DISCUSSION
IA corticosteroids have been the most used treatment for articular lesions for years; however, this is a controversial therapy. This systematic review gathered data regarding its effects on joints (McIlwraith & Lattermann, 2019; Velloso Alvarez et al., 2020; Boorman et al., 2022). The evidence suggests that 1) single IA corticosteroid injections in OA joints tend to show higher effectiveness in SMOA in studies with shorter follow-ups or at periods shorter than the follow-up proposed. DMOA showed no consistent adverse effects, varying according to corticosteroids. Nevertheless, some conflicting results were within the same corticosteroid type across studies. 2) Corticosteroids appear not to induce symptomatic changes in normal joints in horses, although adverse changes to joint tissues (cartilage or SM) and metabolism are usually found. A dose effect, where lower doses have better—or less detrimental—responses, was found in in vitro studies of normal equine articular tissues. The risk of bias in included studies was mainly unclear because the authors did not report the entire methodology. For example, it was not reported whether groups were randomized, and if so, what was the randomization method. Therefore, whether the study presents bias arising from selection, performance, detection, attrition, or reporting is unclear. Despite being implied at times, a clear description of methods should always be included, even if it sounds repetitive. Reporting guidelines for different study designs are currently available. By following these, authors can achieve a lower risk of bias, improving study quality and (consequently) reliable data.
A single injection of corticosteroids for OA joints was evaluated to answer question 1. Articles investigating such topics were more common in humans, where the primary outcomes were usually PROMs. SMOA effects tended to show improvements in shorter follow-ups, having positive effects reported in up to three weeks in horses (lameness) and 3–12 weeks in humans. Beyond this period, one study found detrimental effects to SMOA in horses (de Clifford et al., 2021), while in humans, drawbacks were limited to the loss of positive effects. Despite sharing similarities in the disease process, this finding may point to the possible divergent responses in different species and emphasizes the need for species-specific studies to ensure the application of findings in clinical settings.
Regarding DMOA changes after single injections in OA joints, outcome measures varied according to the corticosteroid; however, in vivoor clinical studies including these outcomes were more scarce than ones including SMOA effects. Findings usually represent changes in joint metabolism, indicating possible disease progression rather than linking the disease progression to corticosteroids. For example, changes from imaging tools were conflicting, ranging from significantly higher chances of joint deterioration to no significant risk of worsening of disease progression compared to controls (Okike et al., 2021; Latourte et al., 2022). A positive effect seen in radiographs was only found for TH in one study, in which patients showed disease progression at a significantly lower rate than controls (Nunes-Tamashiro et al., 2022). Other DMOA changes included SF WBC counts (conflicting results for TA with no changes for MPA), negative results such as increasing synthesis of smaller PG (MPA) inducing changes in chondrocytes phenotype (MPA), transiently increasing GAG concentrations in SF (TA), and increasing in PGE2 concentrations in SF (TA). Positive DMOA were decreases in SF TP (TA), decreases in CCL2 and MMPs activity (TA), and improvements in SM immunohistochemistry (MPA). No other significant changes were found in DMOA of horses or humans injected with corticosteroids for OA treatment.
The effects of corticosteroid exposure on normal joints were also investigated. As expected for any therapy, corticosteroids (MPA and TA) did not cause lameness or joint effusion when these were investigated in three studies (Murray et al., 1998; Robion et al., 2001; Céleste et al., 2005). However, some adverse effects on normal joints were found when MPA and TA were evaluated in vivo . Outcome measures were usually those that can demonstrate disturbances in homeostasis (WBC, TP, PG, GAG, and HA concentrations in SF, PG synthesis in cartilage, biomarkers, and gene expression in SF) while some also showed adverse changes in histology and histochemistry of SM and cartilage and biomechanics of cartilage (Frisbie et al., 1997; Murray et al., 1998; Frisbie et al., 1998). At such levels, more modest positive changes could also be found compared to controls (improvements in histology of cartilage and SM, GAG metabolism and synthesis for TA, improvements in collagen and protein synthesis in cartilage and SM, and increases in lubricating surfactant secretion for MPA). In vitro , adverse effects were common for all corticosteroid types (MPA, BTM, DEX, and TA). These included impairments of cartilage metabolism (synthesis and degradation) with usually dose-dependent effects, where lower doses were associated with less detrimental changes.
Effects of medications on normal joints are usually performed to assess the therapy’s safety; however, findings are frequently extrapolated to clinical use in diseased joints. Although this makes sense, the environment of healthy and osteoarthritic joints is highly divergent; therefore, they may behave differently when exposed to therapy. For example, Todhunter et al. (1998) found that the inhibition of PG synthesis in healthy joints exposed to corticosteroids was not present in lipopolysaccharide-induced joints treated the same way. This finding was also seen in Frisbie et al. (1997), in which GAG content in SF was significantly higher in normal joints exposed to TA but not significantly different from controls when surgically induced joints were treated similarly. Accordingly, some differences in the effects of MPA in healthy and diseased joints were found by Frisbie et al. (1998). This finding suggests that results from normal and diseased joints differ and should not be considered together.
Limitations of this review include heterogeneity of included studies and employed treatment protocols, which hamper comparisons and unclear risk of bias for most of the results presented.
 CONCLUSION
Evidence gathered in this review was mainly rated as unclear risk of bias. Our findings suggest that 1) single IA corticosteroid injections tend to show higher effectiveness in SMOA in studies with shorter follow-ups, while DMOA varied according to corticosteroid with conflicting results regarding the same type of corticosteroid across studies. 2) Corticosteroids do not induce symptomatic changes in normal joints of horses despite adverse effects on joint tissues and metabolism. This finding could indicate pathways that might lead to detrimental effects on joint health; however, direct changes are lacking. A dose-dependent effect, where lower doses have better—or less detrimental—responses, was found in in vitro studies of normal equine articular tissues.