Host Susceptibility Phenotypes and Construction of an Inflammation–Vascular–Metabolic (IVM) Integrated Score for Hard-to-Heal Wounds After Low-Energy Foot Trauma in Older Adults

Lijun XiuTiantian LiuZhaoxia JiXinzhe ZhangGuanghai YuanJian LiPeng YangDepartment of Hand and Foot Surgery, Qingdao Eighth People’s Hospital, Qingdao, Shandong, 266100, People’s Republic of ChinaDepartment of Ophthalmology, Qingdao Eighth People’s Hospital, Qingdao, Shandong, 266100, People’s Republic of ChinaCorrespondence: Peng Yang, Department of Hand and Foot Surgery, Qingdao Eighth People’s Hospital, 84 Fengshan Road, Licang District, Qingdao, Shandong, 266100, People’s Republic of China, Email [email protected]: Older adults with low-energy trauma-initiated foot wounds often become hard-to-heal, but prognostic tools tailored to this scenario are scarce.Methods: In this single-center retrospective cohort, we included consecutive patients aged ≥ 65 years with below-malleoli low-energy trauma-initiated foot wounds. Baseline predictors were the first laboratory values within 48 h and an ordinal perfusion/ischemia category derived from available ABI, pedal Doppler waveform findings, and toe-based measures. Hard-to-heal was failure of complete epithelialization without drainage by 12 weeks (major amputation classified as not healed). Latent class analysis defined susceptibility phenotypes. An interpretable 6-variable inflammation–vascular–metabolic (IVM) model was developed using logistic regression with multiple imputation and bootstrap internal validation.Results: Among 697 patients (median age 74 years), 332 (47.6%) were hard-to-heal at 12 weeks. Phenotypes showed graded risk (lowest to highest): relatively resilient (n=143, hard-to-heal 26/143), renal-hypoalbuminemia (n=165, hard-to-heal 81/165), metabolic–inflammatory (n=206, hard-to-heal 110/206), and ischemia-dominant (n=183, hard-to-heal 115/183). Adjusted odds ratios versus resilient were 3.8 (95% CI 2.3– 6.4), 4.5 (2.8– 7.4), and 6.1 (3.7– 10.1), respectively. The IVM model achieved optimism-corrected AUC 0.80 (95% CI 0.77– 0.83) with calibration slope 0.94 (intercept − 0.03), stratified observed event rates to 19.7% (50/254), 48.8% (137/281), and 89.5% (145/162), and showed net benefit across threshold probabilities ~0.25– 0.70.Conclusion: In this retrospective single-center cohort, distinct susceptibility phenotypes were identifiable and the IVM score provided internally validated early risk stratification. External validation and prospective evaluation are warranted before routine clinical implementation.Keywords:wound healing, older adults, low-energy trauma, perfusion, inflammation, latent class analysis, risk prediction