Cardiovascular

Factors associated with in-hospital heart failure after emergency percutaneous coronary intervention in patients with type 2 diabetes mellitus complicated by ST-segment elevation myocardial infarction: a retrospective study.

TL;DR

Readily available admission variables were independently associated with in-hospital heart failure in T2DM-STEMI patients undergoing emergency PCI, with the internally validated prediction model showing good discrimination (optimism-corrected AUC 0.843) and calibration, though external validation is required before clinical implementation.

Key Findings

Prior myocardial infarction was the strongest independent positive predictor of in-hospital heart failure in T2DM-STEMI patients undergoing emergency PCI.

  • OR = 6.205 (95% CI: 2.692–14.299, P < 0.001)
  • This was the largest odds ratio among all positive predictors in the multivariable model
  • Study included 306 adult patients with T2DM and STEMI, of whom 65 developed in-hospital HF and 241 did not
  • HF was defined as new-onset clinically documented HF during the index hospitalization with maximum Killip class ≥ II

Smoking was independently and strongly associated with increased odds of in-hospital heart failure.

  • OR = 5.552 (95% CI: 2.517–12.247, P < 0.001)
  • Smoking was the second largest positive predictor in the multivariable logistic regression model
  • Analysis was conducted using multivariable binary logistic regression on 306 patients

Age ≥ 75 years was independently associated with nearly fivefold greater odds of in-hospital heart failure.

  • OR = 4.792 (95% CI: 1.394–16.472, P = 0.013)
  • This variable was analyzed as a binary threshold (≥75 years vs. younger) in the multivariable model
  • The wide confidence interval reflects the relatively small number of events (n = 65 HF cases)

Regional wall motion abnormalities on echocardiography were independently associated with increased odds of in-hospital heart failure.

  • OR = 2.415 (95% CI: 1.091–5.344, P = 0.030)
  • Echocardiographic findings were used as predictors in the regression model but natriuretic peptides and echocardiographic findings were also used as supportive evidence for HF classification
  • This was among the smaller positive effect sizes in the multivariable model

Higher white blood cell count at admission was independently associated with greater odds of in-hospital heart failure.

  • OR = 1.143 per 10⁹/L increase (95% CI: 1.072–1.219, P < 0.001)
  • WBC count was treated as a continuous variable in the multivariable logistic regression
  • The per-unit OR suggests a meaningful cumulative effect across the range of observed WBC values

Higher systolic blood pressure at admission was independently associated with increased odds of in-hospital heart failure.

  • OR = 1.018 per mmHg increase (95% CI: 1.001–1.034, P = 0.032)
  • Systolic blood pressure was treated as a continuous variable in the multivariable model
  • The relatively small per-unit OR reflects the continuous scaling but indicates a cumulative association across the range of SBP values

Lower hemoglobin levels were independently associated with greater odds of in-hospital heart failure.

  • OR = 0.971 per 1 g/L increase (95% CI: 0.952–0.990, P = 0.003), indicating that lower hemoglobin confers higher HF odds
  • Hemoglobin was analyzed as a continuous variable
  • This inverse association suggests anemia or lower hemoglobin as a risk factor for post-PCI HF in this population

Lower left ventricular ejection fraction (LVEF) was independently associated with greater odds of in-hospital heart failure.

  • OR = 0.953 per 1 percentage-point increase (95% CI: 0.911–0.998, P = 0.040), indicating higher HF odds at lower LVEF values
  • LVEF was treated as a continuous variable in the multivariable regression
  • LVEF was included as a predictor; echocardiographic findings were also used as supportive evidence for HF classification but natriuretic peptides were not included as predictors

The multivariable prediction model demonstrated good discrimination and calibration with internal validation.

  • Apparent AUC = 0.881; optimism-corrected AUC (bootstrap) = 0.843
  • 10-fold cross-validation was also performed alongside bootstrap optimism correction
  • Hosmer-Lemeshow goodness-of-fit P = 0.826, indicating no significant lack of fit
  • The model was based on 65 HF events among 306 patients in a single-center retrospective design, and external validation was noted as required before clinical implementation

The study population consisted of 306 adult patients with T2DM and STEMI who underwent emergency PCI, with an in-hospital heart failure incidence of approximately 21%.

  • 65 of 306 patients (approximately 21.2%) developed new-onset in-hospital HF
  • Patients were treated at the People's Hospital of Xinjiang Uygur Autonomous Region between January 2022 and February 2026
  • All patients classified as having HF had a maximum Killip class ≥ II
  • Natriuretic peptides and echocardiographic findings were used as supportive evidence for HF classification and were not included as predictors in the model

What This Means

This research looked at patients with both type 2 diabetes and a serious type of heart attack (called STEMI) who received emergency artery-opening procedures (PCI) to restore blood flow to the heart. Among 306 such patients treated at a hospital in China's Xinjiang region, about 1 in 5 developed heart failure during their hospital stay. The researchers wanted to understand which factors, measurable at the time of admission, were linked to this complication. They found that having had a previous heart attack, being a smoker, being 75 years or older, having abnormal heart wall movement on ultrasound, having a higher white blood cell count, and having higher blood pressure on admission were all associated with a greater chance of developing heart failure. Conversely, patients with higher hemoglobin levels (a measure related to blood oxygen-carrying capacity) and better heart pumping function (higher ejection fraction) were less likely to develop heart failure. Using these eight factors together, the researchers built a prediction model that performed well statistically, correctly distinguishing patients who did and did not develop heart failure about 84% of the time after accounting for potential overfitting through internal testing methods. The model's predictions also aligned well with what actually happened to patients, suggesting it captures real patterns in the data rather than just random noise. This research suggests that commonly available clinical information gathered when a high-risk diabetic heart attack patient arrives at the hospital could help identify those most likely to develop heart failure during their stay. However, because this was a single-center study with a relatively small number of heart failure cases, the authors caution that the model needs to be tested in different hospital settings and patient populations before it could be used to guide clinical decisions. The findings may help direct attention toward modifiable and non-modifiable risk factors in this vulnerable patient group.

Have a question about this study?

Citation

Maimaitiniyazi M, Maimaitiniyazi M, Maisuti M, Yisimitila T, Yushanjiang P, Nijiati M. (2026). Factors associated with in-hospital heart failure after emergency percutaneous coronary intervention in patients with type 2 diabetes mellitus complicated by ST-segment elevation myocardial infarction: a retrospective study.. Frontiers in endocrinology. https://doi.org/10.3389/fendo.2026.1913869