Serum β-defensin-3 and inflammatory markers for discriminating diabetic foot infection and infection severity: a retrospective study
Related Experiment Videos
Yali Ma1,2, Zhao Qu1,3, Qianqian Yao1,3
1School of Basic Medical Sciences, China Three Gorges University, Yichang, Hubei, China.
Background:
Human β-defensin-3 (hBD-3) is an epithelial-derived antimicrobial peptide involved in innate immunity and tissue repair, but its clinical association with diabetic foot infection (DFI) remains insufficiently characterized.
Objective:
To evaluate the associations of serum hBD-3 with prevalent DFI and infection severity and to assess its discriminatory performance alone and in combination with conventional inflammatory markers.
Methods:
This retrospective study included 150 participants: 30 healthy controls, 30 patients with diabetes without foot disease, and 90 patients with diabetic foot disease, including 30 without infection and 60 with DFI. Logistic regression and LASSO penalized regression were used to identify factors associated with prevalent DFI and moderate/severe infection. Restricted cubic spline analyses explored nonlinear associations. Discriminatory performance was evaluated using receiver operating characteristic curves, and internal validation was performed with 1,000 bootstrap resamples.
Results:
Serum hBD-3 was higher in patients with DFI than in those with non-infected diabetic foot disease [108.9 (86.3, 141.5) vs. 69.8 (56.1, 84.7) pg/mL, P < 0.001] and increased with infection severity. Higher hBD-3, CRP, PCT, and HbA1c and lower albumin were independently associated with prevalent DFI. The final prevalent-DFI model included HbA1c, albumin, hBD-3, CRP, and PCT and achieved an apparent AUC of 0.928, with an optimism-corrected AUC of 0.895. The moderate/severe DFI model included hBD-3, CRP, and PCT and achieved an apparent AUC of 0.935 and an optimism-corrected AUC of 0.917. RCS analyses suggested nonlinear associations between hBD-3 and both prevalent DFI and infection severity.
Conclusion:
Serum hBD-3 was associated with the presence and severity of DFI and provided complementary discriminatory information beyond conventional inflammatory markers. The combined models showed promising performance but remain exploratory and require prospective multicenter external validation before clinical application.
