A defect in cortisol production in rheumatoid arthritis: why are we still looking?

Abstract

The question of whether patients with rheumatoid arthritis (RA) might have a defective hypothalamo–pituitary–adrenal (HPA) axis was first raised when RA patients who were treated with glucocorticoids in the 1950s showed dramatic improvement in their symptoms. It was initially hypothesized that this was due to an impaired ability of RA patients to synthesize sufficient amounts of endogenous glucocorticoids, but intensive investigations over the next few decades failed to reveal any significant defects in HPA axis activity in RA patients. In our review of the literature [1] we found no compelling evidence for significant differences in either basal or stress-stimulated HPA axis activity in RA compared with normal healthy individuals. However, we did highlight an inherent defect, which resided in the inability of RA patients to mount an appropriately enhanced glucocorticoid response to increased secretion of proinflammatory cytokines such as interleukin (IL)-1, IL-6 and tumour necrosis factor (TNF)-. ‘In other words’, we concluded, ‘the HPA axis response in RA is defective precisely because it is normal’. Since then, studies by ourselves and others have suggested that this statement may be an oversimplification and that subtle differences in HPA axis activity may exist in subpopulations of RA patients. The purpose of this review is to evaluate progress in assessing the integrity of the HPA axis in RA since this area was previously reviewed [1–3], to determine whether generating information about HPA axis dysfunction in RA is clinically useful, and to discuss whether this area of research remains a fruitful line of enquiry. Persuasive evidence for a major dysfunction in basal circadian HPA axis activity in RA patients remains elusive. In a study of 10 postmenopausal females, plasma adrenocorticotrophic hormone (ACTH) and cortisol were slightly decreased around midnight but were not different from those in healthy controls at other circadian time points [4]. In 17 RA patients of mixed sex with recent disease onset, morning ACTH and cortisol concentrations were similar to those of controls, although the correlation between these hormones was weaker in RA [5]. In contrast, Straub et al. observed significantly higher morning serum cortisol in untreated patients with early RA (34 subjects, largely females) compared with healthy subjects, but no difference in ACTH [6]. However, the ratios of ACTH and cortisol in RA patients in relation to the markedly elevated levels of proinflammatory cytokines IL-6 and TNF- were significantly low compared with healthy controls. Furthermore, the number of swollen joints correlated inversely with the ratio of serum cortisol to IL-6. This suggests that measuring alterations in a single hormonal parameter may be less revealing of the underlying processes of inflammation in RA than measuring a change in the overall milieu of pro- and anti-inflammatory modulators. The low ratios of serum cortisol to IL-6 and TNFin RA cannot be explained by an increased metabolic clearance rate of cortisol [7, 8]. Curiously, in more recent studies by the same group, basal serum cortisol concentrations in RA patients with mild to moderate disease activity were slightly lower [9] and ACTH concentrations were significantly lower [7] compared with healthy controls. Serum cortisol and 24-h urinary cortisol were similar in the two groups [7]. These variations reported in the literature, which may be due to differences in disease duration and/or activity, age and sex of the patient, medication, or sampling time, emphasize the difficulty in pinpointing any specific and reproducible defect in HPA axis activity in RA. In a separate study, diurnal salivary cortisol was elevated in RA patients with recent onset compared with healthy controls, and this effect was positively correlated with disease activity but not with sex or age [10]. However, in these patients the awakening