Decoding pain in axial spondyloarthritis: Mechanisms, management, and future directions

Abstract

Axial spondyloarthritis (axSpA) is an immune-mediated inflammatory condition that is associated with significant disease burden. While the use of biologic and targeted synthetic disease modifying antirheumatic drugs have led to significant improvements in disease outcomes, a substantial proportion of patients continue to face challenges with pain despite effective control of their inflammatory state. Chronic pain is one of the leading causes of disability and remains an unmet research and clinical need in axSpA. Phenotyping pain is essential to understand the etiology of symptoms and to aid in selecting the appropriate management plan in axSpA. In this review, we aim to highlight the mechanisms of pain, its phenotypes, and novel assessment tools. We also review our approach to managing chronic pain in axSpA and discuss future directions in this field.

1. Decoding pain in axial spondyloarthritis: mechanisms, management, and future directions

1.1. Axial spondyloarthritis brief overview

Axial spondyloarthritis (axSpA) is an immune-mediated inflammatory condition that predominantly affects the axial skeleton, resulting in pain and functional limitations. The estimated prevalence of axSpA is 1.4% of the US population [1,2]. Sacroiliitis is a hallmark of the disease which presents with features of inflammatory back pain (IBP) in the vast majority of axSpA patients [3]. IBP symptomatology differs from mechanical back pain (MBP) manifesting with prolonged morning stiffness, improvement of pain/stiffness with exercise, and lack of improvement with rest. Advances in the field of axSpA have led to the discovery of new pathogenic mechanisms and novel targeted therapeutics. While these advances have contributed to improvements in symptoms and disease outcomes [4–7], a substantial proportion of patients continue to face challenges with pain despite effective control of their inflammatory disease. More than 60% of patients do not achieve inactive disease state or remission according to composite disease activity scores. This is evident in trials of tumor necrosis factor inhibitors, interleukin-17 inhibitors, and janus kinase inhibitors in patients with axSpA [8–12]. AxSpA patients have significant disease burden [13–16], and chronic pain is one of the most common contributing factors. In September 2023, the fourth Spondyloarthritis Unmet Research Needs conference took place at the NIH [17], and one of the main unmet research topics was pain in axSpA.

1.2. Pain categories

The International Association for the Study of Pain (IASP) defines pain as an unpleasant sensory and emotional experience, that is sometimes associated with actual tissue damage, and is influenced by biological and psychosocial factors [18]. IASP groups pain into 3 categories: nociceptive, neuropathic, and nociplastic. Nociceptive pain arises from an insult to non-neural tissue, which can occur when inflammation in the musculoskeletal system leads to activation of nociceptors. In axSpA, sacroiliitis and peripheral arthritis are key drivers of nociceptive pain. Neuropathic pain is pain caused by a lesion to the somatosensory nervous system. Pain arising from carpal tunnel syndrome and spinal radiculopathy is an example. Nociplastic pain is defined by the IASP as “pain that arises from altered nociception despite no clear evidence of actual or threatened tissue damage causing the activation of peripheral nociceptors, or evidence for disease or lesion of the somatosensory system causing the pain” [19]. Fibromyalgia (FM) is the prototypical nociplastic pain condition. This pain category encompasses pain from central and peripheral mechanisms. Central mechanisms of nociplastic pain include enhanced ascending pain facilitation and inefficient descending pain inhibition. Central sensitization (CS), a key mechanism [20], is a neurophysiologic mechanism that reflects an exaggerated central nervous system (CNS) neural signaling which elicits pain hypersensitivity [21]. Neural circuits involved in pain processing including the ascending facilitatory pathway (from the periphery to the brain), intra-cranial connectivity between different brain networks for modulation, and the descending inhibitory pathway (from the brain back to the periphery) to inhibit the pain signal (Fig. 1). Abnormalities at any of these different stages of pain signaling can lead to pain amplification.

Fig. 1.

Neural circuits involved in pain processing. The ascending facilitatory pathway originates from the periphery (sacroiliitis in axSpA patients), sending the pain signal through the dorsal horn of the spinal cord to the brain where it gets processed intracranially. The descending pathway then gets activated sending inhibitory signals from the brain through the spinal cord and back to the periphery. As shown in the figure, different cytokines and neurotransmitters play different roles in the ascending/descending pathways, which make targeting these markers a possible therapeutic option. SN: Salience Network, DMN: Default Mode Network, GABA: gamma-aminobutyric acid.

1.3. Neurobiology and neuroimmune pathways of pain

Acute nociceptive pain arises when peripheral nociceptors (Aδ or C-fibers) sense chemical, mechanical, or thermal stimuli that exceed a certain threshold. These nociceptors release neurotransmitters (e.g., substance P, glutamate, calcitonin gene-related peptide) from the dorsal root ganglia, which activate second order neurons in the dorsal horn of the spinal cord [22]. The signal then travels through ascending tracts of the spinal cord to activate third-order neurons in the thalamus and brainstem. The signal is then processed and distributed to different brain networks responsible for sensory, emotional, cognitive, and motor responses. After the signal is processed, the brainstem may activate the descending inhibitory pathway to modulate incoming peripheral signals [23].

In axSpA and individuals with other inflammatory arthritides, immune-neural interactions may amplify pain signaling. Peripheral joints including the SIJ are rich in nociceptors. When inflammation in the joints commences, cytokines such as TNF, IL-17, IL-1, IL-5, and IL-6 directly activate the nociceptors triggering the ascending pain pathways [24]. In addition, immune cells promote the production of nerve growth factors which increases local innervation, leading to peripheral sensitization [25].

Central sensitization is believed to involve higher order immune cells in the CNS, such as microglial cells, glial cells, astrocytes, and oligodendrocytes. Microglial cells are the macrophages of the CNS and are thought to orchestrate the interaction between brain-derived neurotrophic factor and tropomyosin receptor kinases in the dorsal horn of the spinal cord which facilitates neuronal activity [26,27]. This has been described in animal models of rheumatoid arthritis and osteoarthritis [28].

There has been reports on the possibility of Janus Kinase/signal transducers and activators of transcription (JAK-STAT) pathway playing a role in non-inflammatory pain modulation. In rat models of diabetic neuropathy [29], peripheral and central sensitization can be promoted by the phosphorylation of N-methyl d-aspartate (NMDA) receptors in the microglia of the dorsal horn of the spinal cord. Administration of a JAKi led to attenuation of pain by downregulating NMDA receptors. Another study showed that, in animal glial cells of the dorsal horn of the spinal cord, inhibition of the JAK-STAT signaling led to the expression of a suppressor of cytokine signaling-3, which lowered IL-6 and CCL2 levels leading to reduction of mechanical allodynia [30].

1.4. Chronic pain and its burden in axSpA

To engage persons with lived experience, we collaborated with the Spondylitis Association of America (SAA) to conduct a survey identifying what patients see as research priorities for the years 2025–2028. The survey, which received 149 responses, highlighted chronic pain as the top (#1) research priority (71.14%) (unpublished data). This was followed by inflammatory therapeutic targets (49.66%) and extra-musculoskeletal manifestations/the role of diet (48.32%) (unpublished data).

Several cohort and observational studies addressed chronic pain in axSpA, from a lens of central sensitization, fibromyalgia, or neuropathic pain. The prevalence of chronic pain differed between studies, taking into consideration that the definitions of chronic pain were very heterogeneous. A meta-analysis by Jones et al. included 16 studies and a total of 5214 patients and estimated the prevalence of fibromyalgia (based on ACR 1990 criteria) to be 16.4% (95% CI 12.3 – 20.5%) [31]. In cohorts using the Fibromyalgia Rapid Screening Test (FiRST), the prevalence of fibromyalgia was higher, ranging between 17.4% and 37.8% [32–34]. A few studies have assessed symptoms of central sensitization, one of the mechanisms of nociplastic pain, using the Central Sensitization inventory (CSI) [35]. The CSI is a patient-reported outcome measure designed to identify symptoms of widespread pain or central sensitization [36–38]. A study from the Netherlands reported that 45% of the included participants with axSpA (n = 178) had a high probability of central sensitization [35]. Another study from Turkey (n = 116) reported 46.6% central sensitization in their axSpA cohort, based on the same subjective questionnaire [39]. Similarly, Karlibel et al. reported that 45.1% of the 82 participants had a component of central sensitization [40].

The burden of chronic pain in axSpA is well recognized. Several studies assessed outcome differences between axSpA patients with and without central sensitization or fibromyalgia, and reported similar results, showing worse disease activity scores (e.g., Bath Ankylosing Spondylitis disease activity index (BASDA) and Ankylosing Spondylitis disease activity score (ASDAS)), functional indices (e.g., Bath Ankylosing Spondylitis functional index (BASFI)), and quality of life indices (e.g., Ankylosing Spondylitis quality of life score (ASQoL)) [32–35,39–50]. There was a strong association between CS/FM and worse biopsychosocial factors such as anxiety, depression, sleep disturbances, absenteeism from work, and disability. Some studies evaluated the retention rates of biologic treatments in patients with axSpA, and concomitant FM or CS was significantly associated with lower drug retention rates [34,51]. Female patients were found to be disproportionately affected by chronic widespread pain, worse outcome measures, and lower drug retention rates.

Patients with axSpA face significant delays in disease diagnosis (~6.7 years on average), and this could influence the development of chronic pain. A study by Fitzgerald et al. showed that axSpA patients who met the FM criteria had significantly longer delay to diagnosis compared to patients who did not meet the criteria (7.4 vs 6.1 years) [33]. Another study by Mogard et al. reported that lower pain tolerance was associated with longer symptom duration [42]. Lee et al. showed that in conditions with peripheral inflammatory arthritis (such as rheumatoid arthritis), the incidence of developing chronic pain is highest within the first 2 years after symptom onset [52]. This could be potentially amplified in a condition where the delay in diagnosis can extend beyond 10 years as in axSpA.

1.5. Outcome measures in axSpA

Biologic (b-) and targeted synthetic (ts-) disease modifying antirheumatic drug (DMARD) clinical trials showed that more than 60% of axSpA patients do not achieve inactive disease state or remission according to composite disease activity scores [8–12]. This is possibly due to the contribution of non-inflammatory pain. The clinical scores currently used in studies (e.g. ASDAS-inactive disease or ASAS partial remission) rely heavily on patient-reported measures. They include items of patient global assessment, pain rating, and subjective questionnaires. Given the lack of objective measures of inflammation, these patient-reported outcomes can be affected by non-inflammatory components, such as nociplastic pain. Molto et al. designed a prospective observational study where they followed 508 axSpA patients newly treated with a tumor necrosis factor inhibitor (TNFi) [32]. The results were reported at 12 weeks, and stratified by the presence or absence of fibromyalgia, screened by the FiRST tool. Comparing patients with vs without FM, both groups had similar normalization in c-reactive protein (CRP) levels, but the group with concomitant FM had persistently elevated disease activity scores and significantly less response based on ASAS outcomes and ASDAS scores (p-value <0.001). Magnetic Resonance Imaging (MRI) of the sacroiliac joint (SIJ) is currently considered as the gold standard to detect active inflammatory lesions in the SIJ (such as bone marrow edema, capsulitis, enthesitis, and joint effusion). An observational study by MacKay et al. showed that sacroiliac joint MRI scores to assess disease activity have very weak or no correlation with clinical scores (r = 0.18 – 0.31) [53]. These findings suggest that there is a disconnect between clinical scores and true active inflammation, which necessitate revision of these scores and incorporating new objective measures of inflammation and chronic pain in axSpA clinical studies.

1.6. Phenotyping pain

Prolonged morning stiffness, alternating buttock pain, improvement with exercise, and lack of improvement with rest constitute inflammatory back pain (IBP) symptoms. A study by Poddubnyy et al. evaluated individual IBP symptoms and found that they lack specificity for the diagnosis of axSpA (specificity range 23 – 58%) [54]. Another study showed that only 15% of patients with IBP symptoms have axSpA [2]. This necessitates phenotyping pain in the clinic to determine the etiology of chronic pain symptoms. A dedicated workup should be performed to determine whether the pain symptoms are coming from a nociceptive, neuropathic, or nociplastic origin as the treatment strategy can differ significantly. Utilizing MRI and advanced imaging technologies can identify inflammatory lesions in the SIJ/spine or locations of nerve compression. Laboratory work including CRP levels can help identify active systemic inflammation in up to 60% of axSpA patients. Nerve conduction studies can sometimes recognize nerve conduction abnormalities.

Identifying nociplastic pain may be more complex given the lack of clinically validated objective measures. In 2010, the American College of Rheumatology published clinical criteria that can aid providers to diagnose fibromyalgia [55]. This includes the Widespread Pain Index (WPI) and Symptom Severity Scale (SSS), which, together, comprise the Fibromyalgia Survey Questionnaire (FSQ). Recent studies reported that the probability of nociplastic pain can be determined based on FSQ scores, which can be used as a surrogate for nociplastic pain when the score is more than or equal 13 [56–59]. Similarly, the Central Sensitization inventory (CSI) has been utilized to aid in detecting CS component of pain. Both measures remain subjective and lack objectivity. Quantitative sensory testing (QST) and functional neuroimaging emerged as possible modalities to assess pain objectively, though they remain largely experimental.

1.7. Quantitative sensory testing (QST)

QST is a tool that characterizes nociplastic pain mechanisms. Three commonly assessed QST modalities are pressure pain threshold (PPT), temporal summation (TS), and conditioned pain modulation (CPM) [60–62]. The PPT is performed using an algometer at articular and non-articular sites. It is defined as the pressure at which pain is first reported. Low PPTs at extraarticular sites represent increased pain sensitivity due to dysregulation of CNS pain mechanisms or other causes of widespread pain sensitivity. Low PPTs at joint sites reflect abnormalities in both central and peripheral processes, including sensitization of the peripheral nociceptor by joint inflammation [60]. TS can be tested using different pain modalities. One method is to use a mechanical stimulus (e.g., Neuropen) which is tapped 10 times on a non-articular site. TS is calculated by subtracting the pain score at the first tap from the pain score at the tenth tap. Higher TS values are indicative of higher levels of ascending pain facilitation [61,63]. CPM is assessed using a painful conditioning stimulus (e.g., hand submersion in cold water) to activate the descending inhibitory pain pathways and a test stimulus to assess pain sensitivity (e.g., PPT). Low CPM is indicative of abnormalities in descending pain inhibition [62]. Multisensory sensitivity, including light sensitivity, is a potential surrogate of central sensitization and facilitated CNS sensory processing [64]. A recent study on patients with chronic pain and history of traumatic brain injury demonstrated an association between photosensitivity and chronic pain and disability [65].

QST has been studied extensively in different conditions including rheumatoid arthritis [63,66,67], but not in axSpA. In rheumatoid arthritis studies, Heisler et al. [67] showed that central pain as evidenced by QST measures (low PPT and high TS) correlated with pain intensity, which was independent of inflammatory activity. Another study by the same group showed that baseline abnormalities in TS and CPM was associated with lower odds of achieving good response to DMARDs (22% vs 53% in those who did not have baseline abnormalities in TS and CPM) [67]. In conditions like chronic pancreatitis and diabetic neuropathy, QST was predictive of response to treatment with pregabalin and duloxetine, respectively [68,69]. TS was predictive of response to pregabalin in chronic pancreatitis [68], suggesting a role for gabapentinoids in controlling ascending pain facilitation. CPM was predictive of response to duloxetine in diabetic neuropathy [69], suggesting a possible role for serotonin-norepinephrine reuptake inhibitors in modulating the descending inhibitory pathway.

Mogard et al. enrolled axSpA patients in the SPARTAKUS cohort and showed that lower PPT and higher TS were associated with ‘unacceptable pain’ and worse clinical axSpA scores [42]. Another study by Yucel et al. showed that PPT at the sacroiliac joints and trapezius muscles were lower in axSpA patients with CS compared to those without CS (p < 0.05) [70]. No significant difference was found in TS measures. A study by van der Kraan et al. showed that patients who score ≥40 on the CSI had low PPT and high TS compared to patients who scored <40 on CSI (p-value <0.004) [71]. None of these studies stratified patients by the presence or absence of objective signs of inflammation which limits its interpretation in relation to the specific phenotype of pain.

1.8. Neuroimaging

Functional (fMRI) and anatomic brain imaging enable researchers to gain insights into the network of brain regions and mechanisms associated with chronic pain. Resting-state fMRI enables researchers to look at functional connectivity between different brain networks in chronic pain states. Chronic pain studies have shown significant altered connectivity between the default-mode network (DMN) and the salience network (SN). The DMN is a group of brain regions that are strongly correlated during periods when external stimulation is absent [72]. It includes the posterior cingulate cortex, medial prefrontal cortex, lateral parietal, and parts of the medial temporal lobe. The SN is associated with external stimuli and their perceived intensity. It includes the anterior insula, dorsal anterior cingulate cortex, temporoparietal junction, and the dorsolateral prefrontal cortex. Altered connectivity between the DMN and anterior insula of the SN have been observed in studies of fibromyalgia, osteoarthritis, and chronic back pain [73–75]. A rheumatoid arthritis study showed that DMN-insula connectivity was significantly associated with FSQ scores of nociplastic pain [76]. Harris et al. studied pregabalin in patients with non-specific chronic pain and showed that treatment can modulate brain connectivity [77], making fMRI a possible imaging biomarker.

A study of 20 axSpA patients reported that connectivity between the DMN and salience network was also altered in comparison to healthy controls [78]. The same group showed that patients with high clinical pain scores on the day of the study exhibited abnormal cross-network functional connectivity between the DMN and the sensorimotor network, and that there was an interaction between within-DMN connectivity and clinical pain reports [79]. A separate study of 27 axSpA patients vs controls also showed altered networks and correlation with inflammatory markers [80]. Osborne et al. studied functional connectivity in 38 axSpA patients (19 men and 19 women) with chronic low back pain and compared them to a sample of healthy controls [81]. Within the axSpA group, women had greater functional connectivity of the subgenual anterior cingulate cortex (sgACC) to the DMN and sensorimotor networks. When comparing axSpA women with chronic pain to healthy women, patients had less connectivity between sgACC and the hippocampus/frontal regions, and stronger connectivity with the precuneus.

Task-based fMRI studies of tactile stimulation provide a means of localizing somatosensory and pain responsive areas of the brain, as well as networks of brain areas involved in the anticipation and evaluation of neutral and painful stimuli (Fig. 2). There is paucity of task-based fMRI data in axSpA. A rheumatoid arthritis study applied mechanical pressure to the metacarpophalangeal joints in patients vs matched controls and showed increases in neural processing in the medial frontal cortex, and that was associated with clinical pain in RA patients [82].

Fig. 2.

An example of a task-based fMRI depicting BOLD signals in response to a thumb pressure stimulus. In this axSpA example, the stimulus led to activation (red signal) of salience network and inhibition (blue signal) of default mode network. fMRI: functional magnetic brain imaging, BOLD: Blood Oxygenation Level Dependent, axSpA: axial spondyloarthritis.

In addition to functional neuroimaging, structural brain imaging is helpful to assess gray matter volume and thickness which has been found to correlate with chronic pain. In patients with chronic low back pain and osteoarthritis, studies have shown that gray matter thickness was decreased in the prefrontal cortex and somatosensory cortex when compared to controls, and that gray matter volume increased after effective treatment [83,84]. One study from 2013 enrolled 17 patients with ankylosing spondylitis and showed gray matter cortical thinning in the primary somatosensory cortex, insula, and anterior cingulate cortex [85]. This cortical thinning was correlated with higher painDETECT scores, a measure to identify neuropathic pain.

All the aforementioned neuroimaging studies in axSpA included a small number of patients and enrolled only patients with radiographic axSpA, which excluded patients with non-radiographic disease. These studies did not include specific measures of chronic nociplastic pain such as FSQ/CSI scores or quantitative sensory testing, as studying chronic pain was not an aim for these studies.

2. Management of pain in axSpA

There are two major international guidelines (ACR/SPARTAN/SAA and EULAR/ASAS) that serve as a reference to clinicians on how to approach axSpA management [86,87]. These guidelines focus on the treatment of inflammation, including cycling/switching advanced biologic therapies in case of primary or secondary failure to the first agent. The lack of treatment response is determined by the treating provider based on global assessment of the case or the use of certain clinical disease activity scores such as (ASDAS-inactive disease and ASAS-partial remission). There is minimal guidance on how to phenotype pain in axSpA and virtually absent guidance on how to target non-inflammatory pain components.

There are 4 medication classes that treat axSpA: non-steroidal anti-inflammatory drugs (NSAIDs), tumor necrosis factor inhibitors (TNFi), interleukin-17 inhibitors (IL-17i), and janus-kinase inhibitors (JAKi). As discussed previously, ~60% of patients enrolled in clinical trials continued to have symptoms of pain and did not achieve remission. Only until recently, the concept of ‘non-nociceptive’ pain has been gaining more attention in axSpA studies. At the 2025 American College of Rheumatology conference, an abstract addressed the effects of Upadacitinib (a JAK1 inhibitor) on non-nociceptive pain in axSpA [88]. This was a post-hoc analysis from the previously published multicenter observational UPSTAND study. Patients were stratified based on the presence or absence of nociplastic and neuropathic pain, according to the Widespread Pain Index (WPI) and painDETECT questionnaires, respectively. The study showed improvements in pain scores across different groups, but the study did not correlate with the presence or absence of objective signs of inflammation, and excluded patients with a clinical diagnosis of fibromyalgia, which limits the interpretation of the study results.

Opioid use disorder has been widely recognized as a national epidemic leading to negative outcomes. The 2019 ACR/SPARTAN/SAA treatment guidelines recommended against the use of opioids in axSpA. Despite this, epidemiologic studies show alarming numbers of opioid use in axSpA (ranging between 23% and 37%) [89–91]. It was replicated in different studies that opioid use was associated with non-inflammatory aspects such as smoking, substance use disorder, depression, history of self-harm, and BASFI, but not with objective measures of inflammation such as CRP levels [92,93]. This necessitates more intensive education on the harms of opioids, and that the focus should shift to targeting the underlying driving factors of non-inflammatory pain such as anxiety, depression, and sleep disturbance.

There have been no published randomized clinical trials (RCT) on the efficacy of gabapentinoids, serotonin-norepinephrine reuptake inhibitors (SNRIs), tricyclic antidepressants (TCAs), or muscle relaxants such as cyclobenzaprine in patients with axSpA. The use of these agents in clinical practice is extrapolated from studies on other chronic pain conditions such as fibromyalgia (prototypical nociplastic pain condition), osteoarthritis, chronic pancreatitis, and diabetic neuropathy. A meta-analysis published in 2009 showed that pregabalin and gabapentin were associated with reduction of pain (SMD −0.28, 95% CI −0.36, −0.20; p < 0.001), improved sleep (SMD −0.39, 95% CI −0.48, −0.39; p < 0.001), improved health-related quality of life (SMD −0.30, 95% CI −0.46, −0.15; p < 0.001), reduction of fatigue (SMD −0.16, 95% CI −0.23, −0.09, p < 0.001) and reduction of anxiety (SMD −0.18, 95% CI −0.27, −0.10; p < 0.001) in patients with fibromyalgia compared to placebo [94]. One case-control study assessed the efficacy of duloxetine (SNRI) in axSpA with concomitant anxiety and depression [95]. Patients who received duloxetine had significantly less spinal pain, BASDAI, and BASFI scores compared to the control group. Several systematic reviews and meta-analyses demonstrated the efficacy of SNRIs such as duloxetine and milnacipran in patients with fibromyalgia [96,97]. The use of muscle relaxants in chronic pain syndromes including fibromyalgia was evaluated in several studies. A 2004 meta-analysis assessed 5 RCTs of oral cyclobenzaprine in fibromyalgia; it showed that individuals treated with cyclobenzaprine had 3 times improvement in pain symptoms, and moderate benefits on sleep [98]. A 2024 systematic review showed that long-term (>4 weeks) use of muscle relaxants including cyclobenzaprine was not beneficial in fibromyalgia and was associated with increased odds of adverse events such as sedation and dry mouth [99]. A recent systematic review and meta-analysis showed that sublingual cyclobenzaprine led to significant improvements in daily diary pain intensity (SMD = −0.25; 95% CI [−0.34, −0.16]; p < 0.001), fibromyalgia impact questionnaire function (SMD = −0.18; 95% CI [−0.35, −0.02]; p = 0.028), and sleep daily diary (SMD = −0.21; 95% CI [−0.32, −0.11]; p < 0.001), when compared to placebo [100]. Sublingual cyclobenzaprine was approved by the U.S. Food and Drug Administration (FDA) to treat fibromyalgia in August 2025 [101].

Non-pharmacologic modalities including physical therapy, yoga, meditation, and mindfulness remain very important adjuvant treatment options. A recent RCT evaluated physical therapy in axSpA patients with controlled inflammation and showed that the intervention was associated with significantly lower pain scores (an average of 13.71 points less on a 0–100 pain scale (CI: improvement by 10.31–17.11 points)) compared to patients not receiving physical therapy [102]. Several RCTs addressed yoga in axSpA, and showed significant improvements in BASDAI, BASFI, ASQoL, and mental health outcomes [103–105].

We recommend a wholistic approach to the management of patients with axSpA (see Fig. 3), incorporating pain phenotyping, controlling inflammation, and targeting non-inflammatory pathways as well as comorbid conditions.

Fig. 3.

Our proposed approach to the management of patients with axial spondyloarthritis.

NSAID: non-steroidal anti-inflammatory drug. B/tsDMARD: biologic or targeted synthetic disease modifying anti-rheumatic drugs. MRI: magnetic resonance imaging. CRP: c-reactive protein level. SNRI: serotonin-norepinephrine reuptake inhibitors. TCA: tricyclic antidepressants. PT: physical therapy.

^#The potential benefit of gabapentinoids in controlling ascending pain facilitation is extrapolated from other chronic pain conditions such as the study on chronic pancreatitis [68].

*if controlled axSpA: either continue with treatment regimen or consider tapering b/tsDMARDs on a case-by-case basis and following a shared-decision making approach.

‡Reassess involves rethinking about the diagnosis to ensure it’s a correct diagnosis and not a mimic. If condition is accurate, consider repeating MRI of the sacroiliac joint or of the affected location to rule out active inflammatory lesions.

**If there is no active inflammation, consider adding SNRI or TCA, unless these agents are needed prior to target other comorbidities. The potential benefit on descending inhibitory pathways is extrapolated from fibromyalgia and diabetic neuropathy studies [69]. Gabapentinoids could be considered, if not utilized earlier.

†Non-pharmacologic options and screening for comorbidities should be implemented starting at the time of diagnosis and throughout the patient’s journey.

^Short-courses of cyclobenzaprine can be considered, if needed, throughout the patient’s journey to assist with pain control.

3. Future directions

Chronic pain in axSpA remains markedly understudied, and much of our current understanding is extrapolated from work done on other rheumatic or chronic pain conditions. There is a significant unmet need in pain research in the field of axSpA. Experimental studies to examine the pathophysiology of pain in axSpA animal models are essential to clarify the biological mechanisms that drive pain. Similarly, mechanistic human studies of pain in axSpA are warranted to help delineate the etiology of symptoms and better assist with evidence-based phenotyping pain.

A deeper exploration of the interaction between the immune system and the neural circuitry, including the role of brain processing, will be central to advancing precision medicine in axSpA. Importantly, epidemiological studies show that a significant percentage (up to 45%) of axSpA patients have chronic pain despite inflammation control; and despite that, most clinical trials exclude patients with concomitant fibromyalgia and do not include outcome measures to assess non-nociceptive pain phenotypes. We encourage researchers and clinical trialists to start implementing nociplastic and neuropathic pain outcome measures in future study designs.

It is of utmost importance to develop a management algorithm that includes pain phenotyping and multi-modal strategies (pharmacologic and non-pharmacologic) to target inflammation, non-inflammatory pain pathways, and comorbid conditions that precipitate pain and contribute to worse patient outcomes (e.g. anxiety, depression, and sleep disturbance). The traditional model of axSpA focuses on inflammation which leads to damage and disability (Fig. 4) [106] but does not take into account nociplastic pain or other comorbid conditions that severely impact the patients’ quality of life. As discussed, while targeting inflammation has improved outcomes for many, it did not fully control symptoms in the majority of axSpA patients. Our proposed model, and the direction of current and future research, broadens the clinical framework to explicitly incorporate assessments of nociplastic pain and psychosocial factors into the management of patients with axSpA (Fig. 4).

Fig. 4.

Traditional (a) and proposed (b) axSpA models.

Practice points:

• Chronic pain continues to be one of the most common causes of morbidity in axial spondylorthritis.

• Phenotyping pain into nociceptive, nociplastic, and neuropathic can help the clinician understand the etiology of the symptoms.

• Quantitative sensory testing and functional neuroimaging represent novel experimental instruments to assess chronic pain in axial spondyloarthritis.

Research Agenda:

• Understanding the biological mechanisms of pain and exploring the interaction between the immune system and the neural circuitry are central to advancing precision medicine in axSpA.

• It is of utmost importance to develop a management algorithm that includes pain phenotyping and multi-modal strategies (pharmacologic and non-pharmacologic) to target inflammation, non-inflammatory pain pathways, and comorbid conditions that precipitate pain and contribute to worse patient outcomes (e.g. anxiety, depression, and sleep disturbance).