Comparison of the cortisol awakening response inwomen with shoulder and neck pain and women withfibromyalgia

Roberto Riva a,b,*, Paul Jarle Mork c, Rolf Harald Westgaard d, Ulf Lundberg a,e

aDepartment of Psychology, Stockholm University, Stockholm, SwedenbCentre for Musculoskeletal Research, University of Gavle, Gavle, SwedencDepartment of Human Movement Science, Norwegian University of Science and Technology, Trondheim, NorwaydDepartment of Industrial Economics and Technology Management, Norwegian University of Science and Technology, Trondheim,NorwayeCHESS (Centre for Health Equity Studies), Stockholm University, Stockholm, Sweden

Received 28 February 2011; received in revised form 16 June 2011; accepted 20 June 2011

Psychoneuroendocrinology (2012) 37, 299—306

KEYWORDSSalivary cortisol;Hypothalamic—pituitary—adrenal axis;Regionalmusculoskeletal pain;Widespreadmusculoskeletal pain;Stress;Fibromyalgia syndrome;Cortisol awakeningresponse

Summary Shoulder and neck pain (SNP) and fibromyalgia syndrome (FMS), two musculoskeletalconditions of unknown pathogenesis, share some common features in terms of altered neuroen-docrine responses, pain and stress perception. However, the pain distribution in SNP is localized,whereas in FMS is more widespread. Because regional musculoskeletal pain may represent anintermediate stage along a continuum towards widespread musculoskeletal pain we comparedthe cortisol awakening response (CAR) in women with SNP with the CAR in FMS patients andhealthy controls (HC) in a controlled hospital—hotel setting. The aim of the study was toinvestigate whether SNP is related to a deviant regulation of the hypothalamic—pituitary—adrenal (HPA) axis. Eighteen women with SNP, 29 female FMS patients, and 27 female HCparticipated in the study. Cortisol samples were collected upon awakening, 30 and 60 min later.Questionnaires measuring pain levels, sleeping problems, perceived stress, and psychologicalcharacteristics were administered to the participants. Compared with HC, women with SNP had atendency towards higher cortisol levels, whereas FMS had lower cortisol levels. Adjustment forpotential confounders did not influence the results. Women with SNP and FMS patients reportedmore health complaints, pain, and perceived stress than the HC, but women with SNP were lessaffected than the FMS patients. Women with SNP showed a tendency towards an elevated HPA axisactivity compared with HC. The current findings may indicate that the hypercortisolism inregional musculoskeletal pain represent an intermediate stage towards the development of ahypocortisolism in widespread musculoskeletal pain.# 2011 Elsevier Ltd. All rights reserved.

* Corresponding author at: Stockholm University, Department of Psychology, 106 91 Stockholm, Sweden. Tel.: +46 08 163892;fax: +46 08 159342.

E-mail address: roberto.riva@psychology.su.se (R. Riva).

a va i l a ble at ww w. sc ie nce di r ect . com

j our na l h omepa g e: www.e l se v ie r.c om/l oca te/ psyne ue n

0306-4530/$ — see front matter # 2011 Elsevier Ltd. All rights reserved.

doi:10.1016/j.psyneuen.2011.06.014

300 R. Riva et al.

1 Introduction

The pathogenesis of widespread and regional chronic mus-culoskeletal pain conditions is still unknown. Psychosocialstress seems to play an important role in the onset and in thedevelopment of these disorders (Linton, 2000; van der Windtet al., 2000; McFarlane, 2007). Chronic shoulder and neckpain (SNP) is highly prevalent in the population of the Wes-tern countries (30—50% in the general population), particu-larly in women (Picavet and Schouten, 2003; Larsson et al.,2007; Cote et al., 2008). The pain distribution in women withSNP is localized in the shoulder and neck area, but thediagnostic criteria are relatively vague and several specificand partly overlapping diagnoses exist in clinical practice andepidemiological research (Larsson et al., 2007). Fibromyalgiasyndrome (FMS) is a condition characterized by chronic wide-spread musculoskeletal pain and multiple tender points withreduced pressure pain threshold (Wolfe et al., 1990). Comor-bid symptoms commonly associated with FMS include poorsleep quality, morning stiffness, fatigue, anxiety, depression,and psychosocial stress (Wolfe et al., 1990; Krag et al., 1994).

Whether widespread and regional musculoskeletal painconditions represent a continuum with a common pathogen-esis or are different entities has been debated in the scien-tific community (McCain and Scudds, 1988; Goldenberg,1999; Buskila, 2001). To compare two pain-related condi-tions, such as SNP and FMS, may provide new insight into thepathogenesis and the development from regional to wide-spread pain conditions. Several authors have demonstratedthat mental and psychosocial stress worsen pain in bothconditions (Bansevicius et al., 2001; Van Houdenhove andEgle, 2004). Moreover, altered responses of the hypothala-mic—pituitary—adrenal (HPA) axis have been observed inindividuals with regional as well as widespread pain (McBethet al., 2007; McFarlane, 2007; Turner-Cobb et al., 2010).Dysregulation of the HPA axis may lead to an inadequateadaptation to stress thereby representing a possible riskfactor for the development of stress-related diseases (deKloet et al., 2008). Moreover, some studies have found thatindividuals with a deviant regulation of the secretion ofcortisol, the end-product of the HPA axis, have an increasedrisk for chronic pain development (Blackburn-Munro, 2004;McBeth et al., 2007). Both abnormal increase and decrease incortisol secretion may lead to pathology (Miller et al., 2007).Hypocortisolism (i.e., low cortisol levels) has been associatedwith the pathogenesis of widespread and regional chronicmusculoskeletal pain conditions, e.g., chronic pelvic pain(Heim et al., 1998) and FMS (Heim et al., 2000). However,findings for FMS are inconsistent also showing higher ornormal cortisol levels compared with healthy controls (Taket al., 2011). In contrast, several authors have found thatwomen with SNP and healthy controls (HC) had comparablecortisol levels in saliva (Sjors et al., 2010), in blood (Theorellet al., 2000; Nilsen et al., 2007), and in urine (Larsson et al.,2008). However, women with SNP showed a tendency towardshigher serum cortisol levels than HC after a prolonged andstressful reaction-time test (Nilsen et al., 2007).

Dysregulation of the HPA axis can be investigated with theassessment of the cortisol awakening response (CAR). Thepeak level of cortisol in healthy individuals generally occursabout 30 min after awakening and deviating responses from

this pattern may indicate altered HPA axis regulation (Wustet al., 2000; Clow et al., 2004; Federenko et al., 2004).

There are still few studies in the literature that havesystematically investigated salivary CAR under controlled con-ditions in groups with such pain conditions. The aim of thisstudy was therefore to compare women with SNP with FMSpatients and HC in terms of free salivary cortisol levels duringthe morning, self-reported pain, sleep disturbances, and psy-chological features. In a previous study, the comparisonbetween FMS patients and HC of the same study sampleshowed that FMS patients presented an attenuated HPA axisactivity with low CAR (Riva et al., 2010). However, in thecurrent study we added new data on women with SNP becauseSNP might represent a preliminary stage towards the devel-opment of FMS, as it has been already hypothesized for chroniclocal and widespread musculoskeletal pain conditions (McE-wen, 1998; Rosmond and Bjorntorp, 2000; Van Houdenhoveand Egle, 2004; Miller et al., 2007). We therefore hypothesizethat the HPA axis in women with SNP is dysregulated andcharacterized by elevated cortisol secretion. Moreover, wehypothesize that women with SNP show more self-reportedpain, sleep disturbances, perceived stress, and health com-plaints than HC, but less than FMS patients.

2 Methods

2.1 Participants

Eighteen women with SNP, 29 female FMS patients and 27female HC participated in the study (Table 1). Women withSNP were recruited among staff at the University and at asecondary school, FMS patients were mainly recruitedthrough the local fibromyalgia association in Trondheim (Nor-way), and the HC were recruited among donors to thehospital blood bank. Eligible women with SNP were includedif they scored 3 or more on a scale from 0 to 6 representingfrequency and intensity of shoulder and neck pain during thelast 6 months. HC included in the study scored below 3 on thesame scale. The cut-off score of 3 was in a previous studyassociated with a 45% chance of seeking medical consultationdue to shoulder and neck symptoms (Westgaard and Jansen,1992; Jensen et al., 1993; Mork and Westgaard, 2006).Inclusion criteria for eligible FMS patients were a verifieddiagnosis of FMS as defined by the American College ofRheumatology (Wolfe et al., 1990). Number of tender points(M = 15.7; SD = 5.9; range 11—18), number of years since firstsymptoms (M = 13.1; SD = 8.6; range 3—36), and number ofyears with confirmed diagnosis (M = 5.5; SD = 6.0; range 0—26) was retrieved from each FMS patient’s medical record.Participants were excluded if they had: (a) cardiorespiratory,cerebrovascular, neurologic, neuromuscular, endocrine,infectious, metabolic, lung, or cancer disease, (b) injurythat affected function, (c) connective tissue disorder, (d)tendinitis or capsular affection of the shoulder joint, (e), highblood-pressure (i.e., systolic pressure > 140 mmHg or dia-stolic pressure > 90 mmHg) or were taking anti-hypertensivemedication. Participants were also excluded if they weretaking medication that may interact with neural, vascular, ormuscular function or the physiological measurements to beperformed (e.g. antidepressants, antiepileptics, b-block-ers). Participants that used analgesics and/or sleep medicine

Table 1 Descriptive statistics for demographic and other background variables of the women with SNP (N = 18), the FMS patients(N = 29), and the HC (N = 27).

SNP FMS patients HCMean (SD), range Mean (SD), range Mean (SD), range

Age (years) 44.3 (10.7), 23—59 52.1 (8.9), 38—66 51.9 (8.2), 37—64Body mass index (kg/m2) 25.7 (3.9), 19.5—32.4 27.1 (5.9), 17.8—40.0 24.9 (3.3), 20.5—35.0Self-reported sleep duration (h) 5.8 (1.0), 4—8 4.7 (1.2), 2.5—7 6.0 (1.1), 3—7.5

SNP FMS patients HC% (n) % (n) % (n)

Employment fraction �50% 84 (15) 24 (7) 93 (25)Smokers 28 (5) 24 (7) 22 (6)Exercise (>1 session per week) 94 (17) 93 (27) 89 (24)

Comparison of the cortisol awakening response in women with shoulder and neck pain and women with fibromyalgia 301

on a regular basis were instructed to cease medication 2 daysprior to the experiment.

The study protocol was approved by the Regional EthicsCommittees in Norway (project no. 4.2005.2728) and Sweden(Dnr 2006/87-31/1), and all participants signed an informedconsent before inclusion. The study was carried out accord-ing to the declaration of Helsinki.

2.2 Procedure

The evening and the night before saliva sampling the partici-pants stayed in a hospital hotel where they had an eveningmeal (bread, salad, fruits) at around 20:30 h and thereafterwere free to choose activity (e.g., reading, watching TV,playing cards) but were instructed to stay inside the hospitalhotel. They also completed a set of questionnaires beforegoing to sleep in a comfortable hotel room. Saliva samplesto assess CAR were collected on three different occasionsduring the morning of the following day: upon awakening(mean time 06:35 h � 00:24 h), 30 and 60 min after awaken-ing. Participants were instructed to take the Salivette (Sar-stedt, Rommelsdorf, Germany) cotton swab out of the testtube and keep it in the mouth for about 2 min. Thereafter theyplaced the swab back in the test tube (alternatively, they couldspit directly in the test tube). In case of dry mouth participantswere instructed to bite their tongue to stimulate salivarysecretion. The samples 30 and 60 min after awakening weredelivered with the experimenter present, while for the sampleupon awakening the participants were instructed to fill in aform with information on the exact time of the sample. Theparticipants were instructed not to brush their teeth, eat (theyhad to also avoid chewing gum or other sweets), drink coffee,or smoke the last 15 min before each sample. They wereallowed to drink only water.

All samples were frozen after collection. The assay used forthe determination of cortisol in saliva was a competitive radio-immunoassay (Spectria Cortisol Coated tube RIA) purchasedfrom Orion Diagnostica, Espoo, Finland. All samples were ana-lysed twice and in randomized order: intra-assay precision < 5%(1.7—4.1%), inter-assay precision < 10% (4.3—9.0%).

2.3 Questionnaires

A set of questionnaires was administered to the participants toassess subjective health complaints, psychological features,

and perceived stress. The Subjective Health Complaints (SHC)inventory consists of 29 questions concerning subjectivesomatic and psychological complaints during the last 30 days(Eriksen et al., 1999). For each item, severity of the complaintis rated on a 4-point scale (0 = none, 3 = severe) and durationby number of days during the last 30 days. An overall score ofsubjective health complaints is determined by adding up thescores for the individual items. The Karolinska Scales of Per-sonality (KSP) is a self-rating questionnaire constructed tomeasure stable personality traits (Schalling and Edman,1987). KSP consists of 135 items grouped into 15 subscales.Three subscales were used in this study: ‘‘muscular tension’’(difficulties in relaxing), ‘‘psychic anxiety’’ (worrying andfeeling socially insecure), and ‘‘somatic anxiety’’ (autonomousdisturbances such as heart pounding, sweating). The PerceivedStress Scale (PSS) is a 14-item questionnaire designed to assessthe extent of perceived stress in various life situations (Cohenet al., 1983). The subscale Neuroticism of the Eysenck Person-ality Questionnaire (EPQ-N) assesses the general tendency toover-responsiveness or over-reactivity (Eysenck and Eysenck,1975). High scores on the EPQ-N are typical for individuals withanxiety, depression, sleeping problems, and various psychoso-matic disorders. Total scores and subscale scores were calcu-lated for all questionnaires. All the questionnairesadministered to the participants had a good internal consis-tency (Cronbach’s alpha coefficients ranging from .86 to .96).Also after splitting the sample and analysing the data sepa-rately for the three groups all the questionnaires had a goodinternal consistency (Cronbach’s alpha coefficients rangingfrom .61 to .91 for women with SNP). Participants alsoresponded to a short questionnaire concerning self-reportedlevel of sensory pain in neck/shoulder and low back during thelast 24 h, last week, and last 6 months on a visual analoguescale (VAS, 0—100 mm), self-reported sleep duration (i.e.,hours slept during the previous night), smoking habit, employ-ment fraction (i.e., percentage of hours worked per weekcompared to the standard full-time employment of 35—40 hper week), and the number of physical exercise sessions theyusually carry out during a week (Table 1).

2.4 Statistical analysis

Cortisol levels had a positively skewed distribution and,therefore, were logarithmically transformed. The logarith-mic transformation reduced the skewness of the distribution,

Fig. 1 Salivary cortisol levels (anti-log values) upon awaken-ing, 30 and 60 min later in women with SNP (N = 18), FMS patients(N = 29), and HC (N = 27). Error bars represent 95% CI. Asterisksindicate differences between groups with single one-way ANO-VAs on the three morning cortisol samples. *p < .05 **p < .01***p < .001.

302 R. Riva et al.

with Shapiro—Wilk values above the significance level( p < .05) for all the cortisol values except for the cortisolvalues of FMS patients 30 min after awakening ( p = .013).The statistical analysis of this variable was calculated on thelog10 scale, but the figures show the anti-log mean values foran easier interpretation of the results. To obtain indices ofthe cortisol awakening response the area under the curvewith respect to ground (AUCG) was calculated (Pruessneret al., 2003). A repeated measures mixed within-betweenanalysis of variance (ANOVA) was used on the three morningcortisol samples (i.e., CAR) with Greenhouse—Geisser cor-rection of degrees of freedom if the sphericity assumptionwas violated. Additionally, a one-way analysis of variance(ANOVA) on each of the three morning cortisol samples and onthe AUCG was conducted. Bonferroni post-hoc test and Dun-nett’s t test were used to assess differences between thethree groups of participants. Dunnett’s t test is specificallydesigned to compare each different group with one referencegroup. Observed power and partial h2 were calculated asindicators of the effect sizes. To test whether the varianceswere adequately homoscedastic, Levene’s test was per-formed for the cortisol values upon awakening ( p = .192),30 min after awakening ( p = .476), and 60 min after awaken-ing ( p = .515). Pearson’s r and Spearman’s r values were usedto discern possible correlations between the variables. Cor-relations were calculated separately for women with SNP,FMS patients, and HC when there was a significant differencebetween the groups. The two methods showed the sameresults and we choose to only report correlations based onthe Pearson’s r. Relevant confounders that correlated withthe dependent variables were introduced as covariates in theANOVAs. Cronbach’s alpha of the questionnaires was calcu-lated as a measure of internal consistency.

3 Results

3.1 Cortisol levels

Fig. 1 shows the cortisol levels upon awakening, 30 and60 min after awakening in women with SNP, FMS patients,and HC. A repeated measures mixed between-within ANOVAfor the CAR (logged data) showed a significant group effect(F(2,62) = 9.05; p < .001; partial h2 = .23; power = .97), asignificant effect of time on the morning cortisol(F(2,124) = 53.96; p < .001; partial h2 = .47; power = 1),and a significant time by group interaction(F(4,124) = 2.89; p = .044; partial h2 = .09; power = .65).Post-hoc tests showed that women with SNP significantlydiffer from the FMS patients ( p < .001 with Bonferroni andDunnett’s t), but not from the HC ( p = .456 with Bonferroni;p = .228 with Dunnett’s t). Moreover, FMS patients signifi-cantly differed from HC ( p < .05 with both tests).

Additionally, one-way between-subjects ANOVAs wereconducted separately on the three cortisol samples and onthe AUCG (logged data). There was a statistically significantgroup effect for the cortisol levels upon awakening(F(2,70) = 9.23; p < .001; partial h2 = .21; power = .97),30 min after awakening (F(2,66) = 5.49; p = .006; partialh2 = .14; power = .83), 60 min after awakening(F(2,66) = 3.25; p = .045; partial h2 = .09; power = .60), andfor the AUCG (F(2,62) = 8.31; p = .001; partial h2 = .21;

power = .96). Post-hoc test showed that upon awakeningwomen with SNP significantly differed from FMS patients( p = .003 with Bonferroni; p = .002 with Dunnett’s t), butnot from HC ( p = 1 with Bonferroni; p = .993 with Dunnett’st), and that FMS patients significantly differed from HC( p < .01 with both tests). At 30 min after awakening, womenwith SNP significantly differed from FMS patients ( p = .005with Bonferroni; p = .003 with Dunnett’s t). The differencebetween women with SNP and HC at 30 min after awakeningwas not significant using Bonferroni post-hoc test ( p = .077),but significant with Dunnett’s t test ( p = .046). Additionally,the difference between FMS patients and HC at 30 min afterawakening was not significant with both tests. At 60 min afterawakening women with SNP significantly differed from FMSpatients ( p = .041 with Bonferroni; p = .025 with Dunnett’st), but not from HC ( p = .289 with Bonferroni; p = .164 withDunnett’s t), and FMS did not significantly differ from HC. Inaddition, the AUCG of women with SNP significantly differedfrom the AUCG of FMS patients ( p < .001 with Bonferroni andDunnett’s t), but not from the AUCG of the HC ( p = .249 withBonferroni; p = .143 with Dunnett’s t). Moreover, FMSpatients significantly differed from HC using Dunnett’s t( p < .05), but did not using Bonferroni.

3.2 Analysis of confounders

Age, body mass index (BMI), number of cigarettes smokedduring the day, employment fraction, awakening time, timebetween awakening and the first morning salivary sample,hours slept during the last night, and average number ofexercise sessions during a week were checked as possibleconfounders for the AUCG in the comparison of the threegroups. Pearson’s correlations were calculated separately forthe variables that showed significant differences with one-way between-subjects ANOVAs (i.e., age, employment frac-tion, awakening time, and hours slept during the last night).We found a weak positive correlation for the average numberof exercise sessions versus the AUCG (Pearson’s r = .292;N = 65; p = .018) and a weak negative correlation for age

Fig. 2 Bar plot showing the mean values in FMS patients (N = 29), women with SNP (N = 18), and HC (N = 27) for the total score on theSubjective Health Complaints (SHC) inventory, the total score on the Karolinska Scales of Personality (KSP) and its three subscales, thePerceived Stress Scale (PSS) and the Neuroticism subscale of the Eysenck Personality Questionnaire (EPQ-N). Error bars represent 95%CI. *p < .05 **p < .01 ***p < .001.

Comparison of the cortisol awakening response in women with shoulder and neck pain and women with fibromyalgia 303

versus AUCG (Pearson’s r = �.258; N = 65; p = .038), but notfor the other potential confounders. The between groupfactor for the AUCG remained significant after adjustingfor the average number of exercise sessions(F(2,61) = 8.61; p = .001; partial h2 = .22; power = .96) andage (F(2,61) = 6.69; p = .002; partial h2 = .18; power = .90).

3.3 Self-ratings

One-way between-subjects ANOVAs performed on the totalscores of the Subjective Health Complaints (SHC), the Kar-olinska Scales of Personality (KSP) and its subscales, thePerceived Stress Scale (PSS) and the Eysenck PersonalityQuestionnaire-Neuroticism (EPQ-N), showed significant dif-ferences between the three groups (Fig. 2). We used Bon-ferroni post-hoc test and Dunnett’s t test to investigatepossible differences between women with SNP, the FMSpatients and the HC. The two post-hoc tests showed thesame results. Women with SNP reported more subjectivehealth complaints than the HC, but less than the FMSpatients. Women with SNP reported significantly more painin the neck/shoulder during the last 24 h ( p < .001), the lastweek ( p < .001), and the last 6 months ( p < .001) than theHC, but no differences were found between women with SNPand FMS patients. In contrast, women with SNP reported nodifferences in low back pain compared with the HC. However,women with SNP reported significantly less pain in the lowback during the last 24 h ( p < .05), the last week ( p < .01),and the last 6 months ( p < .01) than FMS patients. In addi-tion, women with SNP reported no difference in sleep dura-tion compared with HC. However, women with SNP reportedlonger sleep duration ( p < .01) than FMS patients. Finally,women with SNP reported no difference in the employmentfraction compared with HC. However, women with SNPreported to work more hours per week ( p < .001) thanFMS patients.

For the self-reported variables the correlations wereseparately calculated for the three groups. We found asignificant positive correlation for the SHC versus AUCG(Pearson’s r = .51; N = 18; p = .031) among the controls,and a significant negative correlation for the PSS versus AUCG

(Pearson’s r = �.48; N = 19; p = .038) among FMS patients,but no significant correlations were found among women withSNP. In addition, no significant correlations were foundbetween AUCG and self-reported pain or sleep parameters.

4 Discussion

In the present study we found that the CAR was higher inwomen with SNP than in HC and in FMS patients. Significantdifferences between SNP and FMS were found during all thethree morning samples. A tendency towards significant dif-ference between women with SNP and HC was found only30 min after awakening, but not for the other time points.The analysis of the CAR showed that women with SNP had ahigher AUCG than FMS patients and HC but the difference wassignificant only between the first two groups. However, allthree groups showed a normal CAR time course (i.e., peakcortisol level 30 min after awakening). Reduced but normalCAR time course has been demonstrated in other chronic painconditions, such as chronic fatigue syndrome (Roberts et al.,2004), myogenous facial pain (Galli et al., 2009), and chronicpelvic pain (Heim et al., 1998), whereas the CAR has beenfound absent in patients with hippocampal damage (Bucha-nan et al., 2004) and with global amnesia (Wolf et al., 2005).

The results of the present study on the cortisol levels inwomen with SNP do not support findings on hypocortisolism inother regional chronic musculoskeletal conditions (Heimet al., 2000; Turner-Cobb et al., 2010), but rather show atrend towards an increased cortisol levels in this group ofwomen. Moreover, women with SNP reported more healthcomplaints than the HC particularly in self-reported pain, inline with a previous study (Sjors et al., 2010), but were lessaffected than FMS patients. In contrast, women with SNPreported higher perceived stress than HC, consistently withprevious findings (Larsson et al., 2008), and higher than FMSpatients. The elevated stress levels in women with SNP mayreflect an effort to cope with daily demands that in FMSpatients may not be active anymore. Alternatively, it mayreflect a combined load from a pain condition and the dailyworkload because the employment fraction is significantlyhigher in women with SNP than in FMS patients. Taken

304 R. Riva et al.

together, the results of the present study suggest that devia-tions in the regulation of the HPA axis during the progressiontowards FMS might follow a temporal development, in linewith what has been hypothesized for chronic stress-relateddiseases (McEwen, 1998; Rosmond and Bjorntorp, 2000).Some authors have hypothesized that, at the early stageof FMS development the HPA axis may start to be hyperactive,resulting in high levels of cortisol. Then, when the systemreaches a state of exhaustion after long-term hyperactivitythe HPA axis turns to a state of hypoactivity, resulting in lowcortisol level (Miller et al., 2007; Van Houdenhove et al.,2009). Therefore, women with SNP might represent an inter-mediate stage along the continuum towards the developmentof FMS, as previously speculated by other authors (Nilsenet al., 2007). FMS and SNP might thereby share a commonpathogenesis in which a hyperactive HPA axis in women withSNP constitutes a preliminary stage that turns into a hypoac-tive HPA axis in FMS patients. However, the cross-sectionalnature of the study design does not allow drawing conclusionson causality and these therefore remain speculations. Thus, adysregulation of the HPA axis in women with SNP and FMSpatients might be a consequence of the disease rather thanthe cause. This alternative explanation has been discussed byother authors (Glass et al., 2004; Van Houdenhove et al.,2009). Alternatively, the cortisol responses in women withSNP might simply reflect a high workload exposure for thisgroup (i.e., perceived stress and employment fraction werehigher in women with SNP than FMS patients), rather thandemonstrating a temporal development towards FMS.

In the interpretation of the present results it should beconsidered that the experimental procedure for FMS patientsand HC was different from the procedure for women with SNPduring the evening of the day before the morning cortisolsampling. Patients with FMS and HC underwent a mentalstress test (Riva et al., 2010), whereas women with SNPdid not. However, it is unlikely that the experimental pro-cedure the day before influenced the results. Moreover, theparticipants of the three groups had similar activities duringthe evening and the night before the morning cortisol sam-pling, and slept in a private room in the same hotel.

Several authors have demonstrated that salivary cortisollevels can be influenced by possible confounders, such as age,gender, smoking habit, waking time, sleep quality, physicalactivity, persisting pain, and perceived stress (Federenkoet al., 2004; Hellhammer et al., 2007; Chida and Steptoe,2009). Therefore, in the present study, an extensive analysisof potential confounders for the cortisol levels in the threegroups of participants was conducted on the CAR. However,after controlling for the variables examined as potentialconfounders, there was no difference in the results betweenthe groups.

One of the major confounders for the cortisol levels uponawakening is sleep quality and duration. In the literature,findings on the relation between sleep quality and sleepduration, and salivary cortisol are inconsistent. Some authorshave found decreased morning awakening cortisol levelsafter a night of sleep loss or sleep disturbances (Vgontzaset al., 1999; Backhaus et al., 2004), while others have foundincreased morning cortisol levels (Shaver et al., 1997; Vgont-zas et al., 2001). In contrast, other studies have found noassociations between salivary cortisol and sleep quality andduration (Ekstedt et al., 2004; Wright et al., 2007; Dahlgren

et al., 2009). In the present study, we controlled for the self-reported sleep duration and for waking time, and there wasno effect of these variables on the cortisol values.

Another potential confounder for the cortisol levels isobesity. Some authors have reported positive correlationsbetween BMI and the CAR in HC (Wallerius et al., 2003),whereas other studies have found no such association (Step-toe et al., 2004). In the present study, there was no sig-nificant correlation between cortisol levels and BMI.Furthermore, we made an additional analysis on the CARexcluding all women with SNP and FMS with a BMI thatexceeded the range within the group of HC (i.e. above 35)and the differences between the groups remained signifi-cant.

Some authors have found decreased cortisol levels afterphysical exercise (Alen et al., 1988; Tabata et al., 1990),whereas others have found increased values (O’Connor et al.,1989). In our study the number of exercise sessions during aweek had a weak positive correlation with the AUCG, butafter adjusting for this variable the results did not change.

The studies that investigate salivary cortisol levels, par-ticularly CAR, in women with regional chronic pain, such asSNP, are still few in the literature. In the present study wemeasured the CAR on only 1 day. Because measures of theCAR on a single day are mainly determined by situationalfactors and only to a small proportion by trait factors, it hasbeen shown that the CAR should be measured on at least twoconsecutive week days to obtain reliable AUCG measure-ments (Hellhammer et al., 2007). However, reliability mayalso be enhanced using electronic devices to improve thecompliance of the participants (Kudielka et al., 2003) ormonitoring the participants in a controlled hospital, as wedid in the present study. Additionally, assessment of the CARin the lab or in the hospital does not differ from the oneobserved at home (Wilhelm et al., 2007). The controlledhospital setting of the study presents several advantages:participants’ compliance with the salivary sampling proce-dure was easier to control; an experimenter kept track of theparticipants throughout all the experimental procedures andinstructions for the participants were therefore always avail-able. Moreover, particular attention was paid to the CAR withstrict control over awakening time. Participants wereallowed to wake up at their natural time in a hotel roomin order to avoid stressful disturbances, but the complianceto the timing for the second and third cortisol samples waswell controlled. The results showing a normal time pattern ofthe CAR for women with SNP, FMS patients, and HC givefurther support to the assumption that the compliance wasgood. Finally, we did not control for the menstrual cycle,menopausal state or the use of contraceptive pills. However,there is no systematic effect of menstrual cycle on thedifference between groups. Moreover, some authors foundthat the CAR is not influenced by these factors (Kudielka andKirschbaum, 2003).

In the present study, women with SNP and FMS patientsshare some common features in terms of psychological char-acteristics, self-reported pain, and health complaints; how-ever, women with SNP are less affected. Conversely, womenwith SNP showed a tendency towards an elevated HPA axisactivity while FMS patients showed a decreased HPA activity,as reported in a previous paper (Riva et al., 2010). Our resultson the morning cortisol values of women with SNP are in line

Comparison of the cortisol awakening response in women with shoulder and neck pain and women with fibromyalgia 305

with the hypothesis of previous studies suggesting that adeviant HPA axis regulation in the progression from regionalto widespread musculoskeletal pain conditions may follow atemporal development (McEwen, 1998; Rosmond and Bjorn-torp, 2000; Miller et al., 2007; Van Houdenhove et al., 2009).The hyperactive HPA axis in women with SNP (i.e., highercortisol levels) might constitute a preliminary stage of ahypoactive HPA axis in FMS patients (i.e., lower cortisollevels). However, longitudinal studies are needed to furtherelucidate the underlying mechanisms of the temporal devel-opment of a deviant HPA axis regulation in relation to regio-nal and widespread chronic musculoskeletal pain conditions.

Role of funding source

This research was supported by grants to Professor Ulf Lund-berg from the Swedish Research Council and the SwedishCouncil for Working Life and Social Research. This study wasalso supported by the Centre for Musculoskeletal Research atthe University of Gavle.

Conflict of interest

All authors declare that they do not have any conflicts ofinterest.

Acknowledgments

Thanks to Associate Professor Petra Lindfors for her com-ments on the preliminary manuscript, Professor Magne Rø,MD, and Dr Tonje Okkenhaug Johansen, MD, for performingthe clinical examinations on the participants, Mrs Ann-Chris-tine Sjobeck for performing the cortisol assays, and to MrHavard Wuttudal Loras for assisting during data collection.

References

Alen, M., Pakarinen, A., Hakkinen, K., Komi, P.V., 1988. Responses ofserum androgenic-anabolic and catabolic hormones to prolongedstrength training. Int. J. Sports Med. 9, 229—233.

Backhaus, J., Junghanns, K., Hohagen, F., 2004. Sleep disturbancesare correlated with decreased morning awakening salivary corti-sol. Psychoneuroendocrinology 29, 1184—1191.

Bansevicius, D., Westgaard, R.H., Stiles, T., 2001. EMG activity andpain development in fibromyalgia patients exposed to mentalstress of long duration. Scand. J. Rheumatol. 30, 92—98.

Blackburn-Munro, G., 2004. Hypothalamo-pituitary-adrenal axis dys-function as a contributory factor to chronic pain and depression.Curr. Pain Headache Rep. 8, 116—124.

Buchanan, T.W., Kern, S., Allen, J.S., Tranel, D., Kirschbaum, C.,2004. Circadian regulation of cortisol after hippocampal damagein humans. Biol. Psychiatry 56, 651—656.

Buskila, D., 2001. Fibromyalgia, chronic fatigue syndrome, andmyofascial pain syndrome. Curr. Opin. Rheumatol. 13, 117—127.

Chida, Y., Steptoe, A., 2009. Cortisol awakening response and psy-chosocial factors: a systematic review and meta-analysis. Biol.Psychol. 80, 265—278.

Clow, A., Thorn, L., Evans, P., Hucklebridge, F., 2004. The awakeningcortisol response: methodological issues and significance. Stress7, 29—37.

Cohen, S., Kamarck, T., Mermelstein, R., 1983. A global measure ofperceived stress. J. Health Soc. Behav. 24, 385—396.

Cote, P., van der Velde, G., Cassidy, J.D., Carroll, L.J., Hogg-John-son, S., Holm, L.W., Carragee, E.J., Haldeman, S., Nordin, M.,Hurwitz, E.L., Guzman, J., Peloso, P.M., 2008. The burden anddeterminants of neck pain in workers: results of the bone andjoint decade 2000—2010 task force on neck pain and its associateddisorders. Spine 33, S60—S64.

Dahlgren, A., Kecklund, G., Theorell, T., Akerstedt, T., 2009. Day-to-day variation in saliva cortisol–relation with sleep, stress and self-rated health. Biol. Psychol. 82, 149—155.

de Kloet, E.R., Karst, H., Joels, M., 2008. Corticosteroid hormones inthe central stress response: quick-and-slow. Front Neuroendocri-nol. 29, 268—272.

Ekstedt, M., Akerstedt, T., Soderstrom, M., 2004. Microarousalsduring sleep are associated with increased levels of lipids, corti-sol, and blood pressure. Psychosom. Med. 66, 925—931.

Eriksen, H.R., Ihlebaek, C., Ursin, H., 1999. A scoring system forsubjective health complaints (SHC). Scand. J. Public Health 27,63—72.

Eysenck, H., Eysenck, S., 1975. The manual for the Eysenck Person-ality Questionnaire. Hodder & Stoughton, London.

Federenko, I., Wust, S., Hellhammer, D.H., Dechoux, R., Kumsta, R.,Kirschbaum, C., 2004. Free cortisol awakening responses areinfluenced by awakening time. Psychoneuroendocrinology 29,174—184.

Galli, U., Gaab, J., Ettlin, D.A., Ruggia, F., Ehlert, U., Palla, S., 2009.Enhanced negative feedback sensitivity of the hypothalamus-pituitary-adrenal axis in chronic myogenous facial pain. Eur. J.Pain 13, 600—605.

Glass, J.M., Lyden, A.K., Petzke, F., Stein, P., Whalen, G., Ambrose,K., Chrousos, G., Clauw, D.J., 2004. The effect of brief exercisecessation on pain, fatigue, and mood symptom development inhealthy, fit individuals. J. Psychosom. Res. 57, 391—398.

Goldenberg, D.L., 1999. Fibromyalgia syndrome a decade later: whathave we learned? Arch. Intern. Med. 159, 777—785.

Heim, C., Ehlert, U., Hanker, J.P., Hellhammer, D.H., 1998. Abuse-related posttraumatic stress disorder and alterations of thehypothalamic—pituitary—adrenal axis in women with chronicpelvic pain. Psychosom. Med. 60, 309—318.

Heim, C., Ehlert, U., Hellhammer, D.H., 2000. The potential role ofhypocortisolism in the pathophysiology of stress-related bodilydisorders. Psychoneuroendocrinology 25, 1—35.

Hellhammer, J., Fries, E., Schweisthal, O.W., Schlotz, W., Stone,A.A., Hagemann, D., 2007. Several daily measurements are nec-essary to reliably assess the cortisol rise after awakening: state-and trait components. Psychoneuroendocrinology 32, 80—86.

Jensen, C., Nilsen, K., Hansen, K., Westgaard, R.H., 1993. Trapeziusmuscle load as a risk indicator for occupational shoulder-neckcomplaints. Int. Arch. Occup. Environ. Health 64, 415—423.

Krag, N.J., Nørregaard, J., Larsen, J.K., Danneskiold-Samsøe, B.,1994. A blinded, controlled evaluation of anxiety and depressivesymptoms in patients with fibromyalgia, as measured by stan-dardized psychometric interview scales. Acta Psychiatr. Scand.89, 370—375.

Kudielka, B.M., Kirschbaum, C., 2003. Awakening cortisol responsesare influenced by health status and awakening time but not bymenstrual cycle phase. Psychoneuroendocrinology 28, 35—47.

Kudielka, B.M., Broderick, J.E., Kirschbaum, C., 2003. Compliancewith saliva sampling protocols: electronic monitoring revealsinvalid cortisol daytime profiles in noncompliant subjects. Psy-chosom. Med. 65, 313—319.

Larsson, B., Rosendal, L., Kristiansen, J., Sjøgaard, G., Søgaard, K.,Ghafouri, B., Abdiu, A., Kjaer, M., Gerdle, B., 2008. Responses ofalgesic and metabolic substances to 8 h of repetitive manual workin myalgic human trapezius muscle. Pain 140, 479—490.

Larsson, B., Søgaard, K., Rosendal, L., 2007. Work related neck-shoulder pain: a review on magnitude, risk factors, biochemicalcharacteristics, clinical picture and preventive interventions.Best Pract. Res. Clin. Rheumatol. 21, 447—463.

306 R. Riva et al.

Linton, S.J., 2000. A review of psychological risk factors in back andneck pain. Spine 25, 1148—1156.

McBeth, J., Silman, A.J., Gupta, A., Chiu, Y.H., Ray, D., Morriss,R., Dickens, C., King, Y., Macfarlane, G.J., 2007. Moderationof psychosocial risk factors through dysfunction of the hypo-thalamic—pituitary—adrenal stress axis in the onset of chronicwidespread musculoskeletal pain: findings of a population-based prospective cohort study. Arthritis Rheum. 56,360—371.

McCain, G.A., Scudds, R.A., 1988. The concept of primary fibromy-algia (fibrositis): clinical value, relation and significance to otherchronic musculoskeletal pain syndromes. Pain 33, 273—287.

McEwen, B.S., 1998. Protective and damaging effects of stressmediators. N. Engl. J. Med. 338, 171—179.

McFarlane, A.C., 2007. Stress-related musculoskeletal pain. Best.Pract. Res. Clin. Rheumatol. 21, 549—565.

Miller, G.E., Chen, E., Zhou, E.S., 2007. If it goes up, must it comedown? Chronic stress and the hypothalamic—pituitary—adreno-cortical axis in humans. Psychol. Bull. 133, 25—45.

Mork, P.J., Westgaard, R.H., 2006. Low-amplitude trapezius activityin work and leisure and the relation to shoulder and neck pain. J.Appl. Physiol. 100, 1142—1149.

Nilsen, K.B., Sand, T., Westgaard, R.H., Stovner, L.J., White, L.R.,Bang Leistad, R., Helde, G., Rø, M., 2007. Autonomic activationand pain in response to low-grade mental stress in fibromyalgiaand shoulder/neck pain patients. Eur. J. Pain 11, 743—755.

O’Connor, P.J., Morgan, W.P., Raglin, J.S., Barksdale, C.M., Kalin,N.H., 1989. Mood state and salivary cortisol levels followingovertraining in female swimmers. Psychoneuroendocrinology14, 303—310.

Picavet, H.S.J., Schouten, J.S.A.G., 2003. Musculoskeletal pain inthe Netherlands: prevalences, consequences and risk groups, theDMC(3)-study. Pain 102, 167—178.

Pruessner, J.C., Kirschbaum, C., Meinlschmid, G., Hellhammer, D.H.,2003. Two formulas for computation of the area under the curverepresent measures of total hormone concentration versus time-dependent change. Psychoneuroendocrinology 28, 916—931.

Riva, R., Mork, P.J., Westgaard, R.H., Rø, M., Lundberg, U., 2010.Fibromyalgia syndrome is associated with hypocortisolism. Int. J.Behav. Med. 17, 223—233.

Roberts, A.D.L., Wessely, S., Chalder, T., Papadopoulos, A., Cleare,A.J., 2004. Salivary cortisol response to awakening in chronicfatigue syndrome. Br. J. Psychiatry 184, 136—141.

Rosmond, R., Bjorntorp, P., 2000. Low cortisol production in chronicstress. The connection stress-somatic disease is a challenge forfuture research. Lakartidningen 97, 4120—4124.

Schalling, D., Edman, G., 1987. Personality and Vulnerability toPsychopathology: The Development of the Karolinska Scales ofPersonality (KSP). Karolinska Institutet, Stockholm.

Shaver, J.L., Lentz, M., Landis, C.A., Heitkemper, M.M., Buchwald,D.S., Woods, N.F., 1997. Sleep, psychological distress, and stressarousal in women with fibromyalgia. Res. Nurs. Health 20, 247—257.

Sjors, A., Larsson, B., Karlson, B., Osterberg, K., Dahlman, J.,Gerdle, B., 2010. Salivary cortisol response to acute stress andits relation to psychological factors in women with chronic tra-pezius myalgia–a pilot study. Psychoneuroendocrinology 35, 674—685.

Steptoe, A., Kunz-Ebrecht, S.R., Brydon, L., Wardle, J., 2004. Cen-tral adiposity and cortisol responses to waking in middle-agedmen and women. Int. J. Obes. Relat. Metab. Disord. 28, 1168—1173.

Tabata, I., Atomi, Y., Mutoh, Y., Miyashita, M., 1990. Effect ofphysical training on the responses of serum adrenocorticotropichormone during prolonged exhausting exercise. Eur. J. Appl.Physiol. Occup. Physiol. 61, 188—192.

Tak, L.M, Cleare, A.J., Ormel, J., Manoharan, A., Kok, I.C., Wessely,S., Rosmalen, J.G.M., 2011. Meta-analysis and meta-regression ofhypothalamic—pituitary—adrenal axis activity in functional so-matic disorders. Biol. Psychol. 87, 183—194.

Theorell, T., Hasselhorn, H.-M., Vingard, E., Andersson, B., 2000.Interleukin 6 and cortisol in acute musculoskeletal disorders:results from a case-referent study in Sweden. Stress Medicine16, 27—35.

Turner-Cobb, J.M., Osborn, M., da Silva, L., Keogh, E., Jessop, D.S.,2010. Sex differences in hypothalamic—pituitary—adrenal axisfunction in patients with chronic pain syndrome. Stress 13,292—300.

van der Windt, D.A., Thomas, E., Pope, D.P., de Winter, A.F.,Macfarlane, G.J., Bouter, L.M., Silman, A.J., 2000. Occupationalrisk factors for shoulder pain: a systematic review. Occup. Envi-ron. Med. 57, 433—442.

Van Houdenhove, B., Egle, U.T., 2004. Fibromyalgia: a stress disor-der? Piecing the biopsychosocial puzzle together. Psychother.Psychosom. 73, 267—275.

Van Houdenhove, B., Van Den Eede, F., Luyten, P., 2009. Doeshypothalamic—pituitary—adrenal axis hypofunction in chronicfatigue syndrome reflect a crash in the stress system? Med.Hypotheses 72, 701—705.

Vgontzas, A.N., Bixler, E.O., Lin, H.M., Prolo, P., Mastorakos, G.,Vela-Bueno, A., Kales, A., Chrousos, G.P., 2001. Chronic insomniais associated with nyctohemeral activation of the hypothalamic—pituitary—adrenal axis: clinical implications. J. Clin. Endocrinol.Metab. 86, 3787—3794.

Vgontzas, A.N., Mastorakos, G., Bixler, E.O., Kales, A., Gold, P.W.,Chrousos, G.P., 1999. Sleep deprivation effects on the activity ofthe hypothalamic—pituitary—adrenal and growth axes: potentialclinical implications. Clin. Endocrinol. (Oxf.) 51, 205—215.

Wallerius, S., Rosmond, R., Ljung, T., Holm, G., Bjorntorp, P., 2003.Rise in morning saliva cortisol is associated with abdominalobesity in men: a preliminary report. J. Endocrinol. Invest. 26,616—619.

Westgaard, R.H., Jansen, T., 1992. Individual and work relatedfactors associated with symptoms of musculoskeletal complaints.I. A quantitative registration system. Br. J. Ind. Med. 49, 147—153.

Wilhelm, I., Born, J., Kudielka, B.M., Schlotz, W., Wust, S., 2007. Isthe cortisol awakening rise a response to awakening? Psycho-neuroendocrinology 32, 358—366.

Wolf, O.T., Fujiwara, E., Luwinski, G., Kirschbaum, C., Markowitsch,H.J., 2005. No morning cortisol response in patients with severeglobal amnesia. Psychoneuroendocrinology 30, 101—105.

Wolfe, F., Smythe, H.A., Yunus, M.B., Bennett, R.M., Bombardier, C.,Goldenberg, D.L., Tugwell, P., Campbell, S.M., Abeles, M., Clark,P., 1990. The American College of Rheumatology 1990 Criteria forthe Classification of Fibromyalgia. Report of the MulticenterCriteria Committee. Arthritis Rheum. 33, 160—172.

Wright, C.E., Valdimarsdottir, H.B., Erblich, J., Bovbjerg, D.H., 2007.Poor sleep the night before an experimental stress task is associ-ated with reduced cortisol reactivity in healthy women. Biol.Psychol. 74, 319—327.

Wust, S., Wolf, J., Hellhammer, D.H., Federenko, I., Schommer, N.,Kirschbaum, C., 2000. The cortisol awakening response - normalvalues and confounds. Noise Health 2, 79—88.