Table 2.

Individual study characteristics and results

Authors	Design	Aims	Subjects	Intervention	Outcomes	Results
Stricker et al. 1996 [21]	Non-blinded non-randomised controlled trial	The chondroprotective effect of betamethasone was examined to determine if corticosteroids can decrease articular cartilage injury caused by inflammatory exudate in Staphylococcus aureus gonarthritis in rabbits	27 adolescent male New Zealand white rabbits split into 3 groups.	Each rabbit’s left knee was inoculated with Staphylococcus aureus. Group 1 received antibiotics alone (intramuscular Ceftriaxone for 12 days), group 2 received intramuscular corticosteroids (4 days Betamethasone) and antibiotics, group 3 received intra-articular glucocorticoids (single dose of Betamethasone) and antibiotics. All treatment was commenced at 36 h after joint inoculation.	Synovial fluid culture & cell count as well as peripheral leukocyte count at days 2, 6 and 14. Histological and histochemical analysis was undertaken at 14 days.	A significant difference was seen in mean peripheral leukocyte count when comparing group 1 and 2 at 14 days (5,100 vs 8,100 white blood cells, p < 0.05). Group 2 showed statistically greater safranin O stain orthochromasia than Group 1 or 3 (p < 0.05) which was thought to suggest prostaglandin preservation. This was consistent with biochemical results for hexosamine preservation when group 2 was compared to group 1 (p < 0.05)
Sakiniene et al. 1996 [23]	Non-blinded non-randomised controlled trial	To evaluate the combined effect of systemic corticosteroids and antibiotics on the course of septic arthritis	Two separate experimental protocols were used: 1–40 five week old Swiss mice (20 male, 20 female) split into 3 groups. 2 – 43 male Swiss nice (5–6 weeks old) split into 3 groups. Results of these experiments were pooled.	In both studies mice were initially inoculated with Staphylococcus aureus. One group received no treatment, one group received antibiotics alone (Cloxacillin given intraperitoneally every 12 h) and one group received corticosteroids (Dexamethasone given intraperitoneally every 24 h) and antibiotics. Intervention was commenced at 3 days post inoculation.	Frequency and severity of arthritis (Arthritis index) at 0,3,7,14 days. Histopathological, immunohistochemical, bacteriologic, serologic and haematologic manifestations were assessed at 14 days.	Only those treated with antibiotics and corticosteroids showed a significantly difference in the frequency and severity of arthritis compared to controls at 14 days. Serum levels of IFN-γ were 4 fold decreased in the antibiotic treated group and 15 fold decreased in the corticosteroid and antibiotic group. Serum NO3− levels were significantly reduced in both the treatment arms.
Wysenbeek et al. 1998 [24]	Non-blinded non-randomised controlled trial	To assess the effect of intra-articular corticosteroids added to systemic antibiotics in experimental septic arthritis.	30 eight week old New Zealand rabbits split into 3 groups (n = 10 per group).	Rabbits had their left knee inoculated with Staphylococcus epidermis. Group 1 were untreated controls, group 2 were treated with antibiotics alone (daily intramuscular Cefonicid) and group 3 received glucocorticoids (intra-articular Methylprednisolone given 48 h after infection and 24 h after antibiotic commencement) and antibiotics	A composite score of Histopathological-histochemical parameters in both a vertical (A) and horizontal (B) section at follow up (15 days).	Both treatment groups had significantly lower scores than the control. Group 3 had significantly lower scores than group 2 for A (6.5 [1.4] vs 4.0 [1.0], p = 0.001) and B (7.4 [2.6] vs 4.2 [2.2], p = 0.01) [] = Standard Deviation (SD).
Odio et al. 2003 [17]	Double-blind randomised controlled trial	Assess the impact of corticosteroids and antibiotics on joint dysfunction and the acute response in septic arthritis.	123 children with documented septic arthritis aged 3 months to 11 years (with parental consent). 23 children were unevaluable therefore 100 patients were assessed. There were two groups: Dexamethasone (n = 50) and Placebo (n = 50). There were no significant differences in baselines characteristics between groups	Placebo group received antibiotics and IV 0.9 % saline (equivalent of Dexamethasone timing and volume). There was a variety of antibiotic regimes used based on age, identified pathogen and organism resistance. Dexamethasone group received antibiotics and IV Dexamethasone (0.2 mg/kg/dose given every 12 h for 8 doses). Dexamethasone/Placebo given from diagnosis.	Clinical outcomes were assessed at the end of treatment, 6 months and 12 months. During the acute phase time for normalisation of CRP, duration of symptoms, extent of antibiotic therapy and arthrotomy requirements were also assessed.	Levels of residual dysfunction were significantly improved for Dexamethasone vs. Placebo at end of treatment (2 vs 16 [5.54–16.62], p < 0.001), at 6 months (1 vs 19 [1.41–4.22], p < 0.001) and 12 months (1 vs 13 [1.31–4.1], p < 0.001) [] = 95 % CI. Compared to placebo the Dexamethasone group also had quicker normalisation of CRP (2.04 ± 1.25 vs 4.68 ± 6.23 days, p = 0.01); earlier resolution of symptoms (2.34 ± 5.06 vs 7.81 ± 2.04 days, p = 0.001); and shorter duration of IV antibiotics (7.2 ± 1.2 vs 10 ± 5.6, p < 0.05). ± = SD
Harel et al. 2011 [16]	Double-blind randomised controlled trial	Evaluate the effect of adding dexamethasone to antibiotic therapy in the clinical course of septic arthritis in children.	49 children aged 6–161 months with confirmed septic arthritis. There were two groups: Dexamethasone (n = 24) and Placebo (n = 25)	Placebo group received antibiotics (IV Cefuroxime 150 mg/kg/day in 3 divided doses) and placebo (IV 0.9 % saline at equivalent of Dexamethasone timing and volume). Dexamethasone group received antibiotics and IV Dexamethasone (0.15 mg/kg/dose) every 6 h for 16 consecutive doses. Dexamethasone/Placebo given from diagnosis.	Comparison between groups for clinical and laboratory parameters, length of hospital stay, and late sequelae (follow up at 2,6,12 months).	Compared to placebo those treated with Dexamethasone had a shorter duration of fever (1.68 [2.57] vs 2.38 [2.41] for first day without fever, p = 0.021); quicker normalisation of CRP (3.09 [3.80] vs 5.48 [4.69] days, p = 0.029); and duration of IV antibiotics (9.91 [4.84] vs 12.60 [5.20] days, p = 0.007). No formal assessment of late sequelae or length of hospital stay was performed. [] = SD
Arti et al. 2014 [18]	Double-blind non-randomized controlled trial	Evaluation of the effects of IV dexamethasone on the recovery process in septic arthritis	60 patients with septic arthritis separated into two groups (n = 30). Mean age in group 1 was 8.06 ± 0.5 years. 73 % of the group 1 participants were male and 27 % were female. Mean age in group 2 was 8 ± 0.6 years. 70 % of the group 1 participants were male and 30 % were female.	Group 1 received IV antibiotics (dose & timing not detailed) and IV Dexamethasone (0.15 mg/kg/dose every 6 h for 4 days). Group 2 received antibiotics and placebo (IV 0.9 % saline given at same volume and timing as Dexamethasone). Dexamethasone/Placebo given from diagnosis.	Groups were compared regarding clinical (inflammation & redness, range of motion (ROM), and duration of therapy) and laboratory findings (ESR & CRP).	There was a significant difference between groups 1 & 2 regarding redness & inflammation (2.43 ± 0.15 vs 6.53 ± 0.27 days, p = 0.0014); days of hospitalisation (8.9 ± 0.88 vs 12.17 ± 0.54 days, p = 0.005); joint ROM (91.2 ± 2.28 vs 40.53 ± 1.44°, p = 0.008); and reduction of ESR levels (36.03 ± 2.77 vs 16.73 ± 0.83, p = 0.0001) ± = SD