Dataset Information

Clinical Implication of Hypoxic Liver Injury for Predicting Hypoxic Hepatitis and In-Hospital Mortality in ST Elevation Myocardial Infarction Patients.

ABSTRACT:

Purpose

In this study, we aimed to determine the value of hypoxic liver injury (HLI) in the emergency room (ER) for predicting hypoxic hepatitis (HH) and in-hospital mortality in ST elevation myocardial infarction (STEMI) patients.

Materials and methods

1537 consecutive STEMI patients were enrolled. HLI in the ER was defined as a ≥2-fold increase in serum aspartate transaminase (AST). HH was defined as a ≥20-fold increase in peak serum transaminase. Patients were divided into four groups according to HLI and HH status (group 1, no HLI or HH; group 2, HLI, but no HH; group 3, no HLI, but HH; group 4, both HLI and HH).

Results

The incidences of HLI and HH in the ER were 22% and 2%, respectively. In-hospital mortality rates were 3.1%, 11.8%, 28.6%, and 47.1% for groups 1, 2, 3, and 4, respectively. Patients with HLI and/or HH had worse Killip class, higher cardiac biomarker elevations, and lower left ventricular ejection fraction. Multivariate logistic regression analysis showed that HLI in the ER was an independent predictor of HH [odds ratio 2.572, 95% confidence interval (CI) 1.166-5.675, p=0.019]. The predictive value of HLI in the ER for the development of HH during hospitalization was favorable [area under the curve (AUC) 0.737, 95% CI 0.643-0.830, sensitivity 0.548, specificity 0.805, for cut-off value AST >80]. Furthermore, in terms of in-hospital mortality, predictive values of HLI in the ER and HH during hospitalization were comparable (AUC 0.701 for HLI at ER and AUC 0.674 for HH).

Conclusion

Among STEMI patients, HLI in the ER is a significant predictor for the development of HH and mortality during hospitalization (INTERSTELLAR ClinicalTrials.gov number, NCT02800421).

SUBMITTER: Choi SH

PROVIDER: S-EPMC8470566 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:BackgroundMachine learning (ML) risk prediction models, although much more accurate than traditional statistical methods, are inconvenient to use in clinical practice due to their nontransparency and requirement of a large number of input variables.ObjectiveWe aimed to develop a precise, explainable, and flexible ML model to predict the risk of in-hospital mortality in patients with ST-segment elevation myocardial infarction (STEMI).MethodsThis study recruited 18,744 patients enrolled in the 2013 China Acute Myocardial Infarction (CAMI) registry and 12,018 patients from the China Patient-Centered Evaluative Assessment of Cardiac Events (PEACE)-Retrospective Acute Myocardial Infarction Study. The Extreme Gradient Boosting (XGBoost) model was derived from 9616 patients in the CAMI registry (2014, 89 variables) with 5-fold cross-validation and validated on both the 9125 patients in the CAMI registry (89 variables) and the independent China PEACE cohort (10 variables). The Shapley Additive Explanations (SHAP) approach was employed to interpret the complex relationships embedded in the proposed model.ResultsIn the XGBoost model for predicting all-cause in-hospital mortality, the variables with the top 8 most important scores were age, left ventricular ejection fraction, Killip class, heart rate, creatinine, blood glucose, white blood cell count, and use of angiotensin-converting enzyme inhibitors (ACEIs) and angiotensin II receptor blockers (ARBs). The area under the curve (AUC) on the CAMI validation set was 0.896 (95% CI 0.884-0.909), significantly higher than the previous models. The AUC for the Global Registry of Acute Coronary Events (GRACE) model was 0.809 (95% CI 0.790-0.828), and for the TIMI model, it was 0.782 (95% CI 0.763-0.800). Despite the China PEACE validation set only having 10 available variables, the AUC reached 0.840 (0.829-0.852), showing a substantial improvement to the GRACE (0.762, 95% CI 0.748-0.776) and TIMI (0.789, 95% CI 0.776-0.803) scores. Several novel and nonlinear relationships were discovered between patients' characteristics and in-hospital mortality, including a U-shape pattern of high-density lipoprotein cholesterol (HDL-C).ConclusionsThe proposed ML risk prediction model was highly accurate in predicting in-hospital mortality. Its flexible and explainable characteristics make the model convenient to use in clinical practice and could help guide patient management.Trial registrationClinicalTrials.gov NCT01874691; https://clinicaltrials.gov/study/NCT01874691.

Project description:In an era of increasing need for precision medicine, machine learning has shown promise in making accurate acute myocardial infarction outcome predictions. The accurate assessment of high-risk patients is a crucial component of clinical practice. Type 2 diabetes mellitus (T2DM) complicates ST-segment elevation myocardial infarction (STEMI), and currently, there is no practical method for predicting or monitoring patient prognosis. The objective of the study was to compare the ability of machine learning models to predict in-hospital mortality among STEMI patients with T2DM. We compared six machine learning models, including random forest (RF), CatBoost classifier (CatBoost), naive Bayes (NB), extreme gradient boosting (XGBoost), gradient boosting classifier (GBC), and logistic regression (LR), with the Global Registry of Acute Coronary Events (GRACE) risk score. From January 2016 to January 2020, we enrolled patients aged > 18 years with STEMI and T2DM at the Affiliated Hospital of Zunyi Medical University. Overall, 438 patients were enrolled in the study [median age, 62 years; male, 312 (71%); death, 42 (9.5%]). All patients underwent emergency percutaneous coronary intervention (PCI), and 306 patients with STEMI who underwent PCI were enrolled as the training cohort. Six machine learning algorithms were used to establish the best-fit risk model. An additional 132 patients were recruited as a test cohort to validate the model. The ability of the GRACE score and six algorithm models to predict in-hospital mortality was evaluated. Seven models, including the GRACE risk model, showed an area under the curve (AUC) between 0.73 and 0.91. Among all models, with an accuracy of 0.93, AUC of 0.92, precision of 0.79, and F1 value of 0.57, the CatBoost model demonstrated the best predictive performance. A machine learning algorithm, such as the CatBoost model, may prove clinically beneficial and assist clinicians in tailoring precise management of STEMI patients and predicting in-hospital mortality complicated by T2DM.