Taiwan Rheumatology Association consensus recommendations for the management of axial spondyloarthritis

Abstract Aim To establish guidelines for the clinical management of axial spondyloarthritis that take into account local issues and clinical practice concerns for Taiwan. Method Overarching principles and recommendations were established by consensus among a panel of rheumatology and rehabilitation experts, based on analysis of the most up‐to‐date clinical evidence and the clinical experience of panelists. All Overarching Principles and Recommendations were graded according to the standards developed by the Oxford Centre for Evidence Based Medicine, and further evaluated and modified using the Delphi method. Results The guidelines specifically address issues such as local medical considerations, National Health Insurance reimbursement, and management of extra‐articular manifestations. Conclusion It is hoped that this will help to optimize clinical management outcomes for axial spondyloarthritis in Taiwan.


| INTRODUC TI ON
In recent years, several guidelines regarding the management of axial spondyloarthritis (axSpA) have been published, [1][2][3][4][5][6] predominantly by societies and experts in Europe and the USA. These guidelines represent a distillation of current knowledge on axSpA, and can provide valuable guidance to clinicians; but evidence shows that the genetic features of axSpA may vary between Asian and Caucasian patients, [7][8][9][10][11] and when clinical issues such as limited access to biologics, limited reimbursement for treatment, limited awareness, and under-diagnosis are taken into account, it is clear that local perspectives are needed to improve the management of axSpA. Moreover, the incidence and prevalence of tuberculosis, 12 hepatitis B, 13 and hepatitis C 14 are higher in Taiwan as compared to Europe or the USA, and this may limit treatment options for Taiwanese patients, particularly regarding the use of biologics. Therefore, aspects of axSpA that have local relevance were discussed in these guidelines, and recommendations with an emphasis on improving awareness, diagnosis, management, and outcomes in Taiwanese patients were formulated. It is hoped that these guidelines will help to focus attention on under-addressed issues in the management of axSpA, and bring a fresh perspective to the current discussion.
Axial spondyloarthritis is a chronic type of arthritis that primarily affects the sacroiliac joints and the spine. 7 Since the publication of the 2009 Assessment of SpondyloArthritis international Society (ASAS) classification criteria, 15 axSpA has been categorized into radiographic axSpA, which is largely synonymous with ankylosing spondylitis (AS) and presents with radiographically visible structural damage to the sacroiliac joint and axial skeleton; and non-radiographic axSpA (nr-axSpA), a milder form of axSpA that does not exhibit such structural damage but nevertheless imposes a heavy burden of disease. 7,15,16 It has been proposed that the term "axSpA" should preferentially be used in diagnosis rather than nr-axSpA or AS, 16,17 unless medical reasons exist to justify making a distinction. 7,16 In the spirit of this, the term "axSpA" in these guidelines encompasses both nr-axSpA and AS.
Importantly, these guidelines seek to address less explored issues in axSpA that are important for clinical management from both a local and global perspective. There is a recommendation discussing the management of extra-articular manifestations (EAM), primarily uveitis, psoriasis, and inflammatory bowel disease (IBD) but also encompassing other conditions that affect the lungs, kidneys, and heart of axSpA patients, with best management practices mentioned where supported by evidence. Osteoporosis and the risk of spinal fractures has been noted, since motorcycles and bicycles are one of the major modes of transport in Taiwan and can increase fracture risk, which is a serious concern as such fractures are difficult to recover from and may incapacitate a patient for life in worst-case scenarios. Regarding treatment, recommendations for exercise have been broadened to include evidence for yoga, Tai Chi, qigong, and other types of exercise that are common in Taiwan. The latest clinical trial data for novel therapies such as interleukin-17 inhibitors (IL-17i) have also been included. It is hoped that the discussion of these issues will help to provide practical and relevant evidence-based guidance to clinicians in Taiwan and beyond.

| MATERIAL S AND ME THODS
The formulation of these guidelines was undertaken by a committee of rheumatology and rehabilitation experts on behalf of the Taiwan Rheumatology Association (TRA). The structure of the guidelines was modeled on the recently published 2016 update of the ASAS-European League Against Rheumatism (EULAR) management recommendations for axial spondyloarthritis, 1 and also incorporated elements from the UK National Institute for Health and Care Excellence (NICE) 2017 guideline (NG65) on the diagnosis and management of SpA in over 16s, 3

and the British Society for Rheumatology (BSR) and British
Health Professionals in Rheumatology (BHPR) guideline for the treatment of axSpA (including AS) with biologics. 4 The objective was to establish guidelines for the clinical management of axSpA from a local perspective that would take into account issues and concerns in clinical practice that are pertinent to Taiwan. The guideline committee was convened by W.-CT and included 15 experts in rheumatology and rehabilitation, with all members required to disclose any potential conflicts of interest prior to taking up their positions. Each committee member was assigned to conduct a systematic literature review for a specific section of these guidelines, with a special emphasis on recent studies published between 2015 and 2017 and issues not addressed by other guidelines. Collected evidence was presented to the committee for review and discussion by all members, and based on these discussion results, each member prepared the wording for the recommendations and accompanying statements in their responsible section. The recommendations and statements were then presented to the committee for discussion, voting, and revision based on the Delphi method, with a threshold of 75% required for approval of each recommendation and statement. The levels of evidence, grades of recommendation, and levels of agreement were then added to each recommendation (Table 1). Briefly, level Ia refers to evidence derived from the meta-analysis of randomized controlled trials; level Ib refers to evidence derived from at least one randomized controlled trial; level IIa refers to evidence derived from at least one controlled study without randomization; level IIb refers to evidence derived from at least one type of quasi-experimental study; level III refers to evidence from comparative, correlation, case-control, or other non-experimental descriptive studies; and level IV represents evidence from expert committee reports, opinions, or clinical experience from respected authorities. Recommendations and statements based on direct level I evidence were graded as A; those based on direct level II evidence or extrapolated from level I evidence were graded as B; those based on direct level III evidence or extrapolated from level I or level II evidence were graded as C; and those based on direct level IV evidence or extrapolated from levels I, II, or III evidence were graded as D. Levels of agreement were derived through anonymous electronic voting at a committee meeting. The final guidelines and manuscript were reviewed and approved by all committee members, and were then reviewed and ratified by the TRA Executive Committee before submission to this journal.

| RE SULTS
These guidelines are intended for the use of all healthcare professionals involved in the management of axSpA, including rheumatologists, physiatrists, and clinicians of other disciplines. It is also important to explain these guidelines to patients and ensure their informed participation in shared decision-making regarding treatment and care. Considering that these guidelines are intended to focus on the management of axSpA, other aspects of the disease such as classification, diagnosis, and pathogenesis will not be discussed, unless they are relevant to treatment decisions.
As both axSpA and nr-axSpA are relatively new concepts, [15][16][17] it is inevitable that a significant proportion of the evidence in these guidelines was derived from AS patients. However, efforts have been made to identify and include evidence from studies conducted in axSpA patients, and although the term "axSpA" is used in these guidelines to refer to all patients across the spectrum of disease, the terms "nr-axSpA" and "AS" are also used where circumstances call for greater specificity. These guidelines also address the management of EAM and other comorbidities (eg osteoporosis and fractures) in axSpA, and although the guideline committee recognizes that these issues are complex and may even deserve their own guidelines, it is important for healthcare professionals to take these into account when advising patients, selecting treatment, and evaluating risks such as drug-drug interactions.
As with the ASAS-EULAR guidelines, 1 these guidelines begin with a set of overarching principles that are meant to be kept in mind throughout the management of axSpA. These principles define the main considerations, influencing factors, and best approaches regarding axSpA care in Taiwan today. The overarching principles and background statements are presented below, along with their respective levels of evidence (LoE), grade of recommendation (GoR), and levels of agreement (LoA; see Table 1).
Overarching Principle 1: The rheumatologist serves as the main coordinator of care for axSpA, a disease with diverse manifestations that is best managed through multidisciplinary care. (LoA: 100%) In a recent meta-analysis of eight studies (seven longitudinal cohort studies and one cross-sectional study) involving 1242 nr-axSpA patients and 2236 AS patients, 18 20 Populationbased studies conducted in thousands of Taiwanese AS patients also found higher risk of hypertension, 21 acute coronary syndrome, 22 and peptic ulcers 21 as compared to the general population. Taken together, these studies show that axSpA encompasses a diverse range of manifestations and comorbidities that require an integrated, multidisciplinary approach to effectively manage them. The selection of suitable treatment particularly requires an integrated approach in building a regimen that effectively checks musculoskeletal symptoms, EAM, and related comorbidities, all while avoiding potentially debilitating drug-drug interactions. In such a situation, the rheumatologist, having broad knowledge of the axSpA disease spectrum and the patient's condition, represents the ideal candidate to serve as the main coordinator of care with clinicians and health professionals in other specialties.
Rheumatologists should recognize this role and take a proactive approach in securing multidisciplinary support for axSpA patients to achieve better management outcomes. The rheumatologist serves as the main coordinator of care for axSpA, a disease with diverse manifestations that is best managed through multidisciplinary care.
The primary objective of axSpA treatment is to secure health-related quality of life and normalize function for the patient to the greatest extent possible.   36,37 Taken together, the evidence shows that axSpA patients face significant challenges, and the main goal of treatment should therefore be to restore and preserve quality of life and normal function for patients as much as possible.
It is also important to recognize that the cost of axSpA includes both direct costs related to treatment and indirect costs from loss of work productivity, inability to work, and restrictions in daily function, and all these costs should be considered when optimizing treatment for patients. A cross-sectional study has shown that axSpA patients who respond to treatment can achieve comparable health-related quality of life to the general population, and rates of activity impairment (33.3% vs 47.4%, P < .001) were also significantly lower than non-responders, indicating the importance of effective treatment. 24 When quality of life and normal function are achieved through effective management, patients will require less burden of care, and the overall socioeconomic costs are expected to be lower as well. Ideally, shared decision-making will help to bring patients, caregivers, and healthcare professionals together in working toward optimal care and outcomes.

Overarching
An important aspect of shared decision-making for Taiwanese axSpA patients involves family planning and pregnancy management, as many patients are of reproductive age and may face family pressure to conceive. 40 Family planning discussions can fill a key unmet need for both male and female patients. 41 Inflammation in axSpA has been associated with reduced sperm motility and impaired testicular function, 42 and small cross-sectional studies have reported that a majority of married or sexually active male patients felt that the disease had a negative effect on their sexual life, with issues such as low sex drive, premature ejaculation, sexual dissatisfaction, and impotence more frequently occurring. [43][44][45] In addition, a 2016 Swedish case-control study of 388 deliveries among AS patients compared to 1082 matched controls from the general population found that even after adjustment for smoking habits, age, education, and comorbidities, female AS patients had a higher risk of caesarean section, preterm birth, and small-for-gestational-age. 46 Prior to conception, adjustments to therapy may be needed to avoid miscarriage or congenital abnormalities; however, pregnant axSpA patients can still experience active disease, and therefore treatment may still need to be maintained during pregnancy. 47 Recommendations for patients with rheumatic diseases on the use of nonsteroidal anti-inflammatory drugs (NSAIDs), [48][49][50] conventional synthetic disease-modifying antirheumatic drugs (csDMARDs), 49,50 and tumor necrosis factor inhibitors (TNFi) [49][50][51][52][53][54] during pregnancy and lactation are currently available, but there is insufficient data to make any recommendations regarding the use of IL-17i at present. 55 Overarching NSAIDs and cyclooxygenase-2 (COX-2) inhibitors are covered by the NHI for the long-term treatment of inflammation and associated pain in AS patients. 59 In addition, csDMARDs are reimbursed for the treatment of peripheral symptoms in AS patients. 58 The NHI currently covers the use of biologics (bDMARDs, including TNFi and IL-17i) for the treatment of AS, but the reimbursement criteria is quite strict: biologics must be prescribed by clinicians licensed in the fields of rheumatology or immunology, and reimbursement must be applied for and approved prior to treatment initiation. In addition, the patient must be aged 18  >1 mg/dL in two consecutive tests, with at least a 4-week interval in between tests) and fails to respond to extensive treatment with at least two different NSAIDs (must have received continuous treatment at the same clinic or institution for 3 months or more, and must have used each NSAID for at least 4 weeks or more, unless discontinuation due to toxicity or tolerance occurs; those with peripheral symptoms must have undergone extensive treatment with at least two NSAIDs and sulfasalazine). 58 Furthermore, all patients must present a certificate indicating they have received exercise-related patient education (or an affidavit stating that they exercise regularly at home), and must sign a treatment consent form indicating they understand the indications, contraindications, and side effects of therapy. 58 AS patients who receive reimbursed biologics need to undergo BASDAI evaluation after 12 weeks of treatment, and must demonstrate >50% improvement or a decrease of at least two points to continue therapy; efficacy evaluations should subsequently be undertaken every 12 weeks for those who continue treatment. 58 The endemic presence of tuberculosis 12

and hepatitis B 13 in
Taiwan is a critical issue. The incidence and prevalence of tuberculosis in Taiwan are both considerably higher than that seen in the USA or Europe. 12  The burden of hepatitis B in Taiwan is also great, 13   can be used to evaluate prognosis. In addition, the 2009 ASAS classification criteria allows for diagnosis based on positive findings (eg bone marrow edema) from either X-ray or magnetic resonance imaging (MRI). 15 X-ray imaging can reveal syndesmophytes, but since radiologic progression is slow, 68 it is suggested that the interval between spinal X-rays should be no less than 2 years. The modified Stoke Ankylosing Spondylitis Spine Score (mSASSS) can be used to evaluate radiologic changes, but is mostly used for clinical research.
MRI can detect early inflammation before structural damage is radiographically visible, and a high degree of spinal inflammation observed on MRI has been correlated with a successful response to TNFi. 69 Moreover, signs of disease progression (inflammation, bone erosion, fatty change) can be visualized on MRI, and these can be used to support decisions on whether or not an X-ray to assess the presence of syndesmophytes will be needed. 69  In rheumatoid arthritis, T2T is clearly defined as either remission or low disease activity, but this is not the case for axSpA, which also lacks a definition for MDA. 72

| Recommendation 4
Patients with axSpA should be encouraged to stop smoking and start an individualized regular exercise program as soon as possible.
The program should emphasize flexibility training, especially spinal mobility exercises, but aerobic exercise, resistance training, breathing exercises, and physiotherapy are also recommended. A recent study from Taiwan observed a strong association between smoking and poor disease prognosis in AS patients, 78 and it has also been reported that axSpA patients who smoke have a reduced response to TNFi treatment 79 and greater disease activity. 80 Therefore, all axSpA patients should be actively encouraged to stop smoking upon diagnosis, and smoking cessation programs may facilitate this. Regular exercise is recommended in axSpA patients, and is associated with benefits such as improved joint mobility, reduced disease activity, and decreased cardiovascular (CV) risk. [81][82][83] It is important that exercise and physiotherapy programs be individualized so as to improve adherence and minimize the risk of injury. 84 week at moderate intensity). 86 A warm-up of 5-10 minutes is recommended before exercise, as is a cool-down of about 5-10 minutes after exercise, and intensity levels should be increased gradually, in order to minimize the risk of injury. 86 It is also important to avoid any intense strenuous exercise when there is a flare-up or acute inflammation. 86 Postural education and exercise are important to help patients avoid positions that lead to prolonged stooping. 86 85 There is also preliminary evidence that sug- have shown that EAM incidence is comparable in nr-axSpA and AS patients, 18 indicating that EAM is a valid concern even in patients with relatively mild disease. Uveitis is the most common EAM by far, with a reported pooled prevalence rate of 21.4% in Asian AS patients. 19 Acute anterior uveitis is the most frequent type associated with AS, and should be managed as an emergency to avoid complications; a Taiwanese study of 146 AS patients has also reported that acute anterior uveitis is associated with higher disease activity and poor functional ability. 88 Generally, axSpA patients with anterior uveitis should be referred to an ophthalmologist as soon as possible, as inflammation can lead to papillary and lens dysfunction with blurring of vision, and glaucoma and severe impairment of vision can occur in some cases if adequate treatment is delayed. 89,90 However, the prognosis is generally good following treatment with topical mydriatics, cycloplegics, and corticosteroids, and dilating drops (eg scopolamine) may be used to relieve pain. 89,90 Most acute anterior uveitis cases resolve spontaneously within 3 months. 90 It has been reported that sulfasalazine can reduce the incidence of uveitis flares 91,92 and the intensity of new flares, 92 and TNFi have been confirmed to reduce acute uveitis flares in AS patients as well. [93][94][95][96] However, although meta-analysis 93 and retrospective study 94  were not found to be effective in the treatment of Crohn's disease.
A study of 11 701 Taiwanese AS patients has observed a greater risk of comorbidities compared to the general population, 21 and EAM in the lung, kidney, and heart can also develop. 102 EAM in the lung typically include interstitial lung disease, apical fibrosis, and bronchiectasis, but the underlying pathophysiology remains unclear. It has been reported that up to 52% of AS patients demonstrate some type of lung involvement on high-resolution computed tomography (CT) scan. 103 Considering that many patients may have pathological lung changes that will not be discovered unless CT or biopsy tests are conducted, it is suggested that clinicians can be more proactive and conduct these tests to assess the presence and severity of lung EAM in axSpA patients when necessary. Although the prevalence of kidney EAM in AS patients is low, many patients (10%-35%) will have renal comorbidities such as amyloidosis and IgA-associated nephropathy, 21,90,104 which can significantly increase mortality risk. Evidence pertaining to treatment is limited to a report describing possible benefit for amyloidosis with long-term (>1 year) etanercept treatment. 105 It has been reported that 10%-30% of AS patients have heart pathologies (encompassing EAM and comorbidities), and the inflammatory environment of axSpA is known to increase CV risk in patients; this is a key cause of mortality. 106,107 TNFi may help to reduce inflammation and CV risk, but no direct evidence is available to date. It should be noted that NSAIDs can exacerbate CV risk both during treatment and within the immediate weeks after treatment cessation, 108  Regarding CV risk, a long-term placebo-controlled trial of celecoxib 400 and 800 mg/d (a dose not used in rheumatology practice) for the prevention of colonic adenomas demonstrated increased risk of CV events with treatment. 111 By contrast, celecoxib at 400 mg/d did not increase CV risk in two other longterm placebo-controlled trials, for the prevention of adenomatous polyps 112 and Alzheimer's disease. 113 In addition to these reports, two large meta-analyses comparing NSAIDs with placebo treatment revealed that COX-2 inhibitors, diclofenac, and ibuprofen all increased major vascular events to varying degrees, while only naproxen did not increase either major vascular events or mortality. 114,115 It should be noted that, compared to patients with autoimmune diseases, the patients recruited in the above trials did not have diseases with persistent inflammation. Given that chronic inflammation is recognized as a CV risk factor, the anti-inflammatory effect of NSAIDs may reduce the CV risk of autoimmune diseases, and such an effect might counterbalance the increased CV risk from NSAID treatment in arthritis patients.

| Recommendation 8
Although csDMARD monotherapy is not recommended for axSpA, it can be effective against peripheral arthritis and EAM; co-administration of csDMARDs with biologics may be beneficial in axSpA, but further evidence is needed to confirm this. In current practice, biologics are usually initiated with TNFi treatment, but there may be instances for which IL-17i might be a better starting biologic: for example, tuberculosis or HBV reactivation following TNFi treatment is a major concern, and the IL-17i secukinumab appears to have lower risk of tuberculosis reactivation, 61,62 although long-term real-world safety data will be needed to confirm this.

| Recommendation 10
Intra-or inter-class switching between biologics or small molecule therapies may be considered for patients with inadequate response or who become intolerant to therapy. secondary failure). 150 For cases of primary failure characterized by lack of response, it may be worthwhile to re-consider the diagnosis, as the ASAS-EULAR guideline task force has stated that true primary failure is rare in axSpA patients with active disease 1 ; however, patients with true primary failure or intolerance/toxicity to TNFi may benefit from switching to IL-17i. 1 For patients with secondary failure characterized by loss of response, intra-class switching can extend efficacy, but it should be noted that drug survival rates are generally lower for the second (47%-72%) and third (49%) TNFi at 2 years. 150,151 In patients who lose response to TNFi, switching to an IL-17i can be beneficial, but overall efficacy may be less than in TNFnaïve patients. 152 Switching for primary failure may be conducted after the patient has initiated one class of therapy and found to be unresponsive or intolerant within 3-6 months, while switching for secondary failure may be conducted if regular monitoring shows that response to treatment has decreased, or if anti-drug antibodies emerge. 1,150 As small molecule therapies emerge in the near future, it may be possible to consider switching to such therapies in the advent of treatment failure with biologics.
A recent study of 42 axSpA patients who underwent dose reduction of TNFi for 1 year found that 76.2% remained in remission or low disease activity at the end of the study, with shorter duration of remission before dose reduction, shorter duration of treatment with biologics, and shorter disease duration found to be risk factors for relapse. 153 Tapering or dose reduction of etanercept has been conducted successfully in AS patients with >6 months of stable disease, with acceptable efficacy in cases that needed to restart a full dose of treatment 154,155 ; however, in stable patients who completely discontinued etanercept 156 or adalimumab, 157 a greater percentage experienced disease flares as compared to controls. These results indicate that dose reduction may be carried out successfully in long-term stable patients, 158 but evidence for the specific timing of dose reduction and the effect of restoring full doses in relapsed patients remains limited for now.

| Recommendation 11
In patients with refractory pain or disability and radiographically visible structural damage of the hip joint, hip arthroplasty should be considered, while corrective osteotomy may be considered for patients with disabling spinal deformity.
(LoE: III; GoR: C; LoA: 100%) A study of three real-world datasets involving 2718 AS patients found that 24%-36% of patients presented with clinically significant hip involvement, and 5% of all pooled patients required hip surgery. 159 Patients who demonstrate radiographic structural damage and accompanying symptoms should be considered for hip arthroplasty regardless of age, and cementless prostheses are preferred in young patients. 1 Corrective osteostomy is highly specialized and should be undertaken in consultation with a spinal surgeon for patients with severe and disabling spinal deformities. Two recent retrospective studies respectively involving 13 160 and 12 161 AS patients found that kyphosis correction and alleviation of back pain could be achieved by osteotomy through the pathological fracture gap or same-level transpseudarthrosis osteotomy with surgical repair via interbody fusion by a single posterior approach, indicating that these approaches are feasible and potentially effective.

| D ISCUSS I ON
These guidelines have endeavored to include the latest evidence concerning the management of axSpA, and have also incorporated F I G U R E 1 Management algorithm for axSpA issues of local relevance to clinicians in Taiwan. A clinical algorithm covering key aspects of axSpA management is provided in Figure 1 for easy reference. It should be noted that the Overarching Principles and Recommendations should be read with the accompanying statements to derive the best picture of current evidence, and the listed references may need to be consulted when more information is needed. Although significant challenges still remain with the recognition of axSpA as a category of disease that can be used in drug indications and reimbursement criteria, initiatives aimed at increasing clinician acceptance, promoting related research, and advising healthcare policymakers will be undertaken in the near future. These guidelines will continue to be updated in the event of new evidence or the approval of new treatments for axSpA, and feedback will be actively solicited from rheumatologists, clinicians of other disciplines, other healthcare professionals (including pharmacists, physical therapists, and nurses), professional societies, patient groups, healthcare institutions, regulatory authorities, health insurance providers, and the pharmaceutical industry, in order to ensure that these recommendations remain up-to-date and relevant for clinical practice.

ACK N OWLED G EM ENTS
Expert consensus for these recommendations were finalized during 3 committee meetings which were funded by Novartis Taiwan. No funding or sponsorship was received for publication of this article.