Diagnostic value of systemic immune-inflammation composite index combined with triglyceride-glucose index in type 2 diabetes patients with coronary heart disease: a retrospective diagnostic model study

BACKGROUND: The prevalence of coronary heart disease (CHD) among patients with type 2 diabetes mellitus (T2DM) has increased substantially. Early identification of high-risk individuals is critical for improving clinical outcomes. This study aimed to evaluate the diagnostic value of the systemic immune-inflammation composite index (SIICI) and the triglyceride-glucose (TyG) index for CHD in patients with T2DM, and to develop and validate a combined diagnostic model incorporating these indices. METHODS: We retrospectively enrolled 599 patients with T2DM who underwent coronary angiography and divided them into CHD (n = 371) and non-CHD (n = 228) groups based on angiographic findings. Using stratified 7:3 sampling according to CHD status, the cohort was randomly split into training (n = 439) and validation (n = 160) sets. Clinical and laboratory data were collected. Univariate logistic regression (P < 0.1) followed by backward stepwise multivariate logistic regression was performed to construct the diagnostic model. Model performance was evaluated using receiver operating characteristic (ROC) curves, calibration curves, and decision curve analysis (DCA). RESULTS: Compared with the non-CHD group, patients with CHD had significantly higher median levels of SIICI [5.24 (3.54, 8.05) vs. 3.75 (2.63, 4.76), P < 0.001] and TyG index [9.36 (8.96, 9.77) vs. 8.89 (8.60, 9.27), P < 0.001]. Multivariate logistic regression identified SIICI (OR = 1.347, 95% CI: 1.190-1.524) and TyG (OR = 2.843, 95% CI: 1.794-4.507) as independent risk factors for CHD. The combined model achieved an area under the curve (AUC) of 0.840 (95% confidence interval [CI]: 0.803-0.877) in the training set, with a sensitivity of 72.8% and specificity of 80.4%; in the validation set, the AUC was 0.851 (95% CI: 0.794-0.908), with a sensitivity of 75.8% and specificity of 80.0%. Calibration curves demonstrated good agreement (Hosmer-Lemeshow test: P = 0.597). DCA revealed positive net clinical benefit across a threshold probability range of 10-56%. CONCLUSIONS: Both SIICI and TyG are independent risk factors for CHD in patients with T2DM. The combined diagnostic model exhibits excellent discriminative ability, good calibration, and stable generalizability. While head-to-head comparisons with traditional risk scores are needed, this model offers a non-invasive tool that integrates inflammatory and metabolic information for cardiovascular risk stratification in T2DM patients.