AHMF: Conf 2001: A Harvey

Are cortisol levels low in Chronic Fatigue Syndrome?: A meta-analysis

A. Harvey, G. Purdie, J. Bushnell, P. Ellis

Department of Psychological Medicine,
Wellington School of Medicine and Health Sciences,
University of Otago,
NZ

A number of studies suggest that cortisol may be lower in Chronic Fatigue Syndrome (CFS) than in healthy controls, whereas other studies have reported less convincing results.

To conduct a meta-analysis aimed at determining the status of hypocortisolism in CFS.

Literature searches were carried out using Medline, PsycLIT, EMBASE, PsycINFO and relevant articles in peer reviewed journals. Inclusion criteria were; application of either the CDC (1988, 1994) or the Oxford (1991) diagnostic criteria, and comparable healthy controls. To investigate sources of heterogeneity, the test statistic Q was applied to the data. 24-hr UFC cortisol measures were compared with am and pm plasma cortisol samples and clinical variability was examined by performing a separate analysis on studies that had recruited patients through tertiary care providers as opposed to primary care.

14 published studies (k) were found that met the inclusion criteria; n=344 CFS patients and n=272 healthy controls. A significant (p<0.0001) low to moderate effect size was found (-0.36) for reduced basal cortisol levels in CFS. However, there was significant heterogeneity amongst the studies (p<0.0001). Investigation of methodological differences revealed that 24-hr urinary cortisol were better at measuring basal cortisol levels, increasing the effect size (EF = -0.73, k=4), although there was still significant heterogeneity (p<0.01). Morning plasma cortisol levels also showed a significant effect size (EF = -0.35, k=4), substantially reducing the heterogeneity (p=0.50). Afternoon and evening plasma samples were unable to distinguish between groups (EF = 0.01, k=7), the heterogeneity was also reduced (p=0.045). Analysis of 7 studies that reported recruiting patients from tertiary care increased the effect size (EF = 0.63), although there was significant heterogeneity (p<0.01).

Reduced basal cortisol levels are relevant to the experience of Chronic Fatigue Syndrome. Patients seen in tertiary care perhaps show more endocrine abnormalities and clinical methodologies seem to be important in assessment procedures.