Adjunctive corticosteroids may modestly reduce hospital stay for bacterial facial infections but evidence certainty is low

BACKGROUND: Objectives: To evaluate the clinical effectiveness and safety of adjunctive corticosteroid versus standard treatment alone in facial infections. METHODS: Eligibility criteria: Included clinical comparative studies (randomized or non-randomized) evaluating adjunctive corticosteroids versus standard care in patients with bacterial facial infections such as odontogenic/facial space infections, dental abscesses, orbital cellulitis, periorbital cellulitis, and Ludwig's angina. INFORMATION SOURCES: An unrestricted literature search of five databases was conducted up to December 15, 2025. Risk of bias: The quality assessment of the studies was conducted using the Cochrane Risk of Bias Tool (ROBINS-I) for non-randomized and a new risk of bias tool (RoB-2) for randomized studies. SYNTHESIS OF RESULTS: Random effects meta-analyses using mean difference (MD) or standardized mean differences (SMDs) were performed, followed by sensitivity analyses, and assessment of the quality of evidence using GRADE. RESULTS: Included studies: Fifteen comparative studies (three randomized and 12 non-randomized) including 13,905 patients with bacterial facial infections (61.5% male) were included. SYNTHESIS OF RESULTS: Low-certainty evidence showed that the use of adjunctive steroids significantly reduced hospital stay compared with standard treatment alone (10 studies; MD = -1.61 [-3.17, -0.05] days; p = 0.04; I = 99.3%). Low-certainty evidence also showed that pre-treatment CRP was higher in the steroid groups (3 studies; SMD = 0.33 [0.09, 0.58]; p = 0.03; I = 0%). Low-certainty evidence indicated no significant difference in the number of surgeries (p = 0.13), while very low-certainty evidence showed no significant differences in intensive care unit admissions (p = 0.07) or airway compromise (p = 0.18). Sensitivity and subgroup analyses showed no significant differences by study design, infection type, or risk of bias, confirming robust results; only ethnicity had a significant effect on hospital stay (p = 0.0027). DISCUSSION: Limitations of evidence: Certainty of evidence was low to very low for all outcomes, with hospital stay rated low due to methodological limitations, high risk of bias, and the predominance of non-randomized studies. INTERPRETATION: The low level of evidence suggests that adjunctive corticosteroids may provide a reduction in hospital length of stay for patients with bacterial facial infections. Further well-designed, randomized, and prospective comparative studies are needed to confirm these findings. PROPSERO Registration Number: (CRD420261329331).