Objective To construct and validate a nomogram model based on early clinical features to predict the risk of in-hospital cardiac arrest (IHCA) in patients with acute myocardial infarction (AMI).

Methods A total of 152 AMI patients who were admitted to Beijing Friendship Hospital, Capital Medical University from January 2013 to December 2023 and experienced IHCA were retrospectively enrolled (IHCA group), with 608 contemporaneous AMI patients free of IHCA matched as the control gropu at a 4∶1 ratio (non-IHCA group). Patients were randomly divided into a training set and a validation set at a 7∶3 ratio. LASSO regression and multivariable logistic regression were used for variable selection and model construction in the training set. A nomogram was constructed based on the final model, and its performance was evaluated in the training set and validation set.

Results The proportion of patients with Killip class≥Ⅱ was higher in the IHCA group than that in the non-IHCA group in both the training set and validation set (89.6% vs 34.0%, 95.7% vs 36.3%; P＜0.001). The final model included nine variables: troponin T, triiodothyronine (T₃) uptake, blood glucose, pulmonary artery systolic pressure (PASP), total bilirubin, estimated glomerular filtration rate (eGFR), eosinophil percentage, pre-hospital β-blocker use, and in-hospital P2Y12 inhibitor use. The nomogram model achieved an AUC of 0.975 (95%CI 0.963–0.985) in the training set and 0.959 (95%CI 0.932–0.981) in the validation set, higher than the global registry of acute coronary event (GRACE) score in both datasets (P＜0.001).

Conclusion A nomogram model established based on early clinical indicators at admission could provide reference for predicting the risk of IHCA in patients with AMI.