Morning cortisol and fasting glucose are elevated in women with chronic widespread pain independent of comorbid restless legs syndrome

1 Introduction

Chronic widespread pain (CWP) is defined as persistent pain for at least 3 months and with the presence of pain in at least two contralateral body quadrants and the axial skeleton [1], [2]. At least 11% of the general population has been reported suffering from CWP [2]. Dysfunctional neural pain processing mechanisms like sensitisation, long-term potentiation of afferent pain signals and dysfunctional inhibition have been suggested as potential pathogenic mechanisms of CWP [3]. Disturbed sleep has been reported as a hallmark symptom of this chronic pain condition. However, the detailed underlying mechanisms are still not fully understood [4].

Restless legs syndrome (RLS)/Willis-Ekbom disease is a neurological condition that involves sensorimotor disturbance in the legs or the arms [5]. RLS symptom expression follows a circadian rhythm with increased symptom burden during late evening and night. Periodic limb movements (PLMs) occur in approximately 80% of all RLS patients and can considerably disrupt sleep architecture [5]. The pathogenesis of RLS is still not fully understood, and current hypotheses include dopaminergic dysfunction, reduced iron content in the central nervous system, local/systemic hypoxia, genetic associations and/or alteration in neurotransmitters like orexin [6].

RLS is a common comorbidity in patients with CWP, with a reported prevalence of 30–64% compared to approximately 10% in the general population [7], [8], [9], [10]. In a recent population-based study, we demonstrated that the likelihood of RLS is 5–6 times higher in middle aged women with multi-site pain compared to pain free controls independent of weight, age and psychiatric comorbidity [9]. Moreover, in women with CWP there was a strong detrimental influence of comorbid RLS on sleep quality, the degree of daytime sleepiness and body fatigue. The pathophysiological mechanisms to explain the association between CWP and RLS are still widely unknown.

The aim of the current case-control study was to investigate the impact of CWP on stress hormones, glucose metabolism and overnight skin sympathetic nervous activity. The influence of sleep disordered breathing and comorbid conditions (e.g. RLS, depression and anxiety) was also investigated.

2 Methods

2.1 Study design

Study participants were recruited from a population-based questionnaire survey cohort of females aged 18–65 years in the Swedish county Dalarna [9]. In the current study, based on the CWP classification used in the original cohort, 27 women with pain presenting in all five body areas (neck, shoulder-arms, upper back, lower back and legs) were randomly selected from 3,060 participants. Twenty-seven controls without pain in the five areas at baseline were also identified. At follow-up, 4 out of the 27 controls reported multi-site pain and were therefore classified as CWP in the study. The final analysis group included 31 individuals in the CWP and 23 subjects without pain in the control groups (Fig. 1). All 54 subjects signed an informed consent after receiving oral and written information. The study was approved by the regional Ethical Review Board at Uppsala University (Dnr 2010-124).

Fig. 1:

The study flow chart illustrates the selection process of the study participants from the original population-based cohort to the final analysis cohort of the current study.

3 Results

3.2 Blood sample analysis

Morning cortisol levels (464±141 vs. 366±111 nmol/L, p=0.011) and fasting blood glucose (6.0±0.8 vs. 5.4±0.7 mmol/L, p=0.007) were elevated in subjects with CWP compared to the non-CWP group (Table 2). A ferritin threshold of <50 μg/L was fulfilled by one third of those with CWP and one quarter of the controls. There was no difference in mean ferritin levels between the CWP group and controls (89±68 vs. 88±45 μg/L, n.s.) or between women with and without RLS. The strong influence of CWP on morning cortisol levels and fasting blood glucose is illustrated in Figs. 2 and 3. There were no significant associations between RLS severity and cortisol or glucose levels. GLM analysis demonstrated that CWP diagnosis was associated with a 122±47 nmol/L increase in morning cortisol and a 0.89±0.28 mmol/L increase in fasting blood glucose after adjustment for age, BMI, RLS and anxiety status (p=0.009 and 0.001, respectively).

Table 2:

Blood sample analysis.

Variable	Chronic wide-spread pain (n=25)	Controls (n=23)	Between group difference (p-Value)
Morning cortisol (nmol/L)	464±141	366±111	0.011
Fasting blood glucose (mmol/L)	6.0±0.8	5.4±0.7	0.007
Ferritin (μg/L)	89±68	88±45	n.s.
% Ferritin <50 μg/L	36%	26%	n.s.
% Ferritin <30 μg/L	16%	13%	n.s.

Fig. 2:

Association between the chronic widespread pain/restless legs syndrome and morning cortisol (nmol/L).

Fig. 3:

Association between chronic widespread pain/restless legs syndrome and fasting blood glucose (mmol/L).

4 Discussion

From our study we can demonstrate several novel findings. Firstly CWP is associated with an upregulated stress response, both during wakefulness, demonstrated by elevated morning cortisol and impaired glycemic control, as well as during sleep, as shown by activation of sympathetic vascular control (PWA attenuation) and poor subjective sleep quality. Secondly, RLS in women with CWP is not independently associated with metabolic dysfunction, alterations in iron metabolism or overnight hypoxic load as previously described in RLS without CWP [23], [24]. These findings suggest a specific phenotype with RLS in this particular population, characterised by a high degree of neuropathic pain symptoms in CWP patients with RLS. Thirdly, depressive mood and cognitive dysfunction were linked to the RLS diagnosis rather than to CWP. It is also speculated that central nervous system processes linked to depression and sleepiness/fatigue also may be involved in the generation of RLS symptoms.

Acute pain induces a stress response that involves increased catecholamine secretion. Different types of stressors may induce specific biological and behavioural responses and the response pattern may vary significantly [25]. On the other hand, chronic pain conditions like CWP are considered to be associated with increased neurotransmitter activity of the central nervous system [26]. The hypothalamic-pituitary-adrenal (HPA) axis and the sympatho-adrenomedullary systems appear to be involved in both of the processes. In the current study, we found an almost 25% increase in morning cortisol levels irrespective of the confounders assessed. This finding is in line with previous studies which identified an activated HPA axis in fibromyalgia patients compared with matched controls [27], [28]. However, other studies also demonstrated a depressed HPA axis [29], [30], [31] or normal serum cortisol levels [32], [33] in patients with fibromyalgia. The differences in findings remain unexplained but may be related to the population-based setting in our study compared with clinic-based chronic fibromyalgia patients in previous studies. The duration of exposure to stimulation and comorbidity may also play a role in cortisol production and release in various pain conditions [34], [35].

A strong association between chronic pain and psychiatric comorbidity has been observed in several clinical studies, and a combination of HPA axis dysfunction and disturbed monoamine transmission was proposed as a key to understanding the association [36]. A recent study reported an increased odds ratio for depression in subjects with hypercortisolism (OR 1.9–2.4) when compared to eucortisolism [37]. Indeed, in our study we identified an increased prevalence of psychiatric diseases in subjects with CWP. Most interestingly, depression symptoms were linked to a comorbid RLS diagnosis, whereas anxiety symptoms were linked to CWP subjects with elevated cortisol levels.

A number of recent studies conducted in patients with fibromyalgia [7], [8], [9] and other chronic multi-site pain conditions [38] established a strong association between CWP and RLS. Most recently, we extended this association by showing a strong relationship between the number of pain areas and the prevalence of RLS [9]. Shortened and disturbed sleep was a hallmark in the association between CWP and RLS [39]. However, the possible mechanisms relating these two conditions have not been extensively studied. RLS is known to be related to iron deficiency [40], local or systemic hypoxia [23], [24], [41], hypercortisolism [42] and increased motor dysfunction during sleep [43]. Whilst we examined those factors as potential mechanisms for the dose-response relationship between pain localisation and RLS symptoms, none of those factors could be identified in our study. A partial explanation may be the limited sample size of the current investigation or CWP status rather than RLS status as inclusion criteria for entry in the study cohort. Another potential hypothesis for this particular RLS phenotype in CWP patients may be that the opioid system seems to play an important role in RLS. When RLS is associated with pain, endogenous opioids may excite dopaminergic activity that is either specific for reducing PLMs or has an effect sufficient enough to reduce PLMs but not to reduce RLS symptoms. This might partially explain the lower PLM index in women with CWP compared to the pain free group. Alternatively, we cannot exclude that RLS in CWP subjects has a different aetiology and pathogenic mechanism as evidenced by the finding that a neuropathic component had an increased prevalence in CWP subjects with RLS compared to those without [44].

Our study has several strengths and also limitations which need to be addressed. First, we describe a prospective study including individuals from a population-based sample which clearly increases the generalisation of our findings. Second, classification of CWP and RLS status is based on a detailed clinical interview, a physical examination, and questionnaires all performed by a CWP and RLS expert team. Our data also indicate a high reliability of the RLS/CWP status classification over time between baseline and follow-up assessments. Third, while missing data reduced the sample size, objective data were collected for metabolic parameters, causes of sleep fragmentation (apnoeas, hypopnoeas, flow limitation, and PLM), and overnight hypoxic exposures. In addition, in finger pulse wave analysis only 17/3 subjects (CWP/controls) were included in the final analysis. While control subjects did not perform all clinical examinations or interview with the CWP/RLS expert, we argue that the bias is toward a lower incidence of RLS diagnosis in the control group. The low IRLSSG rating score in the control group suggests that undiagnosed RLS in controls is less likely.

5 Conclusion

In this prospective study, we demonstrated that elevated morning cortisol levels, impaired fasting glucose control and markers of increased adrenergic sympathetic activity during sleep are associated with CWP in women. The stimulation of adrenergic sympathetic activity was not further influenced by RLS, although RLS was common in CWP and associated with increased mental and cognitive dysfunction. Future studies are warranted to investigate the influences of HPA abnormality and increased nocturnal sympathetic activity on sleep in patients with CWP.