Untargeted Plasma Lipidomics Identifies Oxidative Stress and Inflammation Signatures Associated with Recurrent Leg Ulcer Status in Sickle Cell Disease

Leg ulcers are among the most burdensome complications of sickle cell disease (SCD), yet the molecular determinants underlying ulcer susceptibility and recurrence remain incompletely characterized. We performed untargeted plasma lipidomics by LC–MS in 129 individuals with SCD from the REDS-III Brazil cohort, comparing participants with a recurrent sickle cell leg ulcer (SCLU) phenotype (n = 72) and SCD controls without ulcer history or follow-up episodes (n = 57). Data were split into a training set (80%; SCLUs n = 58, controls n = 46) and an independent validation set (20%; SCLUs n = 14, controls n = 11), with leakage-resistant preprocessing and variable selection restricted to the training set. After quality filtering, 2,419 features were retained. Univariate screening in the training set yielded 181 candidates, and 44 annotated metabolites were retained for visualization and modelling. Recurrent SCLUs showed higher lysophospholipids and oxylipins and lower sphingolipids and FAHFAs, consistent with inflammation, oxidative stress and membrane dysregulation. Random Forest achieved the best validation discrimination (AUC = 0.85; accuracy = 0.72; sensitivity = 0.86; specificity = 0.55). Untargeted plasma lipidomics defines an interpretable lipid signature linked to recurrent SCLUs and nominates candidates for targeted quantification and external validation to support individualized risk stratification.