[Development and validation of a risk prediction model for cardiogenic shock occurrence in acute myocardial infarction].

Objective

To explore the value of model based on the stress hyperglycemia ratio (SHR) in predicting acute myocardial infarction (AMI) complicated by cardiogenic shock (CS).

Methods

This was a retrospective cross-sectional study. Patients diagnosed with AMI from the MIMIC-Ⅳ 3.0 database in the United States between 2008 and 2022 were included and randomly divided into a training set (1 861 cases) and an internal validation set (799 cases) at a 7∶3 ratio. Additionally, eligible AMI patients from Qingdao Municipal Hospital between January 1, 2021, and February 1, 2025, were included as an external test set (316 cases). Key factors were screened using the Least Absolute Shrinkage and Selection Operator (LASSO) regression. Univariate and multivariate logistic regression models were used to identify factors influencing the occurrence of CS in AMI patients, and a nomogram prediction model based on SHR was established. The predictive performance of the model was evaluated using the area under the receiver operating characteristic (ROC) curve (AUC), calibration curve, and decision analysis (DCA) curve.

Results

In the training set, patients had a median age of 69 (61, 77) years, with 1 293 males, and 16.7% (310/1 861) had concurrent CS. In the internal validation set, patients had a median age of 69 (61, 77) years, with 550 males, and 18.3% (146/799) had concurrent CS. In the external test set, patients had a median age of 72 (64, 80) years, with 199 males, and 11.1% (35/316) had concurrent CS. Multivariate logistic regression analysis indicated that systolic blood pressure (SBP), SHR, white blood cell (WBC) count, hematocrit (HCT), aspartate aminotransferase (AST), anion gap (AG), activated partial thromboplastin time (APTT), heart failure (HF), and acute kidney injury (AKI) were all influencing factors for the occurrence of CS in AMI patients (all P<0.05). A nomogram model based on these nine variables demonstrated an AUC of 0.82 (95%CI: 0.80-0.84), a sensitivity of 0.68 and a specificity of 0.82 in the training set for predicting CS in AMI patients; an AUC of 0.79 (95%CI: 0.75-0.83), a sensitivity of 0.64 and a specificity of 0.77 in the internal validation set; and an AUC of 0.84 (95%CI: 0.77-0.92), a sensitivity of 0.77 and a specificity of 0.80 in the external test set. Calibration curves indicated good consistency across all datasets, and DCA curve demonstrated that the nomogram model had excellent clinical applicability.

Conclusions

SHR is an influencing factor for CS in AMI patients. The nomogram model developed using SBP, SHR, WBC, HCT, AST, AG, APTT, HF and AKI provides a more intuitive method for identifying the risk of CS in AMI patients.