In obese patients with AMI-CS undergoing PCI, heart rate, hypertension, diabetes mellitus, malignant arrhythmia, and multivessel disease were associated with in-hospital mortality, and logistic regression showed more favorable internally validated predictive performance than regularized XGBoost using the same routinely available predictors; findings are exploratory and internally validated only.
Among 265 obese patients with documented AMI-CS treated with PCI, in-hospital mortality was 63 deaths (23.8%).
Higher heart rate, hypertension, diabetes mellitus, malignant arrhythmia, and multivessel disease were associated with in-hospital mortality.
The seven-variable logistic regression model outperformed regularized XGBoost in internally validated discrimination and calibration.
Serial lactate concentrations were consistently higher in non-survivors, and cardiac power output was modestly lower.
A post hoc severity-augmented logistic model had high discrimination but poor calibration, so it was not sufficiently calibrated for clinical deployment.
The authors suggest diabetes, hypertension, and perfusion abnormalities are biologically compatible with endothelial dysfunction, but endothelial function was not measured.
The findings are exploratory and limited by sample size, lack of external validation, and outcome-informed predictor selection, precluding claims of clinical readiness.
This study looked at 265 obese patients who had a heart attack complicated by cardiogenic shock (when the heart cannot pump enough blood) and were treated with a procedure to open blocked arteries (PCI). About 24% (63 patients) died in the hospital. Higher heart rate, high blood pressure, diabetes, dangerous heart rhythm problems, and disease in multiple coronary vessels were linked with a higher chance of dying in the hospital. Blood lactate, a marker of poor tissue blood flow, was higher in those who died at every time point measured over 24 hours. The researchers also compared a traditional statistical model (logistic regression) with a machine-learning method (XGBoost) for predicting death using seven routine measures. The traditional model did somewhat better (AUC 0.767 vs 0.712, where 1.0 is perfect). A further model that added shock severity information looked much more accurate (AUC 0.922), but it was poorly calibrated, meaning its predicted risks were not reliable enough for use in practice. This research suggests that simple, routinely collected information may help identify high-risk patients, but the results are preliminary. The study came from one hospital, was retrospective, had a modest number of patients, and was not tested in other settings. The authors speculate that damage to blood vessel lining (endothelial dysfunction) may play a role, but they did not measure it. Larger multicenter studies are needed before any such model could be used in clinical care.
Maimaitiniyazi M, Yisimitila T, Maimaitiniyazi M, Maisuti M, Haibier A, Yushanjiang P, et al.. (2026). Exploratory clinical predictors and internally validated models of in-hospital mortality after PCI in obese patients with acute myocardial infarction and documented cardiogenic shock: a single-center retrospective study.. BMC cardiovascular disorders. https://doi.org/10.1186/s12872-026-06838-x