Abstract
Early identification of chronic wounds is essential for clinical decision-making in wound care. Biofilm infection is a well-known risk factor for delayed healing, while cytokines in the wound microenvironment play critical regulatory roles throughout the healing cascade. This single-centre prospective cohort study enrolled 66 patients with chronic wounds between 2020 and 2023 to evaluate cytokine biomarkers and biofilm detection tools for predicting wound outcomes. Clinical signs of biofilm (CSB) alone demonstrated limited predictive value, with accuracies of 66.7% for 30-day and 45.5% for 90-day healing. In contrast, the Wound Biofilm Detection Kit (WBDK) showed superior performance, achieving predictive accuracies of 92.4% and 60.6% for 30- and 90-day outcomes, respectively, outperforming both CSB and MolecuLight i:X. Cytokine analysis identified serum CRP, wound CRP and wound MCP-1 as significant predictors, with ROC analysis demonstrating good discriminative ability for wound CRP (AUC = 0.863) and wound MCP-1 (AUC = 0.830). A simplified Lasso regression model incorporating diabetes mellitus, peripheral arterial disease, wound location and WBDK grade achieved an AUC of 0.77 and an accuracy of 76% for 90-day outcomes. External validation was performed in 17 additional patients yielding a predictive accuracy of 76.5%, supporting the robustness of the model. These findings highlight the limited reliability of clinical signs alone and emphasise the value of objective biofilm detection and cytokine profiling in wound prognosis. Our high-accuracy prediction model, based on readily accessible clinical variables and WBDK results, may facilitate precision wound care and improve real-time management.