ABSTRACT
BACKGROUND: Adverse events in COVID-19 are difficult to predict. Risk stratification is encumbered by the need to protect healthcare workers. We hypothesize that artificial intelligence (AI) can help identify subtle signs of myocardial involvement in the 12-lead electrocardiogram (ECG), which could help predict complications. OBJECTIVE: Use intake ECGs from COVID-19 patients to train AI models to predict risk of mortality or major adverse cardiovascular events (MACE). METHODS: We studied intake ECGs from 1448 COVID-19 patients (60.5% male, aged 63.4 ± 16.9 years). Records were labeled by mortality (death vs discharge) or MACE (no events vs arrhythmic, heart failure [HF], or thromboembolic [TE] events), then used to train AI models; these were compared to conventional regression models developed using demographic and comorbidity data. RESULTS: A total of 245 (17.7%) patients died (67.3% male, aged 74.5 ± 14.4 years); 352 (24.4%) experienced at least 1 MACE (119 arrhythmic, 107 HF, 130 TE). AI models predicted mortality and MACE with area under the curve (AUC) values of 0.60 ± 0.05 and 0.55 ± 0.07, respectively; these were comparable to AUC values for conventional models (0.73 ± 0.07 and 0.65 ± 0.10). There were no prominent temporal trends in mortality rate or MACE incidence in our cohort; holdout testing with data from after a cutoff date (June 9, 2020) did not degrade model performance. CONCLUSION: Using intake ECGs alone, our AI models had limited ability to predict hospitalized COVID-19 patients' risk of mortality or MACE. Our models' accuracy was comparable to that of conventional models built using more in-depth information, but translation to clinical use would require higher sensitivity and positive predictive value. In the future, we hope that mixed-input AI models utilizing both ECG and clinical data may be developed to enhance predictive accuracy.
ABSTRACT
AIMS: The degree of cardiovascular sequelae following COVID-19 remains unknown. The aim of this study was to investigate whether cardiac function recovers following COVID-19. METHODS AND RESULTS: A consecutive sample of patients hospitalized with COVID-19 was prospectively included in this longitudinal study. All patients underwent an echocardiographic examination during hospitalization and 2 months later. All participants were successfully matched 1:1 with COVID-19-free controls by age and sex. A total of 91 patients were included (mean age 63 ± 12 years, 59% male). A median of 77 days (interquartile range: 72-92) passed between the two examinations. Right ventricular (RV) function improved following resolution of COVID-19: tricuspid annular plane systolic excursion (TAPSE) (2.28 ± 0.40 cm vs. 2.11 ± 0.38 cm, P < 0.001) and RV longitudinal strain (RVLS) (25.3 ± 5.5% vs. 19.9 ± 5.8%, P < 0.001). In contrast, left ventricular (LV) systolic function assessed by global longitudinal strain (GLS) did not significantly improve (17.4 ± 2.9% vs. 17.6 ± 3.3%, P = 0.6). N-terminal pro-B-type natriuretic peptide decreased between the two examinations [177.6 (80.3-408.0) ng/L vs. 11.7 (5.7-24.0) ng/L, P < 0.001]. None of the participants had elevated troponins at follow-up compared to 18 (27.7%) during hospitalization. Recovered COVID-19 patients had significantly lower GLS (17.4 ± 2.9% vs. 18.8 ± 2.9%, P < 0.001 and adjusted P = 0.004), TAPSE (2.28 ± 0.40 cm vs. 2.67 ± 0.44 cm, P < 0.001 and adjusted P < 0.001), and RVLS (25.3 ± 5.5% vs. 26.6 ± 5.8%, P = 0.50 and adjusted P < 0.001) compared to matched controls. CONCLUSION: Acute COVID-19 affected negatively RV function and cardiac biomarkers but recovered following resolution of COVID-19. In contrast, the observed reduced LV function during acute COVID-19 did not improve post-COVID-19. Compared to the matched controls, both LV and RV function remained impaired.