What are the core recommendations for rheumatoid arthritis care? Systematic review of clinical practice guidelines

Abstract Systematic r eview to evaluate the quality of the clinical practice guidelines (CPG) for rheumatoid arthritis (RA) management and to provide a synthesis of high-quality CPG recommendations, highlighting areas of consistency, and inconsistency. Electronic searches of five databases and four online guideline repositories were performed. RA management CPGs were eligible for inclusion if they were written in English and published between January 2015 and February 2022; focused on adults ≥ 18 years of age; met the criteria of a CPG as defined by the Institute of Medicine; and were rated as high quality on the Appraisal of Guidelines for Research and Evaluation II instrument. RA CPGs were excluded if they required additional payment to access; only addressed recommendations for the system/organization of care and did not include interventional management recommendations; and/or included other arthritic conditions. Of 27 CPGs identified, 13 CPGs met eligibility criteria and were included. Non-pharmacological care should include patient education, patient-centered care, shared decision-making, exercise, orthoses, and a multi-disciplinary approach to care. Pharmacological care should include conventional synthetic disease modifying anti-rheumatic drugs (DMARDs), with methotrexate as the first-line choice. If monotherapy conventional synthetic DMARDs fail to achieve a treatment target, this should be followed by combination therapy conventional synthetic DMARDs (leflunomide, sulfasalazine, hydroxychloroquine), biologic DMARDS and targeted synthetic DMARDS. Management should also include monitoring, pre-treatment investigations and vaccinations, and screening for tuberculosis and hepatitis. Surgical care should be recommended if non-surgical care fails. This synthesis offers clear guidance of evidence-based RA care to healthcare providers. Trial registration The protocol for this review was registered with Open Science Framework (https://doi.org/10.17605/OSF.IO/UB3Y7). Supplementary Information The online version contains supplementary material available at 10.1007/s10067-023-06654-0.


Introduction
Rheumatoid arthritis (RA) is a systemic autoimmune disorder affecting 0.1-2.0% of most populations [1]. It is a long-term condition characterized by joint inflammation, with potential for joint damage and extra-articular manifestations [2]. RA can significantly impact physical, mental, and social health and can increase morbidity and mortality [3,4]. Economic costs, including direct (e.g., drug costs) and indirect costs (e.g., absenteeism and work disability) are estimated to range from US$2,408 to US$83,845 annually [5].
The last 30 years have seen many substantive changes in RA management including expansion in options for pharmacological management, introduction of instruments for clinical monitoring of disease activity and impact, and increased focus on patient-centered care and support for self-management [6,7]. The pharmacological management options have expanded from conventional synthetic diseasemodifying anti-rheumatic drugs (csDMARDs) to biologic DMARDS (bDMARDs) and more recently targeted synthetic DMARDS (tsDMARDS). There are also multiple treatment strategy trials to be considered alongside local medication availability [8]. The evidence underpinning these management approaches are often summarized for clinicians in clinical practice guidelines (CPG).
The aim of CPGs is to support evidence-based clinical decision-making, improve consistency of care and optimize patient outcomes [9]. Robust CPGs comprise a set of management recommendations, created from a systematic review of the literature and consensus by an expert panel [10]. While CPG production has increased in recent decades there are some concerns about quality and implementation into practice [11][12][13]. Low quality CPGs do not improve care, and conflicting recommendations between CPGs can lead to clinician confusion [11,14]. To date, systematic reviews have either appraised CPG quality [15], or provided a narrative summary on RA management options [16], or reported on both CPG quality and content, but were only specific for certain management options (e.g., physiotherapy interventions [12], Chinese medicine [17]). To the best of the authors knowledge, currently no systematic review has appraised CPG quality and synthesized recommendations from high-quality CPGs for all management options. By summarizing high quality CPG recommendations, we can offer healthcare providers clear, simple guidance on evidence-based RA care.
The aims of this systematic review were to (1) evaluate the quality of the CPGs for the management of RA and (2) to provide a synthesis of high-quality CPG recommendations, highlighting areas of consistency and inconsistency.

Materials and methods
This systematic review is reported according to the Preferred Reporting Items for Systematic reviews guidelines and the protocol registered on the Open Science Framework (https:// doi. org/ 10. 17605/ OSF. IO/ UB3Y7) [18,19]. For full details of methods, see Conley et al. [20]. Briefly, five databases (OvidSP MEDLINE, Cochrane, CINAHL, Embase, and Physiotherapy Evidence Database (PEDro) and four guideline repositories were searched from January 2015 to April 1st, 2023. Online Resource 1 provides the details of the search strategy. Search results were exported into Endnote™ and duplicates removed electronically and manually checked before importing titles/abstracts into Covidence systematic review software (Veritas Health Innovation, Melbourne, Australia. Available at www. covid ence. org). Two independent authors (BC and TG or IL) screened the titles and abstracts to identify relevant studies. Then full texts were then screened for eligibility (Table 1). Any discrepancies were resolved by a third reviewer. Deviations from the original protocol included updating the search twice and inclusion of CPGs that addressed one treatment modality (e.g., • Published in the English language or in which a complete English language version is available • Developed using a systematic process that is a guideline based on a systematic review of the literature and developed by an expert, multidisciplinary panel [2] • Represents an original body of work, i.e., not solely an adaptation or systematic review of existing guidelines Exclusion criteria • Does not include interventional management recommendations • Includes other arthritic conditions • Only addresses recommendations for the system/organisation of care • Unavailable via institutional access, i.e., requires additional payment 1 3 medication prescribing) which were originally excluded based on their narrow scope.
The Appraisal of Guidelines for Research and Evaluation (AGREE) II instrument was used to assess CPG quality [21]. This is a valid, reliable tool that is widely used in CPG appraisal, including those for RA management [11,12,15,22,23]. Pairs of reviewers (from SB, PO, JB, JP, TG, IL, BC) independently rated each CPG against the AGREE II items using a 7-point Likert scale, from 1 (AGREE II criteria not addressed) to 7 (all AGREE II criteria addressed). Individual reviewer domain scores were calculated and expressed as a percentage. We defined acceptable inter-rater agreement as domain scores of 80% or above, consistent with excellent intraclass coefficient values [24,25]. If the two reviewer's domain scores varied equal to or greater than 20%, reviewers met to discuss discrepancies and a third reviewer was consulted to resolve any disagreements on the final rating. The AGREE II developers do not provide criteria for CPG quality, rather, they suggest this is at the discretion of the researchers [21]. Consistent with a previous musculoskeletal review, the authors of this study considered the following domains most important when screening high quality RA CPGs: stakeholder involvement (domain 2); rigor of development (domain 3); and editorial independence (domain 6) [13]. Arthritis reviews implementing the AGREE II instrument established an threshold for high quality as equal to or greater than 60% [12,26]. We decided that CPGs that did not meet this definition were excluded (Online Resource 3).
The first author (BC) independently extracted and recorded CPG data on a bespoke Microsoft Excel spreadsheet, based on a previous musculoskeletal review [13] (Online Resource 2). CPG recommendations were extracted and ranked as either a "should do," "could do," "do not do," or "uncertain" (Online Resource 4). Recommendations were classified into these four categories based on language used in the CPGs (Table 2), consistent with a previous musculoskeletal systematic review of CPGs [13]. Two authors (SB and IL) cross-checked the extracted data and recommendation rankings, any inconsistencies were resolved through discussion among the three authors (BC, SB, IL) while referring to the original CPG.
Recommendations were categorized based on the type of intervention (non-pharmacological, pharmacological, and surgical) and then further divided into individual interventions within these categories (e.g., patient education). Narrative summaries were developed for individual interventions and identified which CPGs included a recommendation within that category; and areas of consistency and inconsistency between CPG recommendations (Online Resource 5). The research team developed a consensus process ( Fig. 1), to describe the consistency of recommendations between CPGs, providing a global consensus recommendation on the individual interventions.

Consensus recommendations
Following synthesis, twenty-two common/consistent "should do" recommendations, seven common/consistent "could do" recommendations, two common/consistent "do not do" recommendations and four "no consensus" recommendations were identified ( Fig. 3; Online Resource 5).

Recommendations with "Should do" consensus
The following care elements were strongly recommended by two or more CPGs:

Pre-treatment investigations and vaccinations
Patients should undergo investigations before commencing treatment [29,30]. These can include screening for comorbidities, pregnancy, chest radiography, blood tests, renal, and liver function tests [29,30]. All patients should be assessed for infections and vaccinations (particularly live vaccinations) should be provided ideally 4 weeks prior to bDMARD or tsDMARD therapy and updated as needed [4,28,30]. Killed or recombinant vaccines can be administered before initiating or during csDMARD, bDMARD, or tsDMARD therapy [4].
Special populations All patients should be screened for tuberculosis (TB) infection before commencing bDMARD or tsD-MARD therapy [4,30]. If a patient has active or latent TB this should be adequately treated before commencing bDMARDs or tsDMARDs [4,30]. Patients should be screened for hepatitis B virus or hepatitis C virus infections and if positive, receive further evaluation and treatment [4,30,35].

Surgical
Referral for surgical opinion Surgery should be considered when medical management has not been successful, and the patient meets evidence-based criteria for surgery (Online Resource 5) [31,37].

Recommendations with "Do not do" consensus
The following were specifically recommended not to be done by two or more CPGs: TNF inhibitor should not be recommended in persons with a history of congestive heart failure, unless there is no other reasonable option, and the congestive heart failure is compensated [4,35]. Live vaccines should not be administered while a patient is being treated with bDMARDs or tsDMARDs [4,30].

Recommendations with no consensus
CPGs reporting conflicting recommendations for: Non-pharmacological treatments such as manual therapies (e.g., thermotherapy, massage, dry needling, passive mobilisations [32,33]) and recommendations for patients with RA and serious infections [27,30,35] or cancer [4,27]. For example, one CPG recommended bDMARDs, with no specific treatment over another [27] while the other CPG recommended csDMARDs for previous low-grade melanoma or non-melanoma skin cancer or lymphoproliferative disorder or standard care for previously treated solid organ malignancy [4]. For full details of conflicting recommendations, see Online Resource 5.

Discussion
Clinicians should be aware of and utilize high-quality CPGs to inform decision-making. We determined sixteen of the twenty-seven CPGs were high quality on the AGREE II instrument. CPGs are costly and time-consuming to develop and the significant number of low-quality CPGs is an inefficient use of resources [52]. Local adaption of contemporary high-quality CPGs that includes relevant updates in evidence and a section for area-specific recommendations/considerations such as, availability or cost of treatments may be a more efficient use of resources than development of several original CPGs within a similar time period.
Development of high-quality CPGs are important, although this does not guarantee translation of recommendations into practice [53]. Studies have reported sub-optimal adherence to CPGs, varying from 22 to 100% for rheumatologists with findings being similar across most health disciplines and internationally [54,55]. The AGREE II instrument measures CPG implementation within the "applicability" domain, which was the lowest scored domain across CPGs (Table 3); consistent with previous systematic reviews for RA management [15,[56][57][58]. This highlights the need for CPG developers to focus on addressing implementation in future guidelines. Inclusion of economic evaluations, treatment algorithms and monitoring and auditing criteria can facilitate translation of recommendations into practice [21]. A variety of individual, health system and contextual barriers to CPG implementation have been identified by Correa et al. [59], which include insufficient high quality evidence, contradictory CPG, and patient and physician factors [59]. While clinicians may choose to use treatment recommendations developed by local peak bodies, systematic reviews such as ours are important to address concerns about the quality of recommendations, and conflicting recommendations. By applying these recommendations, clinicians in any setting can be confident that they are offering high quality care that is supported by robust evidence.
Recommendations across CPGs were relatively consistent for non-pharmacological, pharmacological, and surgical care. Non-pharmacological interventions should include patient education, patient-centered care, shared decision-making, exercise, orthoses, and a multi-disciplinary approach to care. Pharmacological interventions should include csDMARDs, with MTX as the first-line choice. Followed by csDMARDs as combination therapy (LEF, SSZ and HCQ), bDMARDS, and tsDMARDS to achieve a treatment target. Other aspects of medical management consistently recommended included monitoring, pretreatment investigations and vaccinations, and screening for tuberculosis and hepatitis. Surgical care should be recommended if non-surgical care fails. TNF inhibitors should not be used in persons with a history of congestive heart failure, unless there is no other reasonable option, and the congestive heart failure is compensated. Live vaccines should be avoided while patients being treated with bDMARDs or tsDMARDs.
Despite the majority of recommendations being consistent across CPGs, they often lacked sufficient detail to guide practice. For example, physiotherapy, psychology, nursing, and rheumatology are professions which provide a range of care options; however, in some CPGs these were broadly classified as interventions. Distinct recommendations for interventions provided by these professions are needed in future CPGs [56]. Similarly, pharmacological recommendations neglected important information needed for implementation into clinical practice such as medication dosage. Dosage for csDMARDs and other medications were often not included with only two CPGs reporting on dosage for MTX, differing from least 10 mg/week to at least 15 mg/ week [27,35]. While dose variation could be attributed to different disease states and consideration of potential side effects, inconsistent dosages, or not including medication dosages are likely barriers to implementation [60]. CPGs often agreed on monitoring frequency [30,34,35,37], and instruments to measure disease activity [4, 8, 27-30, 37, 51] although recommendations on blood monitoring were vague. Medication management is a large component of RA care, and these medications can be associated with adverse effects such as fatigue, nausea, cytopenia, among others [61]. Furthermore, the presence of side effects can prompt changing of treatments. Future CPGs should focus on providing clear, detailed recommendations to improve consistency of care.
No consensus recommendations could be developed due to conflicting recommendations for certain non-pharmacological interventions such as: thermotherapy, recommendations for tapering medications and recommendations for patients with RA and cancer or serious infections. For example, one CPG recommended specific csDMARDs or standard care for patients with RA and certain cancers [4], while the other CPG recommended bDMARDs on a case-by-case basis for patients with RA and cancer [27]. Differences might be attributed to their varied definitions of cancer, either defining it broadly [27] or stating specific cancer types [4]. Guideline development groups interpretation of evidence can influence recommendations [16]. Both CPGs included relevant health professionals in the guideline development group such as rheumatologists and methodological experts, although varied at times for other members, e.g., one CPG included a biostatistician [4] while the other included patients [27]. Another potential reason for the difference could be their methodological process of determining evidence quality, which varied from using either Scottish Intercollegiate Guidelines Network and Osteba critical appraisal tools [27] or the Oxford Levels of Evidence [4].
Our synthesis identified several areas for further development/investigation. Data related to medication tapering is emerging [62]. As such, not all CPGs reported on tapering and those that did differed on timeframes of when to begin tapering. CPGs recommended tapering from 6 [27,30,35,36] or 12 months [30,37], or did not advise on a time frame [4,8,28,34]. It is important that future CPGs include recommendations on tapering as increasingly people are diagnosed early with RA and treated earlier and the potential issues with long term immunosuppression such as increased likelihood of developing infections [63,64]. Further research is warranted to explore when tapering should occur and to inform CPGs through quantifying risk of flare with treatment tapering. tsDMARD evidence is an emerging area of research, with JAK-inhibitors have been supported in a recent CPG [34,65]. Benefits of JAK-inhibitors include their effectiveness and safety, that they can be administered orally and are associated with a lower production cost in comparison to bDMARDs [2]. This highlights the importance of CPGs being updated every 5 years to reflect advances in medicine [10].

Strengths and potential limitations
Strengths of this systematic review include the use of AGREE II tool as a systematic approach to synthesis [21], and selection of a high-level quality cut-off value, that was based on other reviews in the field [12,26]. Additionally, we involved a multi-disciplinary team, including rheumatologists (RG, MN, CB), physiotherapists (IL, SB, JP, BC), and social scientists (TG, PO, JB).
The AGREE II instrument examines CPG methodology, not necessarily content and scores can be influenced by authors' reporting rather than methodological quality [20]. Our search strategy may have failed to identify all relevant CPGs in relation to RA care as non-English language CPGs and CPGs that addressed assessment and/or diagnosis of RA without management or treatment recommendations were excluded. To reduce the likelihood of CPGs being missed, a medical reference librarian assisted in the development of the search strategy and all authors checked the list of full-text CPGs to determine if any were missing to the best of their knowledge. Authors were required to interpret the language used in CPG recommendations, to provide grading of interventions (e.g., "should do" and "could do"). To improve confidence in our interpretations, consensus statements were developed by three authors (BC, SB, and IL) and reviewed by the expert clinicians (MN, RG, CB). The process of interpretation is clearly reported in the methods, and in previous reviews [20].

Conclusion
Sixteen of the twenty-seven CPGs were identified as high quality on the AGREE II instrument.
Thirteen CPGs met the eligibility criteria, and their recommendations were synthesized. Non-pharmacological care should include patient education, patient-centered care, shared decision making, exercise, orthoses, and a multidisciplinary approach to care. Pharmacological care should include csDMARDs, with MTX as the first-line choice. If monotherapy csDMARDs fail to achieve a treatment target, this should be followed by combination therapy csDMARDs (LEF, SSZ, HCQ), bDMARDS, and tsDMARDS. Management should also include monitoring, pre-treatment investigations and vaccinations, and screening for tuberculosis and hepatitis. Surgical care should be recommended if nonsurgical care fails. This synthesis can provide clear, simple guidance of evidence-based RA care to healthcare providers.

Supplementary Information
The online version contains supplementary material available at https:// doi. org/ 10. 1007/ s10067-023-06654-0. Funding Open Access funding enabled and organized by CAUL and its Member Institutions This project is funded by the Australian Commonwealth Government through a grant administered by Arthritis Australia (GA76063). The funders had no role in the study design; in the collection, analysis, and interpretation of data; in the writing of the report; nor in the decision to submit the article for publication.

Declarations
Competing interests Professor Cheryl Barnabe-Past 2 years-speaking fees from Fresenius Kabi, Janssen, Pfizer; advisory board fees from Celltrion Healthcare, Sanofi Genzyme. There are no other competing interests for the remaining authors.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.