Risk of septic arthritis after corticosteroid joint injections: A retrospective propensity score-matched cohort analysis.

The study relied on electronic health records, which lack clinical details like physical exam findings and may contain documentation errors. The injection codes did not specify which large joint was treated, so risk could not be compared between joints like knees versus shoulders. Data from the TriNetX network cannot capture patients who sought care for septic arthritis outside the included health systems. The authors acknowledge potential unmeasured selection bias, where clinicians may have avoided injecting patients with higher infection risks, which matching could not fully correct.

Dr Wilson serves as Deputy Editor of PM&R. No other disclosures were reported.

Do not interpret the negative risk difference as evidence that injections are protective against infection. The authors attribute this finding to unmeasured selection bias, where clinicians likely avoided injecting patients with higher baseline risk factors.