Electrocardiogram-based mortality prediction in patients with COVID-19 using machine learning

Abstract

Background and purpose: The electrocardiogram (ECG) is frequently obtained in the work-up of COVID-19 patients. So far, no study has evaluated whether ECG-based machine learning models have added value to predict in-hospital mortality specifically in COVID-19 patients.

Methods: Using data from the CAPACITY-COVID registry, we studied 882 patients admitted with COVID-19 across seven hospitals in the Netherlands. Raw format 12-lead ECGs recorded within 72 h of admission were studied. With data from five hospitals (n = 634), three models were developed: (a) a logistic regression baseline model using age and sex, (b) a least absolute shrinkage and selection operator (LASSO) model using age, sex and human annotated ECG features, and (c) a pre-trained deep neural network (DNN) using age, sex and the raw ECG waveforms. Data from two hospitals (n = 248) was used for external validation.

Results: Performances for models a, b and c were comparable with an area under the receiver operating curve of 0.73 (95% confidence interval [CI] 0.65-0.79), 0.76 (95% CI 0.68-0.82) and 0.77 (95% CI 0.70-0.83) respectively. Predictors of mortality in the LASSO model were age, low QRS voltage, ST depression, premature atrial complexes, sex, increased ventricular rate, and right bundle branch block.

Conclusion: This study shows that the ECG-based prediction models could be helpful for the initial risk stratification of patients diagnosed with COVID-19, and that several ECG abnormalities are associated with in-hospital all-cause mortality of COVID-19 patients. Moreover, this proof-of-principle study shows that the use of pre-trained DNNs for ECG analysis does not underperform compared with time-consuming manual annotation of ECG features.

Keywords: Arrhythmia; COVID-19; Deep learning; Electrocardiogram; Machine learning; Mortality.

Fig. 1

Variable importance for the manually extracted ECG features in the LASSO model. Coefficients represent the beta coefficients of the normalised variables in the LASSO model and can there be interpreted as importance values. Negative values point at a lower risk of mortality

Fig. 2

Predicted probability for the deep neural network compared with the LASSO algorithm with manually extracted ECG features, with the probability cut-offs of 30 and 25%, respectively. Inspection of the ECGs in the right lower corner (i.e. correct predictions by the DNN and not the LASSO) showed frequent tachycardia and low QRS voltage that did not meet the criteria, while the age was mostly below 70 years. Inspection of the left upper corner showed that these ECGs were normal, but patients had a high age of up to 93. (DNN deep neural network, ECG electrocardiogram, LASSO least absolute shrinkage and selection operator)