Staged management of infected diabetic foot ulcers: a 300-patient cohort study on prognostic grading, pathogen dynamics, and individualized risk prediction

Objective This study aimed to evaluate the IWGDF/IDSA classification for prognostic stratification in diabetic foot infections (DFIs) and to develop and validate a clinical prediction model for individualized risk assessment. Methods In this retrospective study of 300 patients with DFIs (2020–2024), infection severity was graded as Grade 2 (mild), 3 (moderate), 3(O) (moderate with osteomyelitis), or 4 (severe). We analyzed baseline characteristics, microbiological profiles, treatment responses, and performed time-to-event analyses for ulcer healing and limb salvage. A nomogram predicting 6-month non-healing risk was developed using Cox regression, with performance assessed by C-index and calibration. Subgroup and sensitivity analyses (including competing-risk models) were conducted to test robustness. Results Higher infection grades correlated with worse glycemic control (HbA1c: 9.8% in Grade 4 vs. 8.1% in Grade 2), higher peripheral arterial disease (PAD) prevalence (64.3% vs. 22.4%), and a pathogen shift from Gram-positive cocci (87.2% in Grade 2) to Gram-negative bacilli ( Pseudomonas aeruginosa 52.2%) and polymicrobial infections (60.9%) in Grade 4 (all p 0.001). Kaplan–Meier analysis revealed a severity-dependent gradient in outcomes: median healing time prolonged from 6.0 weeks (Grade 2) to 18.0 weeks (Grade 4), while the 24-week cumulative amputation rate escalated to 66.7% in Grade 4. Osteomyelitis (Grade 3(O)) constituted a distinct high-risk subgroup, with outcomes approximating Grade 4. A nomogram incorporating infection grade, PAD, HbA1c ≥ 9%, and revascularization status demonstrated good discrimination (C-index 0.82, bootstrap-corrected 0.80) and calibration (Hosmer–Lemeshow p = 0.342), maintained in temporal validation (C-index 0.79). Subgroup analysis confirmed a significant interaction between infection grade and revascularization, with the greatest benefit in severe PAD patients (interaction p = 0.03). Sensitivity analyses, including competing-risk models, affirmed the robustness of the severity–outcome association. Conclusion The IWGDF/IDSA classification effectively stratifies DFI patients by integrated metabolic, vascular, microbiological, and prognostic risk. Osteomyelitis warrants distinct management. The developed nomogram provides a validated tool for individualized risk prediction, facilitating early targeted intervention. These findings advocate for a severity-guided, multidisciplinary treatment paradigm to improve outcomes.