Understanding the Mechanisms of Pain in Rheumatoid Arthritis

Pain is a debilitating feature of rheumatoid arthritis (RA) and is often described by patients as their most important symptom. Rheumatoid arthritis pain has traditionally been attributed solely to joint inflammation, however despite the advent of increasingly effective disease modifying agents, patients continue to report pain at long term follow up. The cause for ongoing pain is multifactorial and includes joint damage and pain sensitisation. In this book chapter, we will describe the mechanisms underlying the distinct components of pain which are manifest in rheumatoid arthritis and discuss why a thorough assessment of pain is vital to target treatments appropriately.


Introduction
Rheumatoid arthritis (RA) is a systemic inflammatory disease with a prevalence of between 0.5 to 1% in different worldwide populations [1]. Inflammation predominantly affects the joints causing synovitis, pannus formation and if left untreated, joint destruction. Patients with RA classically present with tender and swollen joints, early morning joint stiffness and systemic symptoms such as fatigue. Severe pain is a particularly debilitating feature of RA that is commonly described as patients' most important symptom [2]. In addition to causing a significant impact on quality of life, studies have shown that RA pain is associated with psychological distress, impaired physical and social function and increased healthcare costs [3].
The pathogenesis of pain in RA is multifactorial. Traditionally, pain was entirely attributed to synovitis and consequent joint destruction. With the advent of increasingly effective disease modifying agents, joint inflammation has become a more treatable cause of pain and joint destruction is preventable. Indeed, the randomised controlled trials (RCTs) that supported the use of classical disease modifying anti-rheumatic drugs (DMARDs), showed statistically and clinically significant reduction in pain with treatment [4]. However, despite effective control of inflammation and disease remission, patients have continued to report troublesome pain at follow-up [5]. The same has been shown in patients taking biologic DMARDs [6]. This suggests that pain does not always fully resolve with the effective suppression of synovitis [7]. Observational studies have also highlighted the complex relationship between pain and inflammation in patients with RA. For example, large discrepancies between objective measures of inflammation such as acute-phase proteins and reported pain, have been shown in some patients with RA [8]. Taken together, this evidence suggests that inflammation and joint destruction alone cannot account for the total pain manifesting in RA. Indeed, increasing evidence supports a role for aberrant pain processing, including peripheral and central pain sensitisation, in the pathogenesis of pain in RA. Throughout this book chapter, we will explore the different mechanisms underlying the perception of pain in patients with RA.

Inflammation in RA
RA is a pathologically heterogenous autoimmune condition. The disease can broadly be divided into sero-positive and sero-negative subtypes. In sero-positive patients, the presence of anti-citrullinated peptide antibodies (ACPAs), is associated with more severe joint damage and increased mortality [9]. In these patients, ACPAs bind to citrullinated autoantigens including fibrinogen, vimentin, collagen type 4 and α-enolase, resulting in the formation of immune complexes (ICs) [10]. ICs activate the complement system and trigger inflammatory cell infiltration within the synovium [11].
The pathology of RA is characterised by the activation of cells of both the innate and adaptive immune system within the synovial matrix. The innate immune response consists of macrophages, mast cells and dendritic cells. These cells produce inflammatory mediators including cytokines, chemokines, lipids, proteases and growth factors. These mediators attract neutrophils and activate cells of the adaptive immune system, such as T cells, B cells and plasma cells. The inflammatory cytokines produced during the innate immune response shape the subsequent activation of the adaptive immune system. For example, cytokines produced in the early phases of inflammation regulate the differentiation of naïve T helper cells into T helper cell subsets and the subsequent T cell response.
In RA, the inflammatory milieu within the synovium is characterised by complex cytokine and chemokine interactions. Cytokines including TNF-α and IL-6 appear to be particularly important, and biologic agents targeting these mediators are well-established treatments for RA [12].
Inflammation results in a catabolic state within the joint. One of the pathognomonic features of RA is the synovial pannus, a hypertrophied area of synovium with tissue destructive properties [13]. Within the pannus, synovial fibroblasts assume an inflammatory phenotype resulting in enhanced cartilage catabolism and synovial osteoclastogenesis [14]. Cytokine-mediated chondrocyte activation results in the stimulation of catabolic pathways. Enzymes including matrix metalloproteinases (MMPs) are activated to degrade the cartilage matrix [15]. Bone erosion is stimulated by the interaction between RANK-L on fibroblasts, T and B cells and its receptor RANK on dendritic cells, macrophages and pre-osteoclasts [16]. Ultimately, this process can result in cartilage and bone destruction and joint deformity.
Therapies that target inflammation such as conventional DMARDs and biologic therapies are effective at suppressing synovitis and reducing joint destruction. The treat-to-target approach is widely recommended for the management of RA. This strategy involves regular monitoring of disease activity, using validated scoring measures such as the DAS28, and escalation of treatment if a target is not reached. RCTs have found that this approach substantially improves disease activity, radiographic progression, quality of life and physical function [17]. These immunomodulatory agents have been shown to reduce pain, albeit not completely [18]. Throughout the next section of this chapter, we will discuss the inflammatory basis of pain in RA.

Pain and joint inflammation
Inflammation has long been accepted to cause pain. Indeed, pain was one of the cardinal features of inflammation, as defined by Celsus in the first century [19]. Pain secondary to inflammation can be classified into acute or chronic pain. The neurotransmission of acute pain signals in response to noxious stimulation involves the activation of a specialised subset of sensory neurons called nociceptors. Nociceptors innervate peripheral tissues, including joints, and transmit painful stimuli to the dorsal root ganglion (DRG). There are many subsets of nociceptors, each responding to different types of noxious stimuli. Aδ and C fibres are the two main types of primary afferent nociceptors [20]. Whilst both Aδ and C fibres are found in superficial organs, such as the skin, C-fibres generally supply deeper structures such as joints [20]. C-fibres are activated by thermal, chemical or mechanical stimulation, resulting in poorly localised, dull pain sensation [20].
The activation of nociceptors involves the stimulation of ligand-gated and voltage-gated ion channels including transient receptor potential cation channel, subfamily A, member 1 (TRPV1), transient receptor potential cation channel, subfamily A, member 1 (TRPA1), Na v 1.7, Na v 1.8, and Na v 1.9 channels, which are expressed on peripheral nerve terminals [21]. Activation of these channels results in the stimulation of intracellular signalling pathways and the transmission of acute pain signals [21]. In the longer term, chronic inflammation results in long lasting changes in nociceptor signalling resulting in peripheral pain sensitisation, a phenomenon that we will discuss later in this chapter.

Synovial joint structures and pain
Arthritic pain is thought to be mediated by nociceptors that innervate the synovium and subchondral bone. In contrast, under physiological conditions, cartilage is an aneural and avascular tissue. This is illustrated in Figure 1.
In RA, chronic inflammation is thought to result in structural and functional changes in the peripheral innervation of joints. This has been shown in animal models where chronic synovitis results in increased innervation of the synovium and increased spontaneous and mechanical-induced firing of articular primary afferents [22,23].

Nociceptor pathways
Nociceptor pathways mediating acute pain perception in response to inflammation are well defined. In the periphery, local immune cells release inflammatory mediators, such as cytokines, that act on the peripheral nerve terminals of nociceptor neurons. This activates the nociceptors to transmit signals via the DRG through the spinothalamic tract to the higher cortical centres, resulting in the perception of pain. It is also well accepted that inflammation can result in heightened nociceptor sensitivity to both noxious and innocuous stimuli. In this case, the activation of nociceptors by inflammatory mediators triggers intracellular signalling cascades that reduce the threshold for nociceptor neurons to fire action potentials [21]. This results in heightened pain sensitivity which can manifest as allodynia; the sensation of pain arising from a non-painful stimulus, or hyperalgesia; a heightened sensation of pain in response to painful stimulation. Throughout the next section of this chapter, we will discuss the inflammatory mediators that stimulate nociceptor activation and sensitisation in RA.

Pain and innate immunity
Cells of the innate immune system, including neutrophils, mast cells and macrophages, release noxious inflammatory mediators and have been shown to stimulate pain and pain sensitisation in a wide range of models and systems. For example, in mouse models of carrageenan-induced inflammatory pain, neutrophils migrate to tissues and sustain pain through the production of cytokines and prostaglandin E2 [24]. In incisional wound injury, macrophages (CD11b + myeloid cells) have been shown to mediate acute pain and pain sensitisation [25]. Mast cell degranulation activates nociceptor firing acutely and may also contribute to pathology of chronic pain and mast cells have been shown accumulate in chronic inflammatory conditions such as complex regional pain syndrome [26,27]. Throughout the next part of the chapter, we will discuss the noxious inflammatory mediators that are released by innate immune cells.

Lipid mediators of pain
Pro-inflammatory lipids include cyclooxygenase (COX) dependent molecules such as prostanoids (prostaglandins, prostacyclins and thromboxanes). COXdependent molecules are well known to cause pain and pain sensitisation and inhibition of the COX enzyme, using non-steroidal anti-inflammatory drugs (NSAIDs), is used for the suppression of pain and inflammation. Indeed, NSAIDs are potent analgesic and anti-inflammatory medications which are effective for the treatment of acute inflammatory pain including synovitis [28].
Studies have investigated the mechanism of action underlying the noxious effect of prostaglandins. Prostaglandin E2 (PGE2) has been shown to activate nociceptors through the binding of EP1-EP4 receptors. This stimulates pain and pain sensitisation via multiple mechanisms. PGE2 stimulates proximal ion channels in nociceptive neurons. This sensitises the neurons to painful stimuli [29]. PGE2 activates more persistent pain sensitisation via PKA and PKC-mediated activation of NFκB in the dorsal root ganglion neurons [30].
Many other classes of pro-inflammatory lipids are thought to be involved in the activation of nociceptor activity. For example, lysophosphatidic acid and sphingosine-1-phosphate are produced during inflammation and have been shown to activate nociceptors leading to increased TRPV1 activity [31]. Leukotrienes may also have an noxious effect and the injection of leukotriene B4 has been shown to activate C and Aδ-fibres in rat models and induce hyperalgesia in humans [21,32].
More recent work has also demonstrated a role for anti-inflammatory and proresolving lipids in the silencing of pain. For example, pro-resolving lipids, including lipoxins, resolvins and protectins have generally been shown to have analgesic effects [33]. Further work is required to characterise the underlying molecular pathways but these mediators may represent targets for the future treatment of pain [33].

Neurotransmitters and pain
Innate immune cells release neurotransmitters capable of modulating pain transmission. For example, mast cells contain histamine and serotonin that are released on degranulation. Histamine triggers pain sensitisation through the activation of H1 and H2 receptors expressed on nociceptors [34]. This results in increased expression of Nav1.8 channels and increased sensitivity to noxious stimuli [34,35].

Cytokines and pain
Inflammatory cytokines represent another important class of molecules that stimulate nociceptors and activate pain sensitisation. IL-1β was the first cytokine to be described as hyperalgesic [36]. This finding was seminal in the field of neuroimmunology and represented early evidence for the cross-talk between the immune system and pain sensitisation. Cytokines have now been found to play important roles in pain modulation in most painful conditions, including RA. Notably, proinflammatory cytokines including IL-1β, IL-6, TNF-α, IL-17A and IL-5, have all been shown to active nociceptors directly [21].
IL-1β sensitises nociceptors through different intra-cellular signalling pathways. Firstly, IL-1β activates the p38 MAPK-mediated phosphorylation of Nav1.8 sodium channels resulting in increased action potential generation and an associated mechanical and thermal hyperalgesia [37]. Secondly, the activation of IL-1R1 by IL1-β, has also been shown to result in increased TRPV1 expression on nociceptors, resulting in thermal pain sensitisation in animal models [38].
Increasing work suggests a role for IL-17 in pain sensitisation. Indeed, many painful autoimmune diseases, such as RA and psoriasis, are characterised by a Th17 immune response. IL-17A has been shown to be broadly expressed by nociceptors and IL-17 has been demonstrated to induce a rapid increase in neuronal excitability [43]. In animal models of RA, IL-17 has been shown to induce hyperalgesia, through a mechanism dependent on the amplification of TNF-α, IL-1B, CXCL-1, endothelin 1 and prostaglandins [44].
In summary, IL-1β, IL-6, TNF-α and IL-17 stimulate pain and pain sensitisation through the synthesis of prostaglandins and/or the activation of sodium or TRP  Figure 2 Summarises the inflammatory mediators that have been shown to activate and sensitise nociceptors 21 . As illustrated, innate and adaptive immune cells release inflammatory mediators that act on their respective receptors to activate nociceptors and sensitise pain signalling through Nav and TRP channels. channels. The different cytokines appear to act via different intracellular signalling pathways, however it remains unclear whether different immune responses (e.g. Th1, Th2 or Th17) induce different pain characteristics through the activation of specific nociceptors and pain receptors.

Immune derived growth factors in pain
Innervation by nociceptors is a dynamic process affected by neurotrophic factors. These factors are often upregulated in response to inflammation or tissue injury and are important to restore the density of innervation post-injury [21]. If there is inappropriate or excessive release of neurotrophic factors, heightened pain sensitivity can occur [21]. Nerve growth factor (NGF) is an important neurotrophic factor that is secreted by innate immune cells during the acute phase of inflammation. NGF activates the receptor TrkA on nociceptors, stimulating the P13K/Src kinase pathway and the phosphorylation of TRPV1 and its translocation into the cell membrane [45]. This results in the rapid sensitisation of nociceptors in response to stimulation by NGF. In the longer term, NGF has been shown to stimulate axonal terminal sprouting, contributing to increased pain sensitivity [46].

A role for ACPAs in pain in RA
It is well established that arthralgia can precede overt joint inflammation and that joint pain is often one of the first symptoms of emerging RA. The mechanism underlying arthralgia preceding inflammation remains unclear but a role for ACPAs has been suggested. Observational studies have shown that ACPAs frequently occur in the preclinical phase of disease and can be detected months to years prior to diagnosis [47]. Experimental studies have raised the possibility that ACPAs can induce pain via a pathway independent of joint inflammation. In one study, mice injected with human or murine ACPAs developed increased pain sensitivity, despite no signs of joint inflammation. In this study, ACPAs were shown to bind to osteoclasts in the bone marrow, and induce CXCL1/2 expression and release. Intra-articular injection of CXCL1/2 was shown to evoke pain-like behaviour and this was inhibited by an IL-8 inhibitor, reparixin [48]. Further work is required to confirm this hypothesis. If correct, it could alter the management of ACPA positive arthralgia and offer new therapeutic targets in the management of early RA.
In summary, inflammation is a well-accepted cause of pain in RA and many inflammatory mediators have been shown to stimulate nociceptor activation and sensitisation, as summarised in Figure 2.
Despite the important role for inflammation in pain in RA, the extent of inflammation does not always correlate with the severity of total pain reported RA patients. Indeed, observational studies have shown that changes in inflammation account for only 40% of changes in pain in RA patients [49]. Furthermore, factors associated with the degree of inflammation such as serology, acute phase response and joint damage correlate poorly with pain prognosis in RA patients [7]. Moreover, in common with other chronic pain conditions, psychosocial factors and female gender predict pain prognosis more accurately than the severity of inflammation [7]. Therefore, additional mechanisms must be responsible for the pain experienced in RA. These mechanisms include joint damage and aberrant pain sensitisation.

Joint damage and pain
The contribution of structural joint changes to the total pain in RA is controversial. In patients with advanced RA, erosions and joint space narrowing are associated with disability and make a small but significant contribution to total reported pain [50]. Moreover, patients with advanced disease show an improvement in pain following joint replacement surgery [51]. However, as more effective disease modifying protocols have been developed, structural joint damage in RA has decreased and corresponding rates of orthopaedic surgery have declined [52]. The prevention of joint damage has produced superior pain outcomes but it is not clear how much of this can be attributed to the prevention of structural damage versus the suppression of inflammation or prevention of pain sensitisation. In recent studies, radiographically assessed joint damage appears to make a small contribution to pain in RA patients [53]. However, some of this pain may be explained by coincident osteoarthritis (OA), which occurs in a similar demographic of patients.
The correlation between joint damage and pain severity appears weak, although investigation on this subject has primarily occurred in patients with OA and relatively little data exists for patients with RA. In OA, structural joint changes do not correlate well with joint pain [54]. The severity of radiographic OA has been shown to explain <20% of the variance in pain intensity [54]. Furthermore, postjoint replacement, many patients continue to report pain. 10% of patients posttotal hip replacement (THR) and 20% post-total knee replacement (TKR) report unfavourable long term pain outcomes [55]. This suggests that structural joint damage alone cannot explain the total pain experienced in OA. Like in RA, central pain sensitisation has been proposed to explain the pain not accounted for by joint destruction [56].

Central pain sensitisation and RA
Processing by the central nervous system (CNS) can affect pain reporting, sensitivity, intensity and pain characteristics [57]. Aberrant pain processing can result in central pain sensitisation; an amplified response of the central nervous system to peripheral nociceptive input [58]. The term central sensitisation was coined in 1989 by Woolf and colleagues based on work in the rat model showing hyperexcitability of spinal cord neurons in response to peripheral tissue injury [58]. Physiologically, central sensitisation represents a state of hyperexcitability of spinal and supraspinal structures due to amplified neuronal signalling involving enhanced synaptic and neurotransmitter activities [59].
An increasing abundance of evidence supports the role for central pain sensitisation in RA and an understanding of central sensitisation is important to optimise patient treatment. Clinically, pain secondary to an inflammatory flare must be differentiated from pain secondary to central sensitisation as they require vastly different management approaches. Throughout the next part of this chapter, we will discuss the molecular basis of pain transmission from the periphery to the CNS, clinical evidence supporting a role for pain sensitisation in RA and some proposed mechanisms for pain sensitisation in the DRG and in the cerebral cortex.

Molecular basis of pain sensitisation
As discussed previously, A-δ and C nociceptive neurons are activated by inflammatory mediators in the periphery. These fibres converge at the DRG, along with non-noxious A-β fibres. Following activation, nociceptor fibres release substance P (SP), calcitonin gene-related peptide (CGRP), glutamate, aspartate and NGF at the afferent nerve endings into the synaptic cleft [60]. These neurotransmitters activate their corresponding receptors on post-synaptic neurons. Activation of postsynaptic receptors results in intracellular signalling changes. For example, activation of NMDA receptors results in increased membrane permeability, intracellular entry of calcium, activation of protein kinases and the expression of c-fos [61]. These signalling changes result in the hyperexcitability of the secondary neurons and amplification of the peripheral noxious stimulus. Post-synaptic neurons ascend in the spinothalamic tract to the thalamus, hypothalamus, limbic system and the somatosensory cortex [61]. These signalling pathways are summarised in Figure 3.
Animal models of RA have been used to investigate the molecular mechanisms underlying spinal pain sensitisation. In these models, molecular changes have been shown to occur in the DRG, spinal neurons and spinoreticular neurons. For example, in complete Freund's adjuvant (CFA) induced arthritis models, increased expression of SP, CGRP, NPY, c-fos, TRPV1, P2X3 and Trk-A receptors in the DRG have been demonstrated [62]. These changes are thought to result in hyperexcitability of spinal neurons and enhanced sensitivity to nociceptor signalling.

Clinical evidence for a role of pain sensitisation in RA
Patients with RA show widespread reductions in pain threshold and increased pain sensitivity, not only over inflamed joints but at distant, non-articular sites [62]. Evidence to support this has come from clinical studies using techniques such as quantitative sensory testing (QST). This technique involves the application of stimuli under standardised testing protocols and the quantification of the participants sensory experience. QST employs different tools for the assessment of the perception of vibration, touch, proprioception, pinprick or blunt pressure sensitivity. RA patients have a lower pain threshold than healthy controls with QST [63]. Furthermore, sensitisation has been shown to affect a wide range of sensory modalities, including thermal and mechanical stimulation.
Studies have demonstrated that pain thresholds vary substantially between patients with RA. Multiple factors have been shown to correlate with differences in pain threshold. Importantly, these include high tender joint count and prolonged disease duration [64]. This suggests that the persistence of nociceptive stimulation results in long-term changes in pain processing resulting in central pain sensitisation. Other factors that have been shown to influence pain threshold include sleep quality, psychosocial factors and analgesic use [65].
Repetitive sensory stimulation, also known as temporal summation, is another experimental model that has been used to investigate central sensitisation in RA. Temporal summation occurs when the time between stimuli is short enough to prevent the dissipation of postsynaptic action potentials before re-activation [66]. This results in a higher membrane potential, increasing the probability that further stimulation will result in post-synaptic activation. In healthy controls, repetitive stimulation results in the reduction of pressure pain thresholds [62]. Studies have shown that this response is augmented in RA patients [67]. This has also been demonstrated electrophysiologically through the measurement of action potentials in response to repetitive stimulation. In healthy controls, there is an increase in the amplitude of action potential evoked from repetitive stimulation using noxious stimulation. This response is amplified in RA patients and has been shown to correlate with disease activity scores and high tender joint counts [68].

Neuropathic pain in RA
In addition to measuring pain thresholds, pain characteristics can be analysed to assess the possible contribution of pain sensitisation to overall pain experience. Specifically, pain questionnaires are commonly used to detect the presence of neuropathic-sounding pain. Neuropathic pain is the perception of pain in the absence of nociceptive input or peripheral tissue damage and is caused by pathology of the peripheral or central nerves. A classic example of neuropathic pain is sciatica. This pain has distinct characteristics such as burning, radiation, shooting, tingling and sensitivity to non-painful stimuli (i.e. allodynia). RA can be associated with neuropathic pain through several mechanisms including compression neuropathy (e.g. carpal tunnel syndrome), co-morbidities (e.g. diabetes), vasculitis (resulting in mononeuritis multiplex) or drug therapies (e.g. gold or leflunomide). Nevertheless, emerging evidence suggests that RA itself can result in neuropathic pain through the induction of aberrant pain processing.
The painDETECT questionnaire enables the classification of pain into likely, possibly or unlikely to be of neuropathic origin. Patients with RA often describe pain with neuropathic features and painDETECT questionnaires can yield between 5 to 20% fulfilling criteria for "likely neuropathic pain" [62]. A significant proportion of these patients have no underlying evidence of neuropathy. One study demonstrated that only 33% of RA patients fulfilling clinical criteria for neuropathic pain had clinical evidence of neuropathy [69]. Of the remaining patients, 57% were shown to have subclinical or axonal neuropathy [70]. This left a significant number of patients with RA who reported neuropathic-sounding pain in the absence of objective nerve injury. It has been suggested that this pain occurs secondary to pain sensitisation however, this has not been proven. Nevertheless, neuropathicsounding pain is an important clinical feature as it predicts inferior pain outcomes. Indeed, a positive correlation between VAS pain scores and painDETECT scores has been demonstrated and patients with probable or likely neuropathic pain have been shown to report significantly higher VAS scores than patients without neuropathicsounding pain [71].
Although the painDETECT questionnaire is a useful tool for characterising pain, care must be taken to interpret results based only on questionnaires. Furthermore, confounding effects with pain severity may affect interpretation. Patients with fibromyalgia demonstrate high painDETECT scores, although evidence of pathology in the peripheral or central nervous system has been difficult to demonstrate. This raises the question of whether painDETECT scores identify pain with similar features to neuropathic pain rather than neuronal pathology itself. Further work is required to fully understand the significance of neuropathic sounding pain in RA.

Fibromyalgia-RA
The association between fibromyalgia and RA sheds light on the complex relationship between inflammation, pain and central pain sensitisation. Fibromyalgia (FM) is the prototypical central pain sensitivity syndrome. Clinically, FM is characterised by chronic widespread pain, sleep disturbance and impaired cognition [72]. Observational studies have shown that the prevalence of fibromyalgia in RA patients is much higher than in the general population with estimated prevalence of 18-24%, compared to 2-4% in non-RA cohorts [73,74].
Two groups of fibromyalgia (FM) have been characterised. Patients with "primary" FM report pain in the absence of identifiable nociceptive input [72]. These patients generally report regional pain syndromes that progress to widespread pain phenotypes with time. "Secondary" FM occurs when aberrant centralised pain processing occurs in the context of identifiable nociceptive input, for example in inflammatory arthritis [72]. It is not yet clear whether these conditions represent the same or different diseases.
The co-existence of FM in RA patients is associated with increased pain scores, a poorer quality of life and worse patient-reported outcomes. In a meta-analysis of 18 studies, RA patients with co-morbid FM had significantly higher pooled DAS28 scores than those without FM [73]. When studies reported individual components Understanding the Mechanisms of Pain in Rheumatoid Arthritis DOI: http://dx.doi.org /10.5772/intechopen.93829 of the DAS28, patients with co-existent FM had significantly higher tender joint counts and higher patient global assessment scores than those without FM [73]. Objective measurements including swollen joints and inflammatory markers were not significantly different between RA patients with and without FM [73]. Other scoring systems including the simplified disease activity index (SDAI) and the clinical disease activity index (CDAI) were also higher in RA patients with comorbid FM [75]. A large study of 11,866 RA patients reported that those with comorbid FM had increased pain, poorer quality of life and greater functional limitation [76].
The recognition of FM in RA patients is important for multiple reasons. Firstly, both RA and FM commonly present with pain and fatigue. Differentiation of the conditions and diagnosis of co-morbidity is vital for patient management as different treatment approaches are required. In FM-RA patients, characterisation of the pain is imperative to manage patients appropriately. For example, inflammatory flares must be differentiated from painful flares secondary to FM. Secondly; recognition of patients with secondary FM offers an important insight into the pathogenesis of pain centralisation which is currently poorly understood. Nevertheless, caution should be used when interpreting the association between inflammatory arthritis and FM as the FM diagnostic tools have not been validated in RA. Furthermore, confounding factors including female sex and mental health problems are more prevalent in both FM and RA.

Central mechanisms of pain sensitisation
Central pain sensitisation is thought to occur at both spinal and supraspinal levels [62]. At the level of the DRG, spinal hyperexcitability occurs secondary to ongoing nociceptive input and pain transmission can be modified by inhibitory or facilitating neurones that can be modulated by descending signals from supraspinal levels [72]. Spinal pain facilitation is thought to be responsible for the spread of mechanical allodynia beyond the innervated field of cutaneous neurons. This has been shown in models using the intradermal injection of capsaicin [77]. Both ipsilateral and contralateral pain facilitation is thought to occur secondary to chronic inflammation and in RA patients, enhanced responses to noxious stimulation occurs at sites distal from inflamed joints [78]. Pain sensitisation is also thought to occur at supraspinal levels and brain imaging has demonstrated changes in cerebral activation secondary to chronic pain. Throughout the next section of the chapter, we will discuss the evidence of supraspinal pain sensitisation in RA.

Brain neuroimaging and pain
Imaging studies have attempted to characterise the neuronal circuitry resulting in cerebral sensitisation in RA. Structural MRI studies have shown increased grey matter density in the basal ganglia of RA patients compared to controls. This area is involved in both motor function and in pain processing [79]. Functional imaging has been used to investigate neuronal activation in response to pain. Functional MRI (fMRI) studies have demonstrated differences in resting state functional connectivity between RA patients and controls. In RA patients, there is increased connectivity between frontal midline regions that are implicated in pain processing, including the supplementary motor area and the mid-cingulate cortex, to sensorimotor regions [80]. Moreover, in RA patients, increased EEG activity has been reported in response to repetitive painful stimuli [81]. These studies suggest that aberrant pain cerebral pain processing may occur in RA and therefore, may result in augmented pain responses.
A further level of complexity is introduced when the biopsychosocial model of pain is considered. This suggests that cognitive and emotional processes are also critical contributors to the overall perception of pain. Indeed, the transmission of nociceptive information is influenced by multiple higher-level factors, such as mood, attention and cognitive factors, to form the resulting pain experience [82]. Mood is a particularly important cognitive factor in RA and meta-analysis has revealed that 16.8% of patients with RA meet the criteria for a major depressive episode [83].
In RA patients, depressive symptoms have been found to correlate significantly with tender joint count [84]. The medial prefrontal cortex has been suggested to play an important role in mediating the relationship between pain severity and depressive symptoms. Evidence has demonstrated an association between depressive scores (measured using the Becks depression index), tender joint count and MPFC activation during provoked joint pain. In the same study, MPFC activation co-varied significantly with limbic activation, an area involved in affective processing. This led the authors to suggest that the MPFC engages areas important for self-relevant processing to mediate the relationship between pain and affective symptoms [84]. In summary, pain processing by higher brain centres affects pain perception and the affective response to pain in RA. Although we are beginning to shed light on higher processing using functional imaging studies, more work is required to fully appreciate the complexities of central pain processing in RA.

Management of pain in RA
The cornerstone of RA treatment is the suppression of inflammation using the treat to target approach. However, disease remission will not lead to the complete resolution of pain in all patients and a multi-modal approach to pain management is very important. This approach has been recommended by rheumatology associations. For example, EULAR have recommended a patient centred approach to pain management where a biopsychosocial framework should be adopted [85]. Specifically, clinicians should differentiate between local and generalised pain and should be guided by patient needs, preferences, pain characteristics, inflammation and psychological factors. Treatments should include education, psychical therapies, orthotics, psychosocial interventions, sleep hygiene, pharmacological and joint-specific treatment options. Throughout this section of the review, we will discuss the different facets of pain management.

Pharmacological therapies
Pharmacological treatments include analgesic agents and immunomodulatory medications. Many analgesic agents are used in the management of RA pain although their use is rarely supported by high-quality RCTs [62]. Commonly used analgesic medications include paracetamol, NSAIDs, opioids and tricyclic anti-depressants. Optimal pain management should involve the characterisation of pain phenotype, in particular, differentiation of peripheral and central pain mechanisms. Pain phenotype could alter the choice of analgesic agent. For example, NSAIDs have been shown to reduce inflammatory pain in RA but not central pain in FM [86]. More work is required to define optimal analgesic use in different subsets of RA patients.
The cornerstone of RA management is the suppression of inflammation. Medications that reduce synovial inflammation are well known to reduce pain in RA patients. Immunomodulatory medications used in RA include glucocorticoids, conventional synthetic DMARDs and biologic DMARDs. Glucocorticoids are commonly used to treat acute inflammatory flares and have been shown to provide significant pain relief [87]. Extensive evidence supports the efficacy of traditional DMARDs, including methotrexate, sulfasalazine and leflunomide, in reducing joint pain. The analgesic effect of cDMARDs parallels the suppression over a time course of weeks to months [62]. Combination therapy has been shown to be superior than monotherapy and the addition of a biologic agent has been shown to reduce pain even further [88,89]. Nevertheless, pain improvement may plateau despite effective suppression of inflammation and studies have shown that this plateau is worse that the UK mean [7]. Persisting pain may result from centrally mediated pain hypersensitivity and may respond better to neuropathic agents or non-pharmacological treatments including education, exercise and cognitive behavioural therapy (CBT) than those treatments focusing on management on nociceptive triggers alone.

Neuropathic agents
Neuromodulatory medications used for the treatment of neuropathic pain include antidepressants such as tricyclic antidepressants (e.g. amitriptyline) and serotonin-noradrenaline re-uptake inhibitors (e.g. duloxetine) or anti-convulsants, e.g., pregabalin or gabapentin [90]. The clinical efficacy of these medications well well-established in conditions including neuropathic pain and generalised pain sensitisation syndromes such as fibromyalgia [91,92]. Neuropathic agents are sometimes used for the treatment of pain in RA however evidence from high quality RCTs is lacking [93]. However, in other localised pain conditions such as hand OA, pregabalin has been shown to improve pain and function [94]. Further work is required to establish the role for neuropathic medications in RA patients.

Psychosocial therapies
Psychological pain management programmes, including cognitive behavioural approaches and mindfulness, have an important role in the management of chronic pain. An abundance of evidence supports the efficacy of psychosocial approaches to pain management in chronic pain conditions [95]. In RA, CBT has the best evidence base for the management of pain with multiple meta-analyses confirming efficacy [96,97]. In addition to benefiting pain symptoms, CBT has been shown to improve other symptoms including fatigue in RA patients [98]. Psychosocial therapy may be most efficacious when offered early in the disease course however further work is required to determine which subset of patients should be offered psychosocial therapies and at which time-point in their illness [99].

Exercise based therapies
Exercise based therapies have an important role in the management of RA. Evidence has shown that resistance exercises decrease disability and functional impairment [100]. Furthermore, a meta-analysis of five studies revealed that resistance exercises resulted in a trend towards a small positive effect on VAS pain [100].

Conclusion
In conclusion, pain remains a significant problem for many patients with RA and is associated with psychological distress, fatigue and reduced quality of life.
In RA patients, pain results from a combination of joint inflammation, structural joint changes and pain sensitisation. In order to treat patients effectively, it is vital to differentiate between different types of pain, as each type should be targeted differently. Effective pain management approaches using a multimodal approach are vital to increase patient well-being, functioning and to reduce individual and societal costs [85].