Integrating inflammatory and cardiac biomarkers for predicting long-term mortality in type 2 diabetes: insights from machine learning, platelet function, and bioinformatics analyses

Abstract Introduction IL-6, NT-proBNP, and TNF-α are liquid biomarker that are elevated in individuals with type 2 diabetes mellitus (T2DM), especially those with cardiovascular complications, reflecting increased heart stress due to factors like high blood sugar, inflammation, and hypertension, making it a useful biomarker for detecting cardiovascular risk and mortality. The aim of our study was to determine the ability of IL-6, NT-proBNP, and TNF-α to predict all-cause mortality and their association with platelet function among T2DM patients. Methodology: Three hundred-three patients with diabetes were enrolled in the study. Among the patients included, 34 (11.2 %) patients died within a median observation time of 5.9 years. Serum IL-6, NT-proBNP, and TNF-α concentrations were assessed by ELISA immunoassay and compared between the patients who survived and those who died. Platelet functions were determined via standardized platelet aggregation test, with 515 was taken as a 4th quartile cut-off to specify the patients with hyper vs normal platelet activity. Results Receiver operating characteristic (ROC) curve analysis showed IL-6, NT-proBNP, and TNF-α as mortality predictions (AUC: 0.653, p= 0.004; AUC: 0.688, p0.001; AUC:0.657, p=0.003 respectively). Combining all biomarkers yielded a higher AUC than the value of each individual biomarker (AUC: 0.697, p=0.0002). Adjusted Cox-regression analysis was used for prediction of all-cause mortality. After including all blood-biomarkers into one multivariate Cox regression model, combined value predicted the future occurrence of long-term all-cause mortality as the most significant (HR= 4.10, 95% CI: 1.9-8.8; p= 0.0003). Besides, patients with hyper platelet activity had significantly higher TNF-α concentration compared to patients with normal platelet activity (p= 0.019). We utilized SHAP analysis to uncover how laboratory features contribute to predictions, confirming IL-6 as the top regulator of high mortality risk. Interaction and enrichment analyses identified 34 shared targets for IL-6, TNF, and NPPB, linking them to cardiovascular diseases (CVD) and pathways such as cytokine signaling, vitamin B12 metabolism, and heart-specific processes like dilated cardiomyopathy. Conclusion To conclude, combining different biomarkers of processes underlying cardiovascular pathophysiology might be beneficial for early diagnosis of all-cause mortality and IL-6, NT-proBNP, and TNF-α combination is a strong and independent predictor of long-term all-cause mortality among patients presenting with T2DM. Moreover, TNF-α presents significant interactions with platelet function in patients with stable diabetes.All ResultsFor image description, please refer to the figure legend and surrounding text.