Purpose To evaluate the efficacy of arthrotomy, when compared with arthroscopy, in the treatment of adults with septic arthritis of any joint. Methods MEDLINE, EMBASE, and Scopus were searched to identify studies comparing arthrotomy and arthroscopy as therapeutic approaches in patients with septic arthritis of any joint. The main outcome was the re-infection rate. A metaanalysis was performed using the generic inverse variance method with random or fixed effects model depending on heterogeneity between studies. Heterogeneity was tested with the I 2 statistic index. Results Twenty studies with 10,249 patients treated by arthrotomy or arthroscopy were evaluated. We observed a significant lower risk of re-infection (odds ratio [OR], 1.35 [95% CI, 1.16-1.58]; p = 0.0002) and complications (OR, 1.32 [95% CI, 1.12-1.55]; p = 0.001) rate as well as less hospital stay (mean difference [MD], 0.57 days [95% CI, 0.10-1.05]; p = 0.02) favouring arthroscopic intervention. The subanalysis indicated that patients with knee (OR, 1.50 [95% CI, 1.17-1.92]; p = 0.001) and shoulder (OR, 1.24 [95% CI, 1.00-1.53]; p = 0.04) septic arthritis intervened by arthrotomy had a higher risk of re-infection. A lower number of hospitalization days (MD, 0.89 days [95% CI, 0.31-1.47]; p = 0.003) and a lower risk for complications (OR, 1.26 [95% CI, 1.04-1.52]; p = 0.02) were observed in patients treated with arthroscopy after septic knee arthritis. Conclusions Available evidence suggests that patients with septic arthritis of the knee and shoulder treated by arthroscopy have less risk of re-infection than those treated by arthrotomy. The quality of the body of evidence is still insufficient to reach reliable conclusions. PROSPERO trial registration number CRD42020176044. Date registration: April 28, 2020.