METHODS
This systematic review followed the PRISMA guidelines (Moher et al., 2009). Using three databases (PubMed, CAB Direct, and Web of Science), a systematic search was made of all published articles up to 20 July 2022, employing the terms described in Table 1.
Inclusion criteria were articles with a full text available, published in English, evaluating the effects of IA corticosteroid injections alone on limb/arm/leg joints. In vitro studies were only included when investigating equine tissues. Corticosteroids included methylprednisolone acetate (MPA); betamethasone (BTM), triamcinolone acetonide (TA), triamcinolone hexacetonide (TH); dexamethasone (DEX), and isoflupredone acetate (IPA). Acceptable study designs were pre-clinical studies, randomized clinical trials (RCTs), prospective or retrospective cohorts, case-controls, crossovers, or pilots; Case reports, reviews, expert opinions, meeting abstracts, conference papers, and designs other than those cited previously were excluded. When equine species were the investigated population, at least a comparison from the baseline of the corticosteroid treatment effect was required. For other species, a negative control as the comparison group was required for inclusion. Other exclusion criteria included other joint types (e.g., temporomandibular), studies only containing groups with injections of corticosteroids plus other medications (e.g., local anesthetics, hyaluronic acid, or antibiotics), and patients with rheumatoid arthritis, juvenile idiopathic arthritis, hemarthrosis, or septic arthritis.
Additional criteria were added according to each question in this review. For question one, articles that investigated the effects of a single IA corticosteroid injection as treatment for OA in horses were included. Because veterinary science is often based on human practice, studies with humans as the investigated population were also included. No in vitro studies were considered for this question. For question two, we considered only studies investigating the effects of IA corticosteroid treatment in healthy joints in horses (in vivo ) or exposure of corticosteroids to normal/healthy equine joint tissues (in vitro ).
Risk of bias was assessed for RCTs using the Cochrane Risk of Bias tool, in which possible sources of bias are evaluated separately, and their risks are rated as ”low,” ”unclear,” or ”high” (Higgins et al., 2011). According to the tool, an RCT is considered to have a final “low risk of bias” criteria only if all its items are classified as ”low risk.” If one or more items are classified as ”unclear,” the final criterion of the study is ”unclear risk,” and if at least one of them is classified as ”high risk,” the study is considered to have a ”high risk” of bias. Studies for horses in vivo were assessed similarly. Observational studies (when included) were rated using the methodological index for non-randomized studies (MINORS) tool, which rates items as reported and appropriate (score = 2), reported and inappropriate (score = 1), or not reported (score = 0) (Slim et al., 2003). Scores were then summed and classified as low risk of bias (final score 13–16 for non-comparative studies or 19–24 for comparative studies) or high risk of bias (scores of ≤ 12 for non-comparative studies or ≤ 18 for comparative studies) (Ajrawat et al., 2019; Slim et al., 2003).
Primary outcomes of interest were symptom-modifying changes (lameness/patient-reported outcome measures [PROMs]) and disease-modifying changes (imaging, histology, biochemistry, and others).