Routine blood-derived inflammatory and metabolic indices in diabetic patients with and without documented cardiovascular comorbidities: a large retrospective exploratory study
Year: 2026
Type: article
Study design: Observational Study
Abstract: Routine blood-derived inflammatory and metabolic indices are increasingly investigated as accessible markers of cardiometabolic status in diabetes. However, their clinical interpretation in large retrospective datasets remains challenging, particularly when statistically significant differences may not correspond to meaningful discrimination. This study evaluated cross-sectional differences in routinely available laboratory indices between patients with diabetes with and without documented cardiovascular comorbidity (CVC) and explored their ability to distinguish the two existing clinical groups. A retrospective exploratory analysis was performed using laboratory and diagnostic data from 7034 patients. Patients were classified as diabetes mellitus without documented CVC (DM, n = 5114) or diabetes with documented CVC (DM + CVC, n = 1920). Routine laboratory variables and derived indices, including triglyceride/glucose index (TyG index), monocyte-to-HDL ratio (MHR), triglyceride-to-HDL ratio (THR), systemic inflammation response index (SIRI), systemic immune-inflammation index (SII), pan-immune inflammation value and its haemoglobin normalised index (PIV, PIV/Hb), neutrophil-to-lymphocyte ratio (NLR), monocyte-to-lymphocyte ratio (MLR) and platelet-to-lymphocyte ratio (PLR), were evaluated using univariable comparisons, robust effect-size estimates, FDR-adjusted multivariable linear and logistic regression, exploratory threshold-based odds ratios, and cross-validated machine-learning classification. Exploratory classifiers were evaluated using stratified 10-fold cross-validation with fold-specific preprocessing and class-imbalance-aware performance measures. Given the predominance of primary hypertension within the heterogeneous DM + CVC group, a complementary primary analysis excluded patients whose only documented cardiovascular comorbidity was hypertension. Several parameters differed significantly between DM and DM + CVC groups; however, most effect sizes were negligible or small. MPV and PDW showed comparatively larger haematological differences than the other parameters, but their limited effect sizes indicated substantial overlap between the groups. In adjusted analyses, higher PDW, TyG, MHR, THR, and MLR were associated with higher odds of existing DM + CVC status, whereas higher MPV, SIRI, and NLR were associated with lower odds. Most effect sizes remained negligible or small. After exclusion of hypertension-only patients, MHR, THR, SIRI, NLR, and MLR retained the same direction and statistical interpretation, whereas TyG was attenuated and PIV/Hb became significant. Machine-learning discrimination was poor to modest, with ROC AUC values of 0.544–0.668 and balanced accuracy of 0.526–0.599, with Naïve Bayes modelling being the most balanced among the evaluated models, although still insufficient for individual-level classification. Routine blood-derived indices provide useful population-level information on inflammatory and cardiometabolic laboratory patterns in diabetes, but most markers showed limited stand-alone discriminatory value for documented CVC. These hypothesis-generating findings require prospective validation using longitudinal outcomes, complete clinical and medication data, and formal assessment of incremental clinical utility.
Source: Scientific Reports
Authors: Boglárka Kovács–Deák(Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Resources, Writing – original draft), Béla Kovács, Francisc Boda(Data curation, Formal analysis, Visualization, Writing – original draft), István Szakács Mikes(Data curation, Formal analysis, Funding acquisition, Resources, Visualization, Writing – original draft), Melinda Kolcsár(Project administration, Supervision, Writing – review & editing), Ibolya Fülöp(Project administration, Supervision, Writing – review & editing), Piroska Kelemen(Project administration, Supervision, Writing – review & editing), Előd Ernő Nagy(Funding acquisition, Project administration, Resources, Supervision, Writing – review & editing)
Language: English
Subfield: Oncology
Field: Medicine
Domain: Health Sciences
Keywords: diabetes mellitus, cardiovascular comorbidity, inflammatory markers, metabolic biomarkers, triglyceride-glucose index, monocyte-to-HDL cholesterol ratio, triglyceride to HDL cholesterol ratio, systemic immune-inflammation index, systemic inflammation response index, neutrophil-to-lymphocyte ratio, monocyte-to-lymphocyte ratio, platelet-to-lymphocyte ratio, plateletcrit, mean platelet volume, platelet distribution width, cardiometabolic risk, machine learning classification, essential hypertension
Open Access status: gold
Estimated APC paid: $2,690
OpenAlex ID: W7235158073
