ABSTRACT
OBJECTIVE: To rapidly exclude severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection using artificial intelligence applied to the electrocardiogram (ECG). METHODS: A global, volunteer consortium from 4 continents identified patients with ECGs obtained around the time of polymerase chain reaction-confirmed COVID-19 diagnosis and age- and sex-matched controls from the same sites. Clinical characteristics, polymerase chain reaction results, and raw electrocardiographic data were collected. A convolutional neural network was trained using 26,153 ECGs (33.2% COVID positive), validated with 3826 ECGs (33.3% positive), and tested on 7870 ECGs not included in other sets (32.7% positive). Performance under different prevalence values was tested by adding control ECGs from a single high-volume site. RESULTS: The area under the curve for detection of acute COVID-19 infection in the test group was 0.767 (95% CI, 0.756 to 0.778; sensitivity, 98%; specificity, 10%; positive predictive value, 37%; negative predictive value, 91%). To more accurately reflect a real-world population, 50,905 normal controls were added to adjust the COVID prevalence to approximately 5% (2657/58,555), resulting in an area under the curve of 0.780 (95% CI, 0.771 to 0.790) with a specificity of 12.1% and a negative predictive value of 99.2%. CONCLUSION: Infection with SARS-CoV-2 results in electrocardiographic changes that permit the artificial intelligence-enhanced ECG to be used as a rapid screening test with a high negative predictive value (99.2%). This may permit the development of electrocardiography-based tools to rapidly screen individuals for pandemic control.
Subject(s)
Artificial Intelligence , COVID-19/diagnosis , Electrocardiography , Case-Control Studies , Humans , Predictive Value of Tests , Sensitivity and SpecificityABSTRACT
BACKGROUND: High sensitivity cardiac troponin-T (hs-TnT) has been associated with mortality in patients hospitalized with COVID-19. We aimed to determine if hs-TnT levels and their timing are independent predictors of adverse events in these patients. DESIGN: Retrospective chart review was performed for all patients hospitalized at our institution between 23 March 2020 and 13 April 2020 who were found to be COVID-19-positive. Clinical, demographic, and laboratory variables including initial and peak hs-TnT were recorded. Univariable and multivariable analyses were completed for a primary composite endpoint of in-hospital death, intubation, need for critical care, or cardiac arrest. RESULTS: In the 276 patients analysed, initial hs-TnT above the median (≥17 ng/L) was associated with increased length of stay, need for vasoactive medications, and death, along with the composite endpoint (OR 3.92, p < 0.001). Multivariable analysis demonstrated that elevated initial hs-TnT was independently associated with the primary endpoint (OR 2.92, p = 0.01). Late-peaking hs-TnT (OR 2.19 for each additional day until peak, p < 0.001) was also independently associated with the composite endpoint. CONCLUSIONS: In patients hospitalized with COVID-19, hs-TnT identifies patients at high risk for adverse in-hospital events, and trends of hs-TnT over time, particularly during the first day, provide additional prognostic information.