Animal Models of Steroid-Induced Osteonecrosis of the Femoral Head Established by Different Doses of Lipopolysaccharide Combined with Glucocorticoids: A Systematic Review and Network Meta-Analysis.

Steroid-induced osteonecrosis of the femoral head (SONFH) is commonly modeled using lipopolysaccharide (LPS) combined with glucocorticoids, but the comparative performance of different LPS doses remains unclear. We conducted a network meta-analysis following PRISMA guidelines and registration in PROSPERO (CRD420261346269) to compare different LPS dose regimens combined with glucocorticoids for establishing SONFH animal models. Embase, PubMed, and Web of Science were searched through December 2025 for randomized controlled animal studies. The primary outcome was the empty osteocyte lacunae rate, and the secondary outcome was the modeling success rate. Forty articles comprising 49 studies and 923 animals were included. Compared with normal controls, low-, medium-, and high-dose LPS combined with glucocorticoids all significantly increased the empty osteocyte lacunae rate and modeling success rate. High-dose LPS produced a significantly greater empty lacunae rate than low-dose LPS, whereas most other pairwise comparisons were not statistically significant. Subgroup analyses indicated that modeling efficacy was influenced by animal species, glucocorticoid dose, and modeling duration, while high-dose LPS may be associated with increased mortality risk. Overall, LPS plus glucocorticoids effectively induces SONFH-like pathological changes, but protocol selection should balance modeling efficacy and safety, and no single definitively optimal regimen can yet be concluded.