Noninvasive Wound Swab Proteomics Enables Prognostic Stratification of Diabetic Foot Ulcer Healing: The HEAL-DFU Proof-of-Concept Study
- 1Steno Diabetes Center Copenhagen, Herlev, Denmark
Abstract
ObjectiveDiabetic foot ulcers (DFUs) affect up to 34% of people with diabetes and precede 80% of lower extremity amputations, yet no ulcer-level molecular biomarkers distinguish healing trajectories. We aimed to identify wound swab proteins associated with DFU healing time and characterize their prognostic potential.
Research Design and MethodsLC-MS/MS proteomics was performed on wound swabs from 112 people with diabetes presenting with new DFUs, followed for up to 6 months. Participants were divided into fast-healers ([≤]40), slow-healers (41-180 days) and chronic wounds (>180 days). Proteins associated with "time to healing" were identified using Firth-penalized Cox regression (FDR<0.2; HR<1 indicated slow healing) and used to construct a cross-validated Protein Healing Score (PHS). Sensitivity and causal mediation analyses examined clinical variable effects on protein abundance, with independent dataset validation.
ResultsParticipants had a mean ({+/-}SD) age of 67.5 ({+/-}12.3) years, and diabetes duration of 27.5 ({+/-}14.9) years. Twenty-seven proteins formed two clusters: a slow-healing cluster enriched in neutrophil granule and innate immune proteins (HR<1), and a fast-healing cluster enriched in proteostatic stress proteins (HR>1). Triglyceride levels differed between healing groups (p=0.001); seven slow-healing proteins showed triglyceride-mediated effects (21-28%), suggesting a dyslipidaemia-innate immunity axis. The PHS significantly stratified prognosis (adjusted HR=1.47, 95%CI 1.14-1.89; p=0.003), outperforming clinical variables alone (AUC-OVR 0.71 vs. 0.62). Transcriptomic validation confirmed directional concordance.
ConclusionsWound swab proteomics at DFU onset identified a 27-protein healing signature linked to neutrophil-innate immunity and triglyceride associated dyslipidaemia, offering a potential non-invasive prognostic tool.
HighlightsO_LIWhy did we undertake this study? No prior study has examined wound swab protein associations with DFU healing time as a continuous outcome, offering finer prognostic resolution than discrete healing categories. C_LIO_LIWhat is the specific question(s) we wanted to answer? Can wound swab proteins at DFU presentation predict healing trajectories, and do clinical factors mediate these associations? C_LIO_LIWhat did we find? Twenty-seven proteins formed two prognostic clusters: a slow-healing cluster enriched in neutrophil/innate immune proteins and a fast-healing cluster linked to proteostatic stress. A Protein Healing Score outperformed clinical variables in stratifying healing outcomes (AUC-OVR 0.71 vs. 0.62). Triglyceride-mediated and neuropathy-driven effects on wound proteins revealed mechanistic links to impaired healing. C_LIO_LIWhat are the implications of our findings? Wound swab proteomics offers a non-invasive prognostic tool at DFU onset, potentially enabling earlier targeted intervention. C_LI
Related
- Plasma proteomics for prognostic stratification in newly diagnosed primary central nervous system lymphoma
- A multicohort proteomics-based machine learning score (MASHPredict) for noninvasive diagnosis and long-term prognostication of metabolic dysfunction-associated steatohepatitis
- PET/CT radiomics for noninvasive histologic risk stratification in resected lung adenocarcinoma: a single-system development and internal validation study.
- Identification of hepatitis B virus-associated membrane proteins as prognostic biomarkers for hepatocellular carcinoma based on retrospective analysis of integrated membrane proteomics
- Minute-Scale Sleep-Onset HRV as a Candidate Risk-Stratification Biomarker for Diabetic Retinopathy: An Exploratory Pilot Study
BibTeX
@misc{closgarcia2026noninvasive,
title = {{Noninvasive Wound Swab Proteomics Enables Prognostic Stratification of Diabetic Foot Ulcer Healing: The HEAL-DFU Proof-of-Concept Study}},
author = {M. Clos-Garcia and S. Y. Jain and C. H. Naaman and J. A. Raguz and Z. Schaefer and G. Kaur and V. R. Curovic and K. Kirketerp-Moller and J. A. Andersen and S. Engberg and A. Rasmussen and A. Veves and P. Rossing and T. S. Ahluwalia},
howpublished = {medRxiv},
year = {2026},
month = sep,
doi = {10.64898/2026.09.15.26363101},
url = {https://www.medrxiv.org/content/10.64898/2026.09.15.26363101v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - Noninvasive Wound Swab Proteomics Enables Prognostic Stratification of Diabetic Foot Ulcer Healing: The HEAL-DFU Proof-of-Concept Study
AU - M. Clos-Garcia
AU - S. Y. Jain
AU - C. H. Naaman
AU - J. A. Raguz
AU - Z. Schaefer
AU - G. Kaur
AU - V. R. Curovic
AU - K. Kirketerp-Moller
AU - J. A. Andersen
AU - S. Engberg
AU - A. Rasmussen
AU - A. Veves
AU - P. Rossing
AU - T. S. Ahluwalia
PY - 2026
DA - 2026/09/15
PB - medRxiv
DO - 10.64898/2026.09.15.26363101
UR - https://www.medrxiv.org/content/10.64898/2026.09.15.26363101v1
AB - ObjectiveDiabetic foot ulcers (DFUs) affect up to 34% of people with diabetes and precede 80% of lower extremity amputations, yet no ulcer-level molecular biomarkers distinguish healing trajectories. We aimed to identify wound swab proteins associated with DFU healing time and characterize their prognostic potential. Research Design and MethodsLC-MS/MS proteomics was performed on wound swabs from 112 people with diabetes presenting with new DFUs, followed for up to 6 months. Participants were divided into fast-healers ([≤]40), slow-healers (41-180 days) and chronic wounds (>180 days). Proteins associated with "time to healing" were identified using Firth-penalized Cox regression (FDR<0.2; HR<1 indicated slow healing) and used to construct a cross-validated Protein Healing Score (PHS). Sensitivity and causal mediation analyses examined clinical variable effects on protein abundance, with independent dataset validation. ResultsParticipants had a mean ({+/-}SD) age of 67.5 ({+/-}12.3) years, and diabetes duration of 27.5 ({+/-}14.9) years. Twenty-seven proteins formed two clusters: a slow-healing cluster enriched in neutrophil granule and innate immune proteins (HR<1), and a fast-healing cluster enriched in proteostatic stress proteins (HR>1). Triglyceride levels differed between healing groups (p=0.001); seven slow-healing proteins showed triglyceride-mediated effects (21-28%), suggesting a dyslipidaemia-innate immunity axis. The PHS significantly stratified prognosis (adjusted HR=1.47, 95%CI 1.14-1.89; p=0.003), outperforming clinical variables alone (AUC-OVR 0.71 vs. 0.62). Transcriptomic validation confirmed directional concordance. ConclusionsWound swab proteomics at DFU onset identified a 27-protein healing signature linked to neutrophil-innate immunity and triglyceride associated dyslipidaemia, offering a potential non-invasive prognostic tool. HighlightsO_LIWhy did we undertake this study? No prior study has examined wound swab protein associations with DFU healing time as a continuous outcome, offering finer prognostic resolution than discrete healing categories. C_LIO_LIWhat is the specific question(s) we wanted to answer? Can wound swab proteins at DFU presentation predict healing trajectories, and do clinical factors mediate these associations? C_LIO_LIWhat did we find? Twenty-seven proteins formed two prognostic clusters: a slow-healing cluster enriched in neutrophil/innate immune proteins and a fast-healing cluster linked to proteostatic stress. A Protein Healing Score outperformed clinical variables in stratifying healing outcomes (AUC-OVR 0.71 vs. 0.62). Triglyceride-mediated and neuropathy-driven effects on wound proteins revealed mechanistic links to impaired healing. C_LIO_LIWhat are the implications of our findings? Wound swab proteomics offers a non-invasive prognostic tool at DFU onset, potentially enabling earlier targeted intervention. C_LI
ER -