Background Exercise therapy has proven effective for people with rheumatic and musculoskeletal diseases (RMDs), including those with inflammatory arthritis such as rheumatoid arthritis (RA) or axial spondyloarthritis (axSpA) [1-2]. Exercise therapy is generally considered safe for people with RMDs, although the evidence is scarce. A few reviews reported on the nature and risk of harms of exercise therapy in RMDs, but none of them specifically addressed the quality of reporting of harms of exercise therapy in studies including people with inflammatory arthritis.

Objectives This study aimed to describe the quality of reporting of harms in clinical studies on the effectiveness of exercise therapy in people with RA or axSpA.

Methods RCTs with at least one treatment arm consisting of supervised exercise therapy in people with RA or axSpA were included. Eight electronic databases were searched up to November 2021. Two researchers independently selected studies for inclusion and extracted data and in case of disagreement a third researcher was consulted. Data extraction included study characteristics and fulfillment of a set of quality aspects derived from the Consolidated Standards of Reporting Trials (CONSORT) Extension for Reporting Harms Outcomes [3], predefined on the basis of consensus among authors (Table 1). Harms outcomes were defined as adverse events reported on individual level irrespective of causality or negative effects on group level (only if explicitly designated as measurement of potential harm). We considered the reporting on harms outcomes of sufficient quality if the authors reported at least 1) the methodology for active surveillance of harms outcomes (item 2a); and 2) the observed number and the nature of harms (items 3b and 3c).

Results The search yielded 5921 records, of which 64 studies (n= 41 RA, n=23 axSpA; described in 83 papers) were included. Of those studies in RA and axSpA, 34 (83%) and 15 (65%) included any information on harms, with 12 (29%) and 3 (13%) reporting active surveillance and 22 (54%) and 5 (22%) reporting on harms outcomes in the results section, respectively (see Table 1). In total, 10 of the 41 (24%) RA studies and 2 of the 23 (9%) axSpA studies fulfilled the predefined criteria for sufficient quality of reporting.

Conclusion The quality of reporting on harms outcomes is insufficient in the majority of RCTs on exercise therapy in people with RA or axSpA, with overall poorer quality in studies on axSpA which impedes substantiated conclusions about harms of exercise therapy. Our findings stress the need for consensus on the definition, classification, assessment and reporting of harms outcomes in trials on the effects of exercise therapy.

References [1]Rausch Osthoff et al. RMD Open 2018;4(2):e000713.

[2]Regel et al. RMD Open 2017;3(1):e000397.

[3]Ioannidis et al. Ann Intern Med 2004;141(10):781-8.

Acknowledgements This project is financially supported by Netherlands Organization for Health Research and Development (ZonMw; 852004019/ 852004018), Ministry of Health, Welfare and Sport (VWS), the Royal Dutch Society for Physical Therapy (KNGF) and the Dutch Arthritis Society (ReumaNederland).

Disclosure of Interests None Declared.