey α=0.83), which are scored on a 7-point scale ranging (MBI-GS) and the Utrecht Work Engagement Scale (UWES) met externally validated criteria (40, 41). The Work engagement was measured with the UWES (3, managers were excluded if they used corticosteroids, 40). The UWES includes 17 items that are indicative of had asthma, diabetes, rheumatoid arthritis, cardiovascu- three dimensions, namely, vigor (6 items, for example, lar disease, metabolic or endocrinological abnormalities, “At work, I feel full of energy”; α=0.88), dedication had a body mass index of >30 kg/m2, or if they used (5 items, for example, “I am enthusiastic about my job”; alcohol and drugs excessively. All of these factors po- α=0.94), and absorption (6 items, for example, “When I am working, I forget everything else around me”; The managers were assigned to the burned-out group α=0.74), which are scored on a 7-point scale ranging when their individual score on the burnout questionnaire Table 1. Demographic variables for the burned-out, engaged, and reference group.
Scand J Work Environ Health 2006, vol 32, no 5 indication of cortisol as a subject characteristic, the val-ues of the two workdays were pooled for the analysis.
Saliva was collected on three consecutive workdays and For each day, the cortisol : DHEAS ratio was calculated one nonworkday. For the cortisol analyses, saliva was as follows: first, the total amount of morning cortisol sampled four times to determine the cortisol awakening within 1 hour after awakening, called area under the response, immediately after awakening and 15, 30, and curve (AUC ground) (45), was calculated; then, the 60 minutes thereafter. The participants had to chew AUC-ground values were divided by the DHEAS val- gently for about 60 seconds on cotton rolls, which were ues. This ratio appeared to have a nonnormal distribu- then put into plastic tubes (Sarstedt; Etten-Leur, Neth- tion and was thus normalized using a log(x+1) trans- erlands). In addition, the managers were instructed to take an oral dose of dexamethasone (0.5 mg) on the sec-ond evening at 2230 to examine the dexamethasone-sup-pressed cortisol levels the next morning. For the Cortisol and dehydroepiandrosterone-sulfate analyses DHEAS analysis, saliva was collected by passive drool First, a repeated-measure analysis with the “within fac- (saliva is put into the plastic tube through a short plas- tors” time (0, 15, 30, and 60 minutes after awakening) tic straw) 1 hour after awakening on the first two work- and day (workday and nonworkday), and the “between days and on the nonworkday. No cotton roles were used factor” group (burned-out, work-engaged, and refer- because they may have caused falsely high DHEAS val- ence) was used to test the cortisol morning level and cortisol awakening response. Second, another repeated- All of the managers were instructed to follow the measures analysis was applied in which people with a time schedule strictly. They were asked to complete the negative cortisol increase in the first 30 minutes after sampling before breakfast and refrain from drinking awakening (AUC increase) (45) were excluded. A nega- coffee or tea and brushing their teeth. A short diary was tive cortisol increase is indicative of noncompliance (46, filled out during the saliva collection. The managers re- 47), usually caused by a delay between the wake-up time ported the time of sampling, sleep quality, perceived and and sampling time (14, 48), the result, in this case, be- expected stress, physical activity, and food, drink and ing the exclusion of 6 managers in the burned-out group, 4 in the work-engaged group, and 4 in the reference The managers stored their samples in a refrigerater group on the workday and 7 burned-out, 13 work-en- (at 4ºC) until they finished their sampling schedule.
gaged and 11 reference managers on the nonworkday.
When finished, they sent the samples back to the re- To test cortisol suppression the day after dexametha- searchers by surface mail. All of the samples were sone intake, a repeated-measures analysis with time as stored in a freezer (–20ºC) until the analyses. Before free the “within factor” and group as the “between factor” cortisol was assayed, the samples were thawed and spun was used. Five managers refrained from dexamethasone at 3000 revolutions/minute for 5 minutes to obtain 0.5– intake, of which one was in the burned-out group, one 1.0 milliliters of clear saliva with low viscosity. Corti- was in the work-engaged group, and three were in the sol was analyzed using an immunoassay (DELFIA) (44).
reference group. When the repeated-measures analysis For the used technique, the precision of the intra- and revealed significant effects, posthoc tests were used to interassay variability was 2.9–7.7% and 6.2–11.5%, re- specify the effects. Finally, multivariate analyses of variance were used to test possible differences betweenthe groups on the workdays and the nonworkday with regard to the DHEAS values and the cortisol : DHEAS A multivariate analysis of variance was used to test whether the three groups differed with respect to demo-graphics and the psychological measures. The cortisoland DHEAS data were checked for missing values and outliers. Per sample point, cortisol, and DHEAS valuesthat exceeded three standard deviations of the mean were excluded from further analysis (22). Missing val-ues and outliers made up 3.2% and 1.6% of the dataset, The correlations between burnout and work engagement respectively. All of the data were checked for skewness, and, because of a nonnormal distribution, a logarithmic The mean scores and standard deviations (SD) of the transformation was applied to the cortisol values on the three groups on the psychological measures are shown day of the dexamethasone suppression and to all of the in table 3. The groups did not differ from each other DHEAS values. Within-person cortisol values can show with regard to demographics, the number of slept hours, some variation over days. To get a more reliable and the time of awakening. As is obvious, the groups Scand J Work Environ Health 2006, vol 32, no 5 differed strongly from each other with regard to all of Table 2. Pearson correlations between burnout and work engage-
the psychological measures, multivariate F (df = 12, ment. All of the correlations are significant at the P<0.01 level.
(MBI-GS = Maslach Burnout Inventory-General Survey, UWES = 162) = 22.65, P<0.001. Subsequent univariate tests showed significant differences on all of the subscalesof the psychological measures. Specifically, and as can be deduced from table 3, Bonferroni posthoc tests showed that the burned-out group scored significantly and substantially higher on exhaustion and cynicism than the two other groups and lower on professional ef- ficacy, vigor, and dedication. The work-engaged group scored higher on all of the work engagement scales than Morning cortisol level and the cortisol awakeningresponse did not differ from each other with respect to morning Figure 1 shows the morning cortisol levels of the three cortisol across days. No significant day × group and groups in the first hour after awakening on the work- time × group interactions were found either. In other day and the nonworkday. A significant main effect was words, neither on workdays nor on nonworkdays did the found for time, F (3, 79) = 49.24, P<0.001, indicating three groups differ from each other with regard to morn- that there was a cortisol awakening response in the first ing cortisol. The day × time interaction was significant, hour after awakening across days. The main effect of F (3, 79) = 3.46, P=0.02, indicating that there was a day was also significant, F (1, 81) = 68.77, P<0.001, steeper slope for the cortisol awakening response on the reflecting that morning cortisol levels were higher on workdays than on the nonworkday. The day × time × group the workdays than on the nonworkday. However, the interaction almost reached significance, F (6, 160) = main group effect (between participants) was not sig- 1.85, P=0.09, indicating a trend for the groups to differ nificant, F (2, 81) = 1.25, P=0.29; thus the three groups in slopes on the workdays and the nonworkday.
Table 3. Psychological measures for the burned-out, engaged, and reference group. (MBI-GS = Maslach Burnout Inventory-General
Survey, UWES = Utrecht Work Engagement Scale, GSQS = Groningen Sleep Quality Scale)
a Burnout and work engagement were measured on 7-point scales ranging from 0 (never) to 6 (everyday).
b The engaged and reference group differed significantly from the burned-out group at the P<0.01 level but did not differ from each other.
c The burned-out, engaged and reference groups differed significantly from each other at the P<0.001 level.
d The engaged group differed significantly from the burned-out and reference groups at the P<0.001 level, but the burned out and reference groups did e Sleep quality ranged from 0 (good sleep quality) to 14 (bad sleep quality).
f Standard deviation in number of hours and minutes (hours:minutes).
Scand J Work Environ Health 2006, vol 32, no 5 Cortisol [nmol/l]
Figure 1. Cortisol awakening response on a
workday and nonworkday, all managers in-
cluded. Note: the burned-out, work-engaged, and reference groups consisted of 29, 33, Time after awakening [min]
Figure 2. Cortisol awakening response on a
workday and nonworkday, excluding the man-
agers with a negative AUC increase. Note: the
groups consisted of 18, 19, and 12 manag- Time after awakening [min]
To substantiate these results, analyses were rerun after the dexamethasone intake. The main effect of time, excluding the managers with a negative AUC increase and the time × group interaction were not significant.
in the first 30 minutes after awakening. The morning Thus, as expected, there was no overall increase in cor- cortisol levels and slopes of the three groups, on the tisol after awakening. The burned-out, work-engaged, workdays and the nonworkday, are shown in figure 2.
and reference managers did not differ from each other Consistent with the previous analysis, a significant main with regard to cortisol increase. However, the group effect was found for time, F (3, 41) = 62.34, P<0.001, main effect was significant [F (2, 76) = 3.43, P=0.04], and day, F (1, 43) = 30.13, P<0.001, indicating that there and therefore the groups showed different levels of cor- was a cortisol awakening response in the first hour af- tisol on this day. Bonferroni posthoc tests revealed that ter awakening across days and that the cortisol levels this effect was due to the lower cortisol levels of the were higher on the workdays than on the nonworkday.
work-engaged employees at 15 and 30 minutes after Again, the main group effect was not significant. In ad- dition, none of the interaction effects was significant.
Thus in contrast to the result of the previous analysis, Dehydroepiandrosterone-sulfate levels and the the increase of the cortisol awakening response was the cortisol : dehydroepinandrosterone sulfate ratio same on the workdays as on the nonworkday.
The means and standard deviations of the three groupsare shown in table 4 for DHEAS. The groups did not Feedback sensitivity: the effect of dexamethasone sup- differ from each other with regard to their DHEAS lev- els, either on the workdays or on the nonworkday [mul- Figure 3 shows the morning cortisol levels of the three tivariate F (df = 6, 158) = 1.02, not significant]. Fur- groups in the first hour after awakening on the day thermore, no significant differences were found between Scand J Work Environ Health 2006, vol 32, no 5 Cortisol [nmol/l]
Figure 3. Cortisol awakening response after
dexamethsaone intake in the burned-out,
work-engaged and reference groups. Note:
groups consisted of 26, 31, and 22 manag- Time after awakening [min]
the three groups with regard to the cortisol : DHEAS ra- Table 4. Means and standard deviations for the burned-out,
tio on any of these days [multivariate F (df = 6, 148) = work-engaged, and reference groups as regards the dehydroepi-androsterone-sulfate (DHEAS) levels.
The overall picture of HPA-axis functioning in stress- related disorders is confusing—both hypo- and hyper-function of the HPA axis has been reported (12, 16, 18–21, 23, 25). The idea behind our study was that study- cortisol levels for employees with a high burnout score.
ing the opposite poles of work-related well-being (burn- However, the assessment of morning cortisol levels in out and work engagement) would create enough con- the study of Melamed et al was only based on one trast to find differences in HPA-axis functioning. Ac- sample, taken at 0800, and it does not provide any in- cording to the results, the burned-out, work-engaged, sight into the awakening response. Higher morning cor- and reference group did, however, not differ with re- tisol levels, but no difference in the increase, have also spect to morning cortisol levels, the cortisol awakening been reported by De Vente et al (17), for burnout pa- response, the DHEAS levels, or the cortisol : DHEAS tients. In addition, Pruessner et al (23) also reported no ratio. Although the morning cortisol levels were higher differences in the cortisol awakening response among on the workdays than on the nonworkday, this effect did employees with a high burnout score, but lowered cor- not differ between the groups. The only difference was tisol levels in the first hour after awakening. It is im- a stronger cortisol suppression in the work-engaged portant to note that the high and low burnout groups in group in response to the low-dose dexamethasone sup- the study of Pruessner et al were defined on the basis of median split of burnout scores and not on the basis of Our results with respect to cortisol morning levels validated cut-off scores (as in our study). Consequently, and the cortisol awakening response in burnout were in the groups in Pruessner et al’s study were less extreme agreement with those of Grossi et al (18, 19), (and showed fewer differences regarding burnout symp- Mommersteeg et al (22), and Steptoe et al (24), who re- toms) than the groups in our study. This finding makes ported no deviations in the cortisol levels of burned-out it all the more remarkable that differences were found.
employees, burnout patients, and people confronted with Lowered cortisol secretion in burnout has also been re- high job strain, respectively. Grossi et al (19) only re- ported by Moch et al (21), but this finding was based ported elevated morning cortisol levels in female burn- on 24-hour urine cortisol collection, and only 16 female out patients on sick leave. Women with a high burnout patients were included. On the basis of our findings and score, but not on sick leave, and men (either patients or those of earlier studies, we conclude that, despite some men with a high burnout score but not on sick leave) isolated differences, there is no convincing evidence for abnormalities in HPA-axis functioning in burnout.
Furthermore, our findings are in contrast with HPA-axis functioning also appeared to be normal Melamed et al (20), who reported elevated morning among work-engaged employees. We did not find lower Scand J Work Environ Health 2006, vol 32, no 5 cortisol levels, as sometimes found in studies that fo- nonworkday. These results do not support earlier find- cused on the association of positive well-being (27) and ings that higher levels of DHEA(S) are positively asso- positive affect (28, 29) with cortisol. In addition, we did ciated with a positive affect (32), positive well-being, not find anomalies in the increase in cortisol in the first and better physical and mental health (33), and nega- hour after awakening. This was the first study to date tively associated with depression (34, 35) and with that included two opposite psychological states, creat- chronic fatigue syndrome (35–37). Nor did Grossi et al ing a strong contrast. Nevertheless, our sensitive design (18) report any deviations in DHEAS in participants did not provide evidence for deviations in HPA-axis with a high burnout score. Therefore, we tend to con- functioning in burnout vis-à-vis work engagement.
clude that there is no change in metabolic balance (cor- The dexamethasone suppression test, however, tisol : DHEAS ratio) in burnout or work engagement.
yielded an interesting finding, a stronger dexamethasone Recently, negative results with regard to HPA-axis suppression of the work-engaged managers. The extent functioning in burnout have been more often reported to which cortisol release is inhibited after dexametha- and could lead to the conclusion that the role of the HPA sone intake indicates feedback sensitivity; thus we can axis in the long-term effects of stress on psychological conclude that work-engaged managers have a more sen- well-being is more complex than initially thought. It has sitive HPA-feedback function. Interestingly, high feed- also recently been concluded that endocrinological and back efficiency is usually observed in concordance with, self-report strain measures do not substitute for each and considered to be a cause of, lower overall cortisol other, but may reflect different underlying processes or levels (49). Indeed the work-engaged group in our study different aspects of stress responses (51). Furthermore, consistently showed the lowest morning cortisol levels, the physiological system is presumably able to compen- although the difference from the other groups was not sate itself on several levels in the axis. To reveal more significant. We observed no deviant cortisol suppression subtle disregulations in the future, more sensitive mea- in the dexamethasone suppression test in the burned-out surement techniques may be needed, like the combined group, and this finding is in agreement with the recent dexamethasone/corticotropin-releasing hormone test or findings of Mommersteeg et al (22). However, tests with infusion of the corticotropin-releasing hor- Pruessner et al (23) has reported that burned-out par- mone and the adrenocorticotropic hormone.
ticipants had lower levels of cortisol after dexametha-sone intake. In this case, the lower cortisol levels after dexamethasone intake may have reflected the lower cor-tisol level they observed in general in the burnout group, A limitation of this study was the relatively small size of the group, the more so after those with a negative In all three groups, the morning cortisol levels were AUC increase were eliminated. The elimination needed higher on the workdays than on the nonworkday, con- to be made to correct for noncompliance. For compli- sistent with findings reported by Schlotz et al (50). Fur- ance to be enhanced, it is advisable to use eDEMTM thermore, at first sight, the increase in cortisol in the first (Aardex Ltd, Zug, Switzerland) electronic monitoring 30 minutes after awakening appeared to be greater on caps, through which exact sampling time can be regis- the workdays than on the nonworkday, consistent with tered (47). A credit to our study was the selection of findings of earlier studies (14, 50). This finding would extreme groups, on the basis of validated burnout and support the interpretation of the cortisol awakening re- work engagement scores. As could be expected, the sponse as an indicator of anticipation to activities of the burned-out, work-engaged, and reference groups dif- workday (14). However, the latter result needs to be in- fered strongly from each other on the burnout and work terpreted with caution because it disappeared when we engagement measures. It should be noted, however, that excluded negative cortisol awakening responses that the mean exhaustion score in our burned-out group was probably indicate noncompliance to the prescribed time 2.9, which corresponded with the scoring anchor “regu- schedule. Remarkably, in the study of Kunz-Ebrecht et larly”. Thus, compared with both of the other groups, al (14), the effect remained present after noncompliers the burned-out group reported relatively more symptoms (people who reported more than 10 minutes delay be- of exhaustion. However, this finding does not necessar- tween wake up time and sampling time) were excluded.
ily mean that the burned-out managers in the current However, in this study, only two morning samples were study experienced extremely high burnout levels in ab- examined, on only one workday, which resulted in a limited reliability considering intraindividual variationbetween days.
We observed no differences between the three groups with regard to their DHEAS levels and the cor- In summary, our study documents that burnout and work tisol : DHEAS ratio, either on the workdays or on the engagement cannot be distinguished with regard to Scand J Work Environ Health 2006, vol 32, no 5 HPA-axis functioning, at least not with the techniques 16. Schulz P, Kirschbaum C, Pruessner JC Hellhammer DH. In- that we used. The basal cortisol levels did not differ be- creased free cortisol secretion after awakening in chronically tween the groups, although the work-engaged employ- stressed individuals due to work overload. Stress Med.
ees showed a better cortisol suppression in response to 17. de Vente W, Olff M, van Amsterdam JGC, Kamphuis JH, dexamethasone, this finding indicating a higher feed- Emmelkamp PMG. Physiological differences between burn- back sensitivity for the HPA axis. Although the latter out patients and healthy controls: blood pressure, heart rate, result needs replication in future studies, our study in- and cortisol responses. Occup Environ Med. 2003;60 suppl dicates that the robustness of the HPA axis as a stress- regulating system has been underestimated.
18. Grossi G, Perski A, Evengard B, Blomkvist V, Orth-Gomer K. Physiological correlates of burnout among women. J Psy-chosom Res. 2003;55:309–16.
19. Grossi G, Perski A, Ekstedt M, Johansson T, Lindström M, Holm K. The morning salivary cortisol response in burnout. J References
20. Melamed S, Ugarten U, Shirom A, Kahana L, Lerman Y, Demerouti E, Bakker AB, de Jonge J, Janssen PPM, Schaufeli Froom P. Chronic burnout, somatic arousal and elevated sali- WB. Burnout and engagement at work as a function of de- vary levels. J Psychosom Res. 1999;46:591–8.
mands and control. Scand J Work Environ Health. 2001; 21. Moch SL, Panz VR, Joffe BI, Havlik I, Moch JD. Longitudi- nal changes in pituitary-adrenal hormones in South African 2. Schaufeli WB, Bakker AB. Job demands, job resources and women with burnout. Endocrine. 2003;21:267–72.
their relationship with burnout and engagement: a multi-sam- 22. Mommersteeg PMC, Heijnen CJ, Verbraak MJPM, van ple study. J Organ Behav. 2004;25:293–315.
Doornen LJP. Clinical burnout is not reflected in the cortisol 3. Schaufeli WB, Salanova M, Gonzalez-Roma V, Bakker AB.
awakening response, the day-curve or the response to a low- The measurement of engagement and burnout: a two sample dose dexamethasone suppression test. Psychoneuroendocri- confirmatory analytic approach. J Happiness Stud. 2002;3:71– 23. Pruessner J, Hellhammer DH, Kirschbaum C. Burnout, per- 4. Halbesleben JRB, Buckley MR. Burnout in organizational ceived stress and salivary cortisol upon awakening. Psycho- 5. Maslach C, Schaufeli WB, Leiter MP. Job burnout. Annu Rev 24. Steptoe A, Cropley M, Griffith J, Kirschbaum C. Job strain and anger expression predict early morning elevations in sali- 6. González-Romá V, Schaufeli WB, Bakker AB, Lloret S. Burn- vary cortisol. Psychosom Med. 2000;62:286–192.
out and work engagement: Independent factors or opposite 25. Mommersteeg PMC, Keijsers GPJ, Heijnen CJ, Verbraak poles? J Vocat Behav. 2006;68:165–74.
MJPM, van Doornen LJP. Cortisol deviations in people with 7. Appels A, Schouten EGW. Burnout as a risk factor for coro- burnout before and after psychotherapy; a pilot study. Health nary heart disease. Behav Med. 1991;17:53–9.
8. Mohren DCL, Swaen GMH, Kant IJ, van Amelsfoort LGPM, 26. Heim C, Ehlert U, Hellhammer DH. The potential role of Borm PJA, Galama JMD. Common infections and the role of hypocortisolism in the pathophysiology of stress-related bodi- burnout in a Dutch working population. J Psychosom Res.
ly disorders. Psychoneuroendocrinology. 2000;25:1–35.
27. Lindfors P, Lundberg U. Is low cortisol release an indicator of 9. Shirom A, Melamed S, Toker S, Berliner S, Shapira I. Burn- positive health? Stress Health. 2002;18:153–60.
out and health review: current knowledge and future research 28. Polk DE, Cohen S, Doyle WJ, Skoner DP, Kirschbaum C.
directions In: Hodgkinson GP, Ford JK, editors. International State and trait affect as predictors of salivary cortisol in healthy review of industrial and organizational psychology; vol 20.
adults. Psychoneuroendocrinology. 2005;30:261–72.
Chichester (UK): Wiley; 2005. p 269–309.
29. Smyth J, Ockenfels C, Porter L, Kirschbaum C, Hellhammer 10. Nicolson N, van Diest R. Salivary cortisol patterns in vital DH, Stone AA. Stressors and mood measured on a momentary exhaustion. J Psychosom Res. 2000;49:335–43.
basis are associated with salivary cortisol secretion. Psycho- 11. Demitrack M. Neuroendocrine correlates of chronic fatigue neuroendocrinology. 1998;23:353–70.
syndrome: a brief review. J Psychiatr Res. 1997;31(1):69–82.
30. Cole MA, Kim PJ, Kalman BA, Spencer RL. Dexamethasone 12. Ehlert U, Gaab J, Heinrichs M. Psychoneuroendocrinological suppression of corticosteroid secretion: evaluation of the site contributions to the etiology of depression, posttraumatic of action by receptor measures and functional studies. Psycho- stress disorder and stress-related bodily disorders: the role of neuroendocrinology. 2000;25:151–67.
the hypothalamus-pituitary-adrenal axis. Biol Psychol.
31. Wolf OT, Kirschbaum C. Actions of dehydroepiandrosterone and its sulphate in the central nervous system: effects on 13. Schmidt-Reinwald A, Pruessner JC, Hellhammer DH, Feder- cognition and emotion in animals and humans. Brain Res enko I, Rohleder N, Schurmeyer TH, et al. The cortisol re- sponse to awakening in relation to different challenge tests 32. McCraty R, Barrios-Choplin B, Rozman D, Atkinson M, Wat- and a 12-hour cortisol rhythm. Life Sci. 1999;64:1653–60.
kins AD. The impact of a new emotional self-management 14. Kunz-Ebrecht SR, Kirschbaum C, Marmot M, Steptoe A.
program on stress, emotions, heart rate variability, DHEA and Differences in cortisol awakening response on work days and cortisol. Integr Physiol Behav Sci. 1998;33:151–70.
weekends in women and men from the Whitehall II cohort.
33. Barrett-Connor E, Khaw KT, Yen SS. A prospective study of Psychoneuroendocrinology. 2004;29:516–28.
dehydroepiandrosterone sulfate, mortality and cardiovascular 15. Lundberg U, Hellström B. Workload and morning salivary disease. N Engl J Med. 1986;315:1519–24.
cortisol in women. Work Stress. 2002;16:356–63.
34. Barrett-Connor E, Von Muhlen D, Laughlin GA, Kripke A.
Scand J Work Environ Health 2006, vol 32, no 5 Endogenous levels of dehydroepiandrosterone sulfate, but not burger CJ. Synthesis of a cortisol-biotin conjugate and evalua- other sex hormones, are associated with depressed mood in tion as a tracer in an immunoassay for salivary cortisol mea- older women: The Rancho Bernardo Study. J Am Geriatr Soc.
surement. J Steroid Biochem Mol Biol. 1992;43:683–92.
45. Pruessner JC, Kirschbaum C, Meinlschmid G, Hellhammer 35. Scott LV, Salahuddin F, Cooney J, Svec F, Dinan TG. Differ- DH. Two formulas for computation of the area under the ences in adrenal steroid profile in chronic fatigue syndrome, curve represent measures of total hormone concentration ver- in depression and in health. J Affect Disord. 1999;54:129–37.
sus time-dependent change. Psychoneuroendocrinology.
36. Kuratsune H, Yamaguti K, Sawada M, Kodate S, Machii T, Kanakura Y, et al. Dehydroepiandrosterone sulfate deficiency 46. Kudielka BM, Broderick JE, Kirschbaum C. Compliance with in chronic fatigue syndrome. Int J Mol Med. 1998;1:143–6.
saliva sampling protocols: electronic monitoring reveals in- 37. Van Rensburg SJ, Potocnik FC, Kiss T, Hugo F, van Zijl P, valid cortisol daytime profiles in noncompliant subjects. Psy- Mansvelt E, et al. Serum concentrations of some metals and steroids in patients with chronic fatigue syndrome with refer- 47. Broderick JE, Arnold D, Kudielka BM, Kirschbaum C. Saliva ence to neurological and cognitive abnormalities. Brain Res cortisol sampling compliance: comparison of patients and healthy volunteers. Psychoneuroendocrinology. 2004;29:636– 38. Hechter O, Grossman A, Chatterton RT Jr. Relationship of dehydroepiandrosterone and cortisol in disease. Med Hypoth- 48. Kupper N, De Geus EJC, Van Den Berg M, Kirschbaum C, Boomsma DI, Willemsen G. Familial influences on basal 39. Parker LN, Levin E, Lifrak E. Evidence for adrenocortical salivary cortisol in an adult population. Psychoneuroendocri- adaptation to severe stress. J Clin Endocrinol Metab.
49. Huizenga NATM, Koper JW, de Lange P, Pols HAP, Stolk 40. Schaufeli WB, Bakker AB. Utrechtse Bevlogenheidschaal: RP, Grobbee DE, et al. Interperson variability but intraperson Voorlopige Handleiding [The Utrecht Work Engagement stability of baseline plasma cortisol concentrations, and its Scale: manual]. Utrecht (Netherlands): Department of Social relation to feedback sensitivity of the hypothalamo-pituitary- & Organizational Psychology; 2003.
adrenal axis to a low dose of dexamethasone in elderly indi- 41. Schaufeli WB, van Dierendonck D. Handleiding van de Utre- viduals. J Clin Endocrinol Metab. 1998;83:47–54.
chtse Burnout Schaal (UBOS) [Manual Utrecht Burnout 50. Schlotz W, Hellhammer J, Schulz P, Stone AA. Perceived Scale]. Lisse (Netherlands): Swets & Zeitlinger; 2000.
work overload and chronic worrying predict weekend-week- 42. Brenninkmeijer V, van Yperen N. How to conduct research day differences in the cortisol awakening response. Psycho- on burnout: advantages and disadvantages of a unidimension- al approach to burnout. Occup Environ Med. 2003;60 suppl 51. Sonnentag S, Fritz C. Endocrinological processes associated with job stress: catecholamine and cortisol responses to acute 43. Schaufeli WB, Leiter MP, Maslach C, Jackson SE. Maslach and chronic stressors. In: Perrewé PL, Ganster DC, editors.
burnout inventory—general survey. In: Maslach C, Jackson Research in organizational stress and well-being: employee SE, Leiter MP, editors. The Maslach burnout inventory–test health, coping, and methodologies. Amsterdam: Elsevier; manual. 3rd ed. Palo Alto (CA): Consulting Psychologists 44. Dressendörfer RA, Kirschbaum C, Rohde W, Stahl F, Stras- Received for publication: 28 November 2005 Scand J Work Environ Health 2006, vol 32, no 5
Publications 1. Cintora-Gonzalez, O., Estournes, C., Guille, J.L., Grob, J.J., Honerlage, B., Lemoigne, J., Levy, R., Lutz, T., Merle, J.C., Muller, D., Richard, M., Rehspringer, J.L., Schell, J., Viart, N. "Aggregates in silica based matrices." Analusis 28 109 - 113, 2000. 2. Felder, D., Guillon, D., Levy, R., Mathis, A., Nicoud, J.F., Nierengarten, J.F., Rehspringer, J.L., Schell,