The combined prediction model integrating GLS with clinical indicators improved early prediction of post-PCI heart failure, increasing AUC from 0.803 to 0.861 (P < 0.001) and yielding NRI 0.216 and IDI 0.057 in patients with STEMI and type 2 diabetes.
Cohort (n=328)
No
Does integrating global longitudinal strain (GLS) with clinical indicators improve the early prediction of post-PCI heart failure in patients with STEMI and type 2 diabetes?
Effect estimate: AUC 0.861 vs 0.803; NRI 0.216; IDI 0.057 (95% CI Model 2 AUC 0.811–0.911; model 1 AUC 0.743–0.862; NRI 95% CI 0.107–0.605; IDI 95% CI 0.015–0.078)
p-value: p=<0.001
Background: Patients with ST-segment elevation myocardial infarction (STEMI) and type 2 diabetes mellitus (T2DM) are at increased risk of heart failure after percutaneous coronary intervention (PCI). Early identification of high-risk individuals remains challenging. This study aimed to develop a prediction model integrating two-dimensional speckle tracking imaging (2D-STI) and clinical variables to improve risk stratification. Methods: A total of 328 T2DM patients with STEMI who underwent PCI were retrospectively analyzed. Clinical, laboratory, and 2D-STI parameters were collected within one week after PCI. Heart failure within one year was the study endpoint. LASSO regression followed by Boruta analysis was used to identify key predictors. A multivariable logistic model was established, visualized by a nomogram, and evaluated using ROC curves, reclassification indices, calibration, and decision curve analysis. Results: Heart failure occurred in 62 patients (18.9%). Six variables—GLS, HbA1c, BMI, eGFR, hs-CRP, and diabetes duration—were identified as core predictors. GLS showed the highest individual discriminative ability (AUC = 0.798). The combined model achieved an AUC of 0.861, significantly outperforming the base model (AUC = 0.803, P < 0.001). Adding GLS improved reclassification (NRI = 0.216; IDI = 0.057). The model demonstrated good calibration and favorable clinical utility. Conclusion: Integrating GLS with clinical, metabolic, inflammatory, and renal indicators significantly improves early prediction of post-PCI heart failure in T2DM patients with STEMI, offering a practical tool for individualized risk assessment. Keywords: global longitudinal strain, two-dimensional speckle tracking imaging, ST-segment elevation myocardial infarction, type 2 diabetes mellitus, heart failure
Su et al. (Sun,) conducted a cohort in Patients with type 2 diabetes mellitus and ST-segment elevation myocardial infarction (STEMI) undergoing percutaneous coronary intervention (PCI) (n=328). Prediction model integrating global longitudinal strain (GLS) measured by 2D speckle tracking imaging and clinical indicators (HbA1c, BMI, eGFR, hs-CRP, diabetes duration) vs. Prediction model with clinical indicators only (HbA1c, BMI, eGFR, hs-CRP, diabetes duration) was evaluated on Occurrence of heart failure within one year after PCI, defined by hospitalization due to heart failure, elevated NT-proBNP with clinical signs, or new or worsening left ventricular dysfunction (AUC 0.861 vs 0.803; NRI 0.216; IDI 0.057, 95% CI Model 2 AUC 0.811–0.911; model 1 AUC 0.743–0.862; NRI 95% CI 0.107–0.605; IDI 95% CI 0.015–0.078, p=<0.001). The combined prediction model integrating GLS with clinical indicators improved early prediction of post-PCI heart failure, increasing AUC from 0.803 to 0.861 (P < 0.001) and yielding NRI 0.216 and IDI 0.057 in patients with STEMI and type 2 diabetes.