Prediction of major adverse cardiovascular events in patients with acute coronary syndrome: Development and validation of a non-invasive nomogram model based on autonomic nervous system assessment

Abstract

Background: Disruption of the autonomic nervous system (ANS) can lead to acute coronary syndrome (ACS). We developed a nomogram model using heart rate variability (HRV) and other data to predict major adverse cardiovascular events (MACEs) following emergency coronary angiography in patients with ACS.

Methods: ACS patients admitted from January 2018 to June 2020 were examined. Holter monitors were used to collect HRV data for 24 h. Coronary angiograms, clinical data, and MACEs were recorded. A nomogram was developed using the results of Cox regression analysis.

Results: There were 439 patients in a development cohort and 241 in a validation cohort, and the mean follow-up time was 22.80 months. The nomogram considered low-frequency/high-frequency ratio, age, diabetes, previous myocardial infarction, and current smoking. The area-under-the-curve (AUC) values for 1-year MACE-free survival were 0.790 (95% CI: 0.702-0.877) in the development cohort and 0.894 (95% CI: 0.820-0.967) in the external validation cohort. The AUCs for 2-year MACE-free survival were 0.802 (95% CI: 0.739-0.866) in the development cohort and 0.798 (95% CI: 0.693-0.902) in the external validation cohort. Development and validation were adequately calibrated and their predictions correlated with the observed outcome. Decision curve analysis (DCA) showed the model had good discriminative ability in predicting MACEs.

Conclusion: Our validated nomogram was based on non-invasive ANS assessment and traditional risk factors, and indicated reliable prediction of MACEs in patients with ACS. This approach has potential for use as a method for non-invasive monitoring of health that enables provision of individualized treatment strategies.

Keywords: acute coronary syndrome; autonomic nervous system; major adverse cardiovascular events; nomogram; outcomes; prediction nomogram.

FIGURE 1

Disposition of patients with ACS (n = 1,368), and establishment of the development cohort (n = 439) and external validation cohort (n = 241).

FIGURE 2

Significant predictors of MACEs in the development cohort from multivariate Cox regression analysis.

FIGURE 3

Nomogram for predicting 1 and 2-year MACE-free survival, based on multivariate Cox regression analysis. For each patient, each clinical characteristic was assigned points by drawing a vertical line from its value to the top row. The points for all five characteristics were added to determine total points middle row. Then, the total points were used to determine the 1 and 2-year probabilities of MACE-free survival by drawing a vertical line to the bottom two rows.

FIGURE 4

ROC analysis of the accuracy of the nomogram in predicting 1-year MACE-free survival in the development cohort (A) and external validation cohort (B) and 2-year MACE-free survival in the development cohort (C) and external validation cohort (D).

FIGURE 5

Calibration curves for the prediction of the risk for 1-year MACE-free survival in the development cohort (A) and the external validation cohort (B) and 2-year MACE-free survival in the development cohort (C) and the external validation cohort (D). Each plot shows the relationship of the nomogram-predicted probability (abscissa) and the actual probability (ordinate) Each light blue diagonal line represents the ideal reference line (in which predicted survival probabilities match observed survival rates) and each red line was calculated by bootstrap resampling (500 times), and represents the performance of the nomogram. Thus, a greater similarity of the red and blue lines indicates more accurate predictions of survival.

FIGURE 6

Decision curve analysis (DCA) for predicting 1-year MACE-free survival in the development cohort (A) and the external validation cohort (B), and 2-year MACE-free survival in the development cohort (C) and the external validation cohort (D). Each plot shows the relationship of threshold probability (abscissa) with the net benefit (ordinate) of the prediction model (blue line), the proportion of patients with MACEs (red line), and the proportion of patients with MACE-free survival (green line).

FIGURE 7

Cumulative MACE-free survival of patients in different prognostic index (PI) tertiles. (A) Two year survival in the development cohort. (B) Two year survival in the external validation cohort. (C) One year survival the development cohort. (D) One year survival the external validation cohort. (E) One year survival in the complete cohort. (F) Two year survival in the complete cohort. Central illustration. Smart wearable electronic devices that use visual and personalized model-based risk assessment and an integrated approach that considers heart rate variability and clinical data (traditional risk factors) could allow patients to receive 24-h real-time personalized home telemonitoring and receive improved clinical management.