Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Artificial intelligence–enabled electrocardiograms for identification of patients with low ejection fraction: a pragmatic, randomized clinical trial

Abstract

We have conducted a pragmatic clinical trial aimed to assess whether an electrocardiogram (ECG)-based, artificial intelligence (AI)-powered clinical decision support tool enables early diagnosis of low ejection fraction (EF), a condition that is underdiagnosed but treatable. In this trial (NCT04000087), 120 primary care teams from 45 clinics or hospitals were cluster-randomized to either the intervention arm (access to AI results; 181 clinicians) or the control arm (usual care; 177 clinicians). ECGs were obtained as part of routine care from a total of 22,641 adults (N = 11,573 intervention; N = 11,068 control) without prior heart failure. The primary outcome was a new diagnosis of low EF (≤50%) within 90 days of the ECG. The trial met the prespecified primary endpoint, demonstrating that the intervention increased the diagnosis of low EF in the overall cohort (1.6% in the control arm versus 2.1% in the intervention arm, odds ratio (OR) 1.32 (1.01–1.61), P = 0.007) and among those who were identified as having a high likelihood of low EF (that is, positive AI-ECG, 6% of the overall cohort) (14.5% in the control arm versus 19.5% in the intervention arm, OR 1.43 (1.08–1.91), P = 0.01). In the overall cohort, echocardiogram utilization was similar between the two arms (18.2% control versus 19.2% intervention, P = 0.17); for patients with positive AI-ECGs, more echocardiograms were obtained in the intervention compared to the control arm (38.1% control versus 49.6% intervention, P < 0.001). These results indicate that use of an AI algorithm based on ECGs can enable the early diagnosis of low EF in patients in the setting of routine primary care.

This is a preview of subscription content, access via your institution

Access options

Rent or buy this article

Prices vary by article type

from$1.95

to$39.95

Prices may be subject to local taxes which are calculated during checkout

Fig. 1
Fig. 2: Subgroup analyses for primary endpoint.

Similar content being viewed by others

Data availability

The data are not publicly available because they are electronic health records, consented for research use by Mayo Clinic investigators. Making the data publicly available without additional consent or ethical approval might compromise patients’ privacy and the original ethical approval. If other investigators are interested in performing additional analyses, requests can be made to the corresponding author (X.Y.) and analyses will be performed in collaboration with the Mayo Clinic.

Code availability

The AI algorithm, which was previously published, cannot be made publicly available because it is proprietary intellectual property (patent pending). The AI algorithm cannot be used in routine practice before getting FDA approval, and this algorithm is currently undergoing a submission/review process with the FDA. The AI algorithm is available upon request for research studies.

References

  1. Wang, T. J. et al. Natural history of asymptomatic left ventricular systolic dysfunction in the community. Circulation 108, 977–982 (2003).

    Article  Google Scholar 

  2. McDonagh, T. A. et al. Symptomatic and asymptomatic left-ventricular systolic dysfunction in an urban population. Lancet 350, 829–833 (1997).

    Article  CAS  Google Scholar 

  3. Redfield, M. M. et al. Burden of systolic and diastolic ventricular dysfunction in the community: appreciating the scope of the heart failure epidemic. JAMA 289, 194–202 (2003).

    Article  Google Scholar 

  4. McDonagh, T. A., McDonald, K. & Maisel, A. S. Screening for asymptomatic left ventricular dysfunction using B-type natriuretic peptide. Congest. Heart Fail. 14, 5–8 (2008).

    Article  CAS  Google Scholar 

  5. Morgan, S. et al. Prevalence and clinical characteristics of left ventricular dysfunction among elderly patients in general practice setting: cross sectional survey. Br. Med. J. 318, 368–372 (1999).

    Article  CAS  Google Scholar 

  6. Jong, P., Yusuf, S., Rousseau, M. F., Ahn, S. A. & Bangdiwala, S. I. Effect of enalapril on 12-year survival and life expectancy in patients with left ventricular systolic dysfunction: a follow-up study. Lancet 361, 1843–1848 (2003).

    Article  CAS  Google Scholar 

  7. Yusuf, S., Pitt, B., Davis, C. E., Hood, W. B. Jr & Cohn, J. N. Effect of enalapril on mortality and the development of heart failure in asymptomatic patients with reduced left ventricular ejection fractions. New Engl. J. Med. 327, 685–691 (1992).

    Article  CAS  Google Scholar 

  8. Yancy, C. W. et al. 2013 ACCF/AHA guideline for the management of heart failure: a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines. J. Am. Coll. Cardiol. 62, e147–e239 (2013).

    Article  Google Scholar 

  9. Wang, T. J., Levy, D., Benjamin, E. J. & Vasan, R. S. The epidemiology of ‘asymptomatic’ left ventricular systolic dysfunction: implications for screening. Ann. Intern. Med. 138, 907–916 (2003).

    Article  Google Scholar 

  10. Hetmanski, D. J., Sparrow, N. J., Curtis, S. & Cowley, A. J. Failure of plasma brain natriuretic peptide to identify left ventricular systolic dysfunction in the community. Heart 84, 440–441 (2000).

    Article  CAS  Google Scholar 

  11. Landray, M. J., Lehman, R. & Arnold, I. Measuring brain natriuretic peptide in suspected left ventricular systolic dysfunction in general practice: cross-sectional study. Br. Med. J. 320, 985–986 (2000).

    Article  CAS  Google Scholar 

  12. Goetze, J. P. et al. Plasma pro-B-type natriuretic peptide in the general population: screening for left ventricular hypertrophy and systolic dysfunction. Eur. Heart J. 27, 3004–3010 (2006).

    Article  CAS  Google Scholar 

  13. Redfield, M. M. et al. Plasma brain natriuretic peptide to detect preclinical ventricular systolic or diastolic dysfunction: a community-based study. Circulation 109, 3176–3181 (2004).

    Article  CAS  Google Scholar 

  14. Nielsen, O. W., Hansen, J. F., Hilden, J., Larsen, C. T. & Svanegaard, J. Risk assessment of left ventricular systolic dysfunction in primary care: cross sectional study evaluating a range of diagnostic tests. Br. Med. J. 320, 220–224 (2000).

    Article  CAS  Google Scholar 

  15. Rihal, C. S., Davis, K. B., Kennedy, J. W. & Gersh, B. J. The utility of clinical, electrocardiographic, and roentgenographic variables in the prediction of left ventricular function. Am. J. Cardiol. 75, 220–223 (1995).

    Article  CAS  Google Scholar 

  16. Davie, A. P. et al. Value of the electrocardiogram in identifying heart failure due to left ventricular systolic dysfunction. Br. Med. J. 312, 222 (1996).

    Article  CAS  Google Scholar 

  17. Attia, Z. I. et al. Screening for cardiac contractile dysfunction using an artificial intelligence-enabled electrocardiogram. Nat. Med. 25, 70–74 (2019).

    Article  CAS  Google Scholar 

  18. Attia, Z. I. et al. Prospective validation of a deep learning ECG algorithm for the detection of left ventricular systolic dysfunction. J. Cardiovasc. Electrophysiol. 30, 668–674 (2019).

    Article  Google Scholar 

  19. Kaggal, V. C. et al. Toward a learning health-care system—knowledge delivery at the point of care empowered by big data and NLP. Biomed. Inf. Insights 8, 13–22 (2016).

    Google Scholar 

  20. Simon, G. E., Platt, R. & Hernandez, A. F. Evidence from pragmatic trials during routine care—slouching toward a learning health system. New Engl. J. Med. 382, 1488–1491 (2020).

    Article  Google Scholar 

  21. Wen, A. et al. Desiderata for delivering NLP to accelerate healthcare AI advancement and a Mayo Clinic NLP-as-a-service implementation. NPJ Digit. Med. 2, 130 (2019).

    Article  Google Scholar 

  22. Yao, X. et al. ECG AI-Guided Screening for Low Ejection Fraction (EAGLE): rationale and design of a pragmatic cluster randomized trial. Am. Heart J. 219, 31–36 (2019).

    Article  Google Scholar 

  23. Yao, X. et al. Clinical trial design data for electrocardiogram artificial intelligence-guided screening for low ejection fraction (EAGLE). Data Brief 28, 104894 (2019).

    Article  Google Scholar 

  24. Liu, X. et al. Reporting guidelines for clinical trial reports for interventions involving artificial intelligence: the CONSORT-AI extension. Nat. Med. 26, 1364–1374 (2020).

    Article  CAS  Google Scholar 

Download references

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Xiaoxi Yao.

Ethics declarations

Competing interests

Mayo Clinic has licensed the AI-ECG algorithm to EKO, a maker of digital stethoscopes with embedded ECG electrodes. At no point will Mayo Clinic benefit financially from the use of the AI-ECG for the care of patients at the Mayo Clinic. P.A.F., F.L.-J., S.K. and Z.I.A. may receive financial benefit from this agreement for use of the AI-ECG outside of the Mayo Clinic. All other authors declare no competing interests.

Additional information

Peer review information Nature Medicine thanks Jill Waalen, Giorgio Quer and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Michael Basson was the primary editor on this article and managed its editorial process and peer review in collaboration with the rest of the editorial team.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Extended data

Extended Data Fig. 1 Subgroup analyses for echocardiogram performed.

Outcome was echocardiogram being performed. An odds ratio greater than 1 means a higher likelihood of echocardiogram being performed. Error bars indicate 95% confidence intervals; mixed effect logistic regressions were used for the statistical test; it was two-sided; no adjustment for multiple comparison was made.

Extended Data Fig. 2 Subgroup analyses for echocardiogram performed among patients with positive screening results.

Outcome was echocardiogram being performed. An odds ratio greater than 1 means a higher likelihood of echocardiogram being performed. Error bars indicate 95% confidence intervals; mixed effect logistic regressions were used for the statistical test; it was two-sided; no adjustment for multiple comparison was made.

Extended Data Fig. 3

Clinician-facing action recommendation report for screening results.

Extended Data Fig. 4

Internal resources website for clinicians.

Supplementary information

Supplementary Information

Supplementary Tables 1–10.

Reporting Summary

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Yao, X., Rushlow, D.R., Inselman, J.W. et al. Artificial intelligence–enabled electrocardiograms for identification of patients with low ejection fraction: a pragmatic, randomized clinical trial. Nat Med 27, 815–819 (2021). https://doi.org/10.1038/s41591-021-01335-4

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1038/s41591-021-01335-4

This article is cited by

Search

Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing