Journal of Affective Disorders 47 (1998) 191–194

Preliminary communication

A comparison of salivary cortisol in chronic fatigue syndrome,community depression and healthy controls

*Paul Strickland, Richard Morriss , Alison Wearden, Bill Deakin

University of Manchester and (Guild NHS Trust), Department of Community Psychiatry, Royal Preston Hospital, Sharoe Green Lane,Fulwood, Preston PR2 9HT, UK

Received 1 May 1997; accepted 4 August 1997

Abstract

Background: Previous studies reporting cortisol hyposecretion in chronic fatigue syndrome may have been confounded byvenepuncture, fasting and hospitalisation. Methods: Morning and evening salivary cortisol were obtained on consecutivedays in the first 3 days of the menstrual cycle and compared in three samples of women taking no medication and matchedfor age: 14 patients with chronic fatigue syndrome, 26 community cases of ICD-10 current depressive episodes and 131healthy community controls. Results: The mean evening cortisol was significantly lower in the chronic fatigue syndromepatients compared to controls with depression (P 5 0.02) and healthy controls (P 5 0.005). Chronic fatigue syndromepatients without psychiatric disorder had significantly lower morning salivary cortisols compared to controls (P 5 0.009).Conclusion: Chronic fatigue syndrome patients display cortisol hyposecretion in saliva as well as plasma compared topatients with depression and healthy controls. Limitations: Small samples of female patients with cortisol estimated at onlytwo time points in the day. Cortisol secretion may be secondary to other neurotransmitter abnormalities or otherphysiological or lifestyle factors in chronic fatigue syndrome patients. Clinical relevance: Chronic fatigue syndrome isbiochemically distinct from community depression. 1998 Elsevier Science B.V.

Keywords: Chronic fatigue syndrome; Depression; Cortisol; Saliva

1. Introduction causes have been suggested (Bearn and Wessely,1994). Up to 50% of sufferers fulfil diagnostic

Chronic fatigue syndrome (CFS) is a disorder of criteria for mild or moderate depression (Abbey andunknown aetiology characterised by medically un- Garfinkel, 1991; Wessely and Powell, 1989), butexplained severe fatigue lasting at least 6 months rarely severe depression. Some authors have sug-(Sharpe et al., 1991). The aetiology of the disorder is gested that CFS is a variant of depression but it isunknown but both biological and psychological unclear if the depressive symptoms are a cause,

consequence or covariate of this chronic disabling* illness (Abbey and Garfinkel, 1991).Corresponding author. Tel.: 1 44 1772 710769; fax: 1 44

1772 710772. The hypothalamic–pituitary–adrenal axis has been

0165-0327/98/$19.00 1998 Elsevier Science B.V. All rights reserved.PII S0165-0327( 97 )00134-1

192 P. Strickland et al. / Journal of Affective Disorders 47 (1998) 191 –194

investigated in both CFS and depression. Reduced recruited from consecutive attenders at a hospitalevening cortisol and normal morning cortisol were medical out-patient clinic. There were no refusals.reported in CFS patients (Demitrack et al., 1991) but All of the patients participated in a prospectivethe confounding effects of psychiatric disorder, double-blind placebo-controlled treatment trial ofvenepuncture, hospital admission and lifestyle fac- fluoxetine and graduated exercise. Female subjects intors were not examined. Reduced morning cortisol in the community who met ICD-10 criteria for a currentCFS patients without psychiatric disorder and ele- mild or moderate depressive episode (World Healthvated morning cortisol in depressed patients referred Organisation, 1992a) and patients without any psy-to psychiatrists has also been reported (Cleare et al., chiatric disorder were recruited as primary depres-1995), but evening cortisol was not examined in this sion and healthy control groups. All of these womenstudy. Increased morning cortisol has been reported were taking part in a community study investigatingconsistently in hospital in-patients and out-patients biological and social factors in depression. Inclusionwith severe depression (Charlton et al., 1987), criteria were age 18–50 years inclusive (to ensuresuggesting that the hypothalamic–pituitary–adrenal matching for age and reduce the likelihood ofaxis abnormalities of hospital referred CFS may be organic causes for depressive symptoms) and abilitydistinct from hospital referred severe depression. to comprehend written English. Exclusion criteria for

The measurement of salivary cortisol provides a all groups were any concurrent medical diagnosismore accurate and valid measure of biologically (other than CFS) and any medication including oralactive free cortisol than estimation from plasma or contraceptives or hormone replacement therapy. Allserum (Kirschbaum and Hellhammer, 1994). Sam- subjects gave informed consent and the studiesples are obtained simply and painlessly with the received ethical approval from the relevant medicalpatient at home during a normal day, unlike the ethics committee.potentially confounding, stressful and atypical cir- Psychiatric diagnoses were made (World Healthcumstances in which plasma cortisol is obtained Organisation, 1992a) from symptoms elicited usinginvolving fasting, venepuncture and sometimes standardised semi-structured psychiatric interviewshospital admission (Demitrack et al., 1991; Cleare et in CFS patients (Lewis et al., 1992), and theal., 1995; Maes et al., 1994). depression and control groups (World Health Organi-

The aim of this study was to determine if CFS sation, 1992b). Fatigue was excluded as a depressivepatients display low salivary evening or morning symptom. In addition all subjects were asked aboutcortisol levels compared to a community sample of recent dieting and sleep disturbance. CFS patientsdepressed and healthy controls. The nature of depres- completed detailed self-rated questionnaires con-sion in patients recruited from the community is cerning the number of times they were awoken fromsimilar to that encountered in CFS patients with both sleep, the amount of time spent awake at night,groups presenting four to six symptoms of depres- regularity of sleep and waking habits and socialsion required to obtain a diagnosis of ICD-10 current routine (Morriss et al., 1997).mild or moderate depressive episodes (World Health The CFS patients were also physiologically as-Organisation, 1992a). ICD-10 severe depressive epi- sessed by measuring grip strength in the right andsodes requiring eight symptoms of depression or the left hands using a dynamometer at midday and theirsomatic syndrome which is analogous to DSM-IV peak functional work capacity determined using amelancholia are relatively rare in the community or cycle ergometer as a measure of physical fitness andCFS patients. willingness to exercise. Social function, activities of

daily living, occupational functioning and generalhealth perception were self-rated (Stewart et al.,1988).

2. Method Saliva samples were taken at 11:00 and 09:00 overtwo consecutive days at home. Premenopausal

Female CFS patients fulfilling the Oxford consen- women provided samples during the follicular phasesus research criteria (Sharpe et al., 1991) were of the menstrual cycle. All women provided at least

P. Strickland et al. / Journal of Affective Disorders 47 (1998) 191 –194 193

2 ml of saliva 30 min after brushing their teeth and patient with moderate depression who had a meanthen immediately froze the samples awaiting collec- cortisol of 4.0 mmol / l. There is a trend for CFStion. Samples were analysed using an in-house patients to have lower morning cortisol levels thancompetitive radioimmunoassay with a correlation the healthy control group. On further inspection ofbetween salivary and plasma cortisol of 0.835. Inter- the data, the morning cortisol of CFS patientsassay precision was 12.7% and intra-assay precision without mild or moderate depression was signifi-was , 6% across the whole assay range. Mean cantly lower than CFS patients with mild or moder-evening and morning cortisol levels were calculated ate depression (median (interquartile range) 3.3 (1.6)for each subject over the two day collection. vs. 6.3 (4.4) mmol / l, respectively, Mann-Whitney

Quantitative data which was normally distributed U-test: U 5 1.5, W 5 11.5, P 5 0.009). Morning cor-across the CFS, depression and healthy control tisol levels in CFS patients were not related togroups were analysed using one-way analysis of weight, dieting, sleep disturbance, grip strength,variance. Cortisol data could not be transformed to a physical fitness, social function, occupational func-normal distribution and were analysed using the tion, activities of daily living nor the patients’ ownKruskal-Wallis analysis of variance. Quantitative perception of their health. There was too small adata comparing morning or evening cortisol with variance in evening cortisol to examine any associa-physiological, lifestyle or psychiatric data were tion with physiological, functional or lifestyle vari-analysed using Spearman correlations or the Mann- ables.Whitney U-test.

4. Discussion3. Results

For the first time a non-invasive method ofThere were 14 CFS patients in the whole sample, sampling in a community setting confirms that CFS

of whom 10 had current ICD-10 mild or moderate patients have significantly lower evening cortisoldepressive episodes (World Health Organisation, levels in contrast to subjects with a similar severity1992a) and four had no current ICD-10 psychiatric of depression or healthy controls. Chronic fatiguediagnosis. There were 26 subjects in the primary patients with or without mild or moderate depressiondepression group with mild or moderate depressive display evening cortisol hyposecretion confirming anepisodes and 131 healthy controls. Table 1 shows earlier report (Demitrack et al., 1991). The results inthat the CFS, primary depression and healthy control a small sample of females confirm that CFS onlygroups taking no active medication do not signifi- patients without psychiatric disorder also have morn-cantly differ in age. ing cortisol hyposecretion (Cleare et al., 1995). Both

Table 1 shows that the CFS patients have previous studies involved plasma cortisol estimationsignificantly lower evening cortisol levels than both obtained by venepuncture after fasting in a hospitalthe depression and healthy control groups. In fact all setting.CFS patients had a mean evening cortisol of 1.0 This study adds to growing evidence that sponta-mmol / l, except one patient with mild depression neous cortisol hypersecretion is seen in more severewho had a mean cortisol of 1.5 mmol / l and one depressive episodes treated in hospitals (Maes et al.,

Table 1Median (interquartile range) cortisol in chronic fatigue syndrome (CFS) on no active medication, depression and healthy controls

CFS (n 5 14) Depression (n 5 26) Controls (n 5 131) ANOVA

Age (years), mean (SD) 36 (11) 34 (6) 34 (7) F 5 0.32, df 5 170, P 5 0.73a 2Evening cortisol (mmol / l) 1.0 (0) 1.5 (2.1) 1.5 (1.5) x 5 7.96, df 5 2, P 5 0.019

b 2Morning cortisol (mmol / l) 5.8 (4.4) 6.0 (6.1) 7.0 (3.5) x 5 5.80, df 5 2, P 5 0.056aCFS vs. depression Mann-Whitney U 5 6, P 5 0.02; CFS vs. controls U 5 321, P 5 0.005.bCFS vs. depression Mann-Whitney U 5 6, P 5 0.52; CFS vs. controls U 5 667, P 5 0.094.

194 P. Strickland et al. / Journal of Affective Disorders 47 (1998) 191 –194

Cleare, A.J., Bearn, J., Allain, T. et al., 1995. Contrasting1994) and not in non-psychiatric or communityneuroendocrine responses in depression and chronic fatiguesamples of mild and moderate depression. Oursyndrome. J. Affect. Disord. 35, 283–289.

findings support the argument that CFS is biochemi- Charlton, B., Leake, A., Wright, C., Griffiths, H., Ferrier, I.A.,cally distinct from depression in the community. The 1987. A combined study of cortisol, ACTH and dexamethasonestudy did not examine patients with ICD-10 severe concentrations in major depression: multiple timepoint sam-

pling. Br. J. Psychiatry 150, 791–796.depressive episodes or the somatic syndrome (melan-Demitrack, M.A., Dale, J.K., Straus, S.E. et al., 1991. Evidencecholia). The possibility that CFS is biochemically

for impaired activation of the hypothalamic-pituitary-adrenalsimilar to some form of atypical severe depressive axis in patients with chronic fatigue syndrome. J. Clin.episode has not been excluded (Demitrack et al., Endocrinol. Metab. 73, 1224–1234.1991). Further research in larger community samples Kirschbaum, C., Hellhammer, D.H., 1994. Salivary cortisol in

psychoneuroendocrine research: recent developments and ap-of CFS patients will need to delineate furtherplications. Psychoneuroendocrinology 19, 313–333.whether abnormal hypothalamic–pituitary–adrenal

Lewis, G., Pelosi, A.J., Araya, R., Dunn, G., 1992. Measuringfunction is a primary abnormality in CFS, secondary psychiatric disorder in the community: a standardised assess-to other central neurotransmitter abnormalities in ment for use by lay interviewers. Psychol. Med. 22, 465–486.CFS, such as serotonin or noradrenaline (Bearn and Maes, M., Calabrese, J., Meltzer, H.Y., 1994. The relevance of the

in-versus outpatient status for studies on the HPA axis inWessely, 1994), or an epiphenomenon brought aboutdepression: spontaneous hypercortisolism is a feature of majorby the lifestyle, physiological or psychiatric conse-depressed in-patients and not of major depression per se. Prog.

quences of CFS. Neuropsychopharmacol. Biol. Psychiatry 18 (3), 503–517.Morriss, R.K., Wearden, A.J., Battersby, L., 1997. The relation of

sleep complaints to fatigue, mood and disability in chronicfatigue syndrome. J. Psychosom. Res. 42 (6), 597–605.Acknowledgements

Sharpe, M.C., Archard, L.C., Banatvala, J.E. et al., 1991. Areport—chronic fatigue syndrome guidelines for research. J. R.

Funding: The Linbury Trust and The Wellcome Soc. Med. 84, 118–121.Trust. Stewart, A.L., Hays, R.D., Ware, J.E., 1988. The MOS short-form

general health survey: reliability and validity in a patientpopulation. Med. Care 26, 724–734.

Wessely, S., Powell, R., 1989. Fatigue syndromes: a comparisonReferencesof chronic ‘post-viral’ fatigue with neuromuscular and affectivedisorders. J. Neurol. Neurosurg. Psychiatry 52, 940–948.

Abbey, S.E., Garfinkel, P.E., 1991. Chronic fatigue syndrome and World Health Organisation, 1992a. International Classification ofdepression: cause, effect or covariate. Rev. Infect. Dis. 13 Disease. 10th ed. World Health Organisation, Geneva.(Suppl. 1), 573–583. World Health Organisation, 1992b. SCAN. World Health Organi-

Bearn, J.A., Wessely, S., 1994. Neurobiological aspects of chronic sation, Geneva.fatigue syndrome. Eur. J. Clin. Invest. 24, 79–90.