Construction and validation of predictive models for major adverse cardiovascular events in patients with acute coronary syndrome

Objective To establish a nomograph model based on heart rate variability (HRV) and evaluate the risk of major adverse cardiovascular events (MACEs) in patients with acute coronary syndrome (ACS).

Methods A total of 322 patients with confirmed ACS were retrospectively included as the modeling set, and 164 patients with ACS were selected as the validation set. The median follow-up time was 23.5 months. The modeling set was divided into MACEs group(n=30) and non-MACEs group(n=292). The risk factors were screened by single-factor comparison and multi-factor Cox regression analysis respectively, and the nomogram model was established. Receiver operator characteristic (ROC) curve, calibration curve and decision curve were used to verify the model, and Kaplan-Meier survival curve was used for risk stratification.

Results There were statistically significant differences in gender, age, ACS type, target vessel diameter stenosis rate, stent placement and incidence of MACEs between the modeling set and the validation set (P < 0.05). Univariate comparison showed that the age of the MACEs group was significantly higher, the proportion of ST-segment elevation myocardial infarction (STEMI) and history of myocardial infarction, admission creatine kinase isoenzyme (CK-MB), troponin I (cTnI) and B-type natriuretic peptide (BNP), target vessel diameter stenosis rate and standardized high frequency (HF) value were significantly higher, the standardized low frequency (LF) value in HRV index and the standardized LF/HF were significantly lower than those in the non-MACEs group (P < 0.05). Multivariate Cox regression analysis showed that age ≥65 years (HR=1.425; 95%CI, 1.124 to 1.758; P=0.001) and previous history of myocardial infarction (HR=1.326; 95%CI, 1.102 to 1.659; P=0.003) and standardized LF/HF < 1.32 (HR=2.203; 95%CI, 1.568 to 2.659; P < 0.001) were independent risk factors for MACEs in ACS patients followed up for two years. R software was used to establish a nomogram model, with a total score of 200 points. ROC showed that the area under the curve (AUC) of the 1-year and 2-year MACEs predicted by the nomogram model was 0.845 and 0.902, respectively, and the AUC of 1-year and 2-year MACEs for the prediction validation set was 0.802 and 0.856, respectively. Both the calibration curve and the decision curve showed that the model had good consistency and net benefit ratio. Kaplan-Meier survival curve showed that the cumulative incidence of MACEs at 1-year and 2-year high-risk (≥80 points) in modeling set and validation set was higher than that at medium-risk (60 to 80 points) and low-risk (< 60 points), and low-risk was the lowest (P < 0.05).

Conclusion Early detection of HRV in ACS patients is beneficial to accurately assess the risk of MACEs in the medium- and long-term follow-up after intervention. The age of ACS patients, past history of MI combined with standardized LF/HF map model has important application potential in assessing prognostic risk stratification.