The triglyceride-glucose index and neutrophil-to-lymphocyte ratio jointly predict the no-reflow phenomenon in T2DM patients with STEMI after primary PCI

Figures

• 

  Fig 1. Patient flowchart. STEMI, acute ST-segment elevation myocardial infarction; PPCI, primary percutaneous coronary intervention; T2DM, type 2 diabetes mellitus.

• 

  Fig 2. Scatter plot analyzing the correlation between TyG and NLR. Pearson correlation analysis reveals a negligible and non-significant positive linear association between the two biomarkers (r = 0.05, 95% CI: −0.04 to 0.14, P = 0.25). The absence of substantial collinearity justifies their simultaneous entry into the multivariable logistic regression model as independent predictors. NLR, neutrophil-to-lymphocyte ratio; TyG, triglyceride-glucose.

• 

  Fig 3. Receiver operating characteristic (ROC) curves of the NLR and TyG for predicting the no-reflow phenomenon. The TyG index (red line) demonstrated an area under the curve (AUC) of 0.70 (95% CI: 0.63-0.78), with an optimal cut-off value of 9.347. The NLR (blue line) yielded an AUC of 0.65 (95% CI: 0.57-0.72), with an optimal cut-off of 2.831. NLR, neutrophil-to-lymphocyte ratio; TyG, triglyceride-glucose.

• 

  Fig 4. Receiver operating characteristic (ROC) curves of sequential multivariable models for predicting the no-reflow phenomenon. The baseline model of age and pain-to-balloon time (Model 1, AUC = 0.700) was incrementally improved by adding NLR (Model 2, AUC = 0.744) and the TyG index (Model 3, AUC = 0.757). The full model (Model 4), containing all predictors, yielded the highest AUC of 0.785. Model 1:age and pain-to-balloon time; Model 2: Model 1 + NLR; Model 3: Model 1 + TyG；Model 4: Model 1 + NLR + TyG.

• 

  Fig 5. Calibration and decision curve analyses for the final model (Model 4). Model 4 included four predictors: age, pain-to-balloon time, neutrophil-to-lymphocyte ratio (NLR), and triglyceride-glucose (TyG) index. (A) Calibration plot. The solid line shows good agreement between predicted and observed probabilities of no-reflow, indicating good calibration. (B) Decision curve analysis. The model provides positive net benefit across threshold probabilities of 0–10%, outperforming both “treat all” and “treat none” strategies, supporting its clinical utility.