BACKGROUND: Subtle, prognostically important ECG features may not be apparent to physicians. In the course of supervised machine learning, thousands of ECG features are identified. These are not limited to conventional ECG parameters and morphology. We aimed to investigate whether neural network-derived ECG features could be used to predict future cardiovascular disease and mortality and have phenotypic and genotypic associations. METHODS: We extracted 5120 neural network-derived ECG features from an artificial intelligence-enabled ECG model trained for 6 simple diagnoses and applied unsupervised machine learning to identify 3 phenogroups. Using the identified phenogroups, we externally validated our findings in 5 diverse cohorts from the United States, Brazil, and the United Kingdom. Data were collected between 2000 and 2023. RESULTS: In total, 1 808 584 patients were included in this study. In the derivation cohort, the 3 phenogroups had significantly different mortality profiles. After adjusting for known covariates, phenogroup B had a 20% increase in long-term mortality compared with phenogroup A (hazard ratio, 1.20 [95% CI, 1.17-1.23]; P<0.0001; phenogroup A mortality, 2.2%; phenogroup B mortality, 6.1%). In univariate analyses, we found phenogroup B had a significantly greater risk of mortality in all cohorts (log-rank P<0.01 in all 5 cohorts). Phenome-wide association study showed phenogroup B had a higher rate of future atrial fibrillation (odds ratio, 2.89; P<0.00001), ventricular tachycardia (odds ratio, 2.00; P<0.00001), ischemic heart disease (odds ratio, 1.44; P<0.00001), and cardiomyopathy (odds ratio, 2.04; P<0.00001). A single-trait genome-wide association study yielded 4 loci. SCN10A, SCN5A, and CAV1 have roles in cardiac conduction and arrhythmia. ARHGAP24 does not have a clear cardiac role and may be a novel target. CONCLUSIONS: Neural network-derived ECG features can be used to predict all-cause mortality and future cardiovascular diseases. We have identified biologically plausible and novel phenotypic and genotypic associations that describe mechanisms for the increased risk identified.

Department of Cardiology Chelsea and Westminster Hospital NHS Foundation Trust London United Kingdom

Department of Cardiology Imperial College Healthcare National Health Service Trust London United Kingdom

Department of Cardiology Royal Brompton and Harefield Hospitals Guy's and St Thomas' NHS Foundation Trust London United Kingdom

Department of Electrical and Electronic Engineering Imperial College London United Kingdom

Department of Infectious Diseases School of Medicine and Institute of Tropical Medicine University of São Paulo Brazil

Department of Information Technology Uppsala University Sweden

Department of Internal Medicine and Cardiology University Hospital Brno and Masaryk University Czech Republic

Department of Internal Medicine Faculdade de Medicina and Telehealth Center and Cardiology Service Hospital das Clínicas Universidade Federal de Minas Gerais Belo Horizonte Brazil

Department of Preventive Medicine School of Medicine and Hospital das Clínicas Empresa Brasileira de Serviços Hospitalares Universidade Federal de Minas Gerais Belo Horizonte Brazil

Faculty of Medicine and Health Sciences Center for Medical Genetics University of Antwerp and Antwerp University Hospital Antwerp Belgium

Harvard Thorndike Electrophysiology Institute Beth Israel Deaconess Medical Center Harvard Medical School Boston MA

Medical Research Council Laboratory of Medical Sciences Imperial College London United Kingdom

National Heart and Lung Institute Imperial College London United Kingdom

Research Department of Epidemiology and Public Health University College London United Kingdom

Richard A and Susan F Smith Center for Outcomes Research in Cardiology Beth Israel Deaconess Medical Center Harvard Medical School Boston MA

Komentář v

Circ Cardiovasc Qual Outcomes. 2024 Dec;17(12):e011483. doi: 10.1161/CIRCOUTCOMES.124.011483. PubMed

Zobrazit více v PubMed

Imanishi R, Seto S, Ichimaru S, Nakashima E, Yano K, Akahoshi M. Prognostic significance of incident complete left bundle branch block observed over a 40-year period. Am J Cardiol. 2006;98:644–648. doi: 10.1016/j.amjcard.2006.03.044 PubMed

Desai AD, Yaw TS, Yamazaki T, Kaykha A, Chun S, Froelicher VF. Prognostic significance of quantitative QRS duration. Am J Med. 2006;119:600–606. doi: 10.1016/j.amjmed.2005.08.028 PubMed

Verdecchia P, Angeli F, Reboldi G, Carluccio E, Benemio G, Gattobigio R, Borgioni C, Bentivoglio M, Porcellati C, Ambrosio G. Improved cardiovascular risk stratification by a simple ECG index in hypertension. Am J Hypertens. 2003;16:646–652. doi: 10.1016/s0895-7061(03)00912-9 PubMed

Raghunath S, Ulloa Cerna AE, Jing L, vanMaanen DP, Stough J, Hartzel DN, Leader JB, Kirchner HL, Stumpe MC, Hafez A, et al. . Prediction of mortality from 12-lead electrocardiogram voltage data using a deep neural network. Nat Med. 2020;26:886–891. doi: 10.1038/s41591-020-0870-z PubMed

Attia ZI, Kapa S, Lopez-Jimenez F, McKie PM, Ladewig DJ, Satam G, Pellikka PA, Enriquez-Sarano M, Noseworthy PA, Munger TM, et al. . Screening for cardiac contractile dysfunction using an artificial intelligence-enabled electrocardiogram. Nat Med. 2019;25:70–74. doi: 10.1038/s41591-018-0240-2 PubMed

Khurshid S, Friedman S, Reeder C, Di Achille P, Diamant N, Singh P, Harrington LX, Wang X, Al-Alusi MA, Sarma G, et al. . ECG-based deep learning and clinical risk factors to predict atrial fibrillation. Circulation. 2022;145:122–133. doi: 10.1161/CIRCULATIONAHA.121.057480 PubMed PMC

Zvuloni E, Read J, Ribeiro AH, Ribeiro ALP, Behar JA. On merging feature engineering and deep learning for diagnosis, risk-prediction and age estimation based on the 12-lead ECG. IEEE Trans Biomed Eng. 2023;70:2227–2236. doi: 10.1109/TBME.2023.3239527 PubMed

Ribeiro AH, Ribeiro MH, Paixao GMM, Oliveira DM, Gomes PR, Canazart JA, Ferreira MPS, Andersson CR, Macfarlane PW, Meira W, Jr, et al. . Automatic diagnosis of the 12-lead ECG using a deep neural network. Nat Commun. 2020;11:1760. doi: 10.1038/s41467-020-15432-4 PubMed PMC

Barish M, Bolourani S, Lau LF, Shah S, Zanos TP. External validation demonstrates limited clinical utility of the interpretable mortality prediction model for patients with COVID-19. Nat Mach Intell. 2021;3:25–27. doi: 10.1038/s42256-020-00254-2

Lima EM, Ribeiro AH, Paixão GMM, Ribeiro MH, Pinto-Filho MM, Gomes PR, Oliveira DM, Sabino EC, Duncan BB, Giatti L, et al. . Deep neural network-estimated electrocardiographic age as a mortality predictor. Nat Commun. 2021;12:5117. doi: 10.1038/s41467-021-25351-7 PubMed PMC

Marmot MG, Smith GD, Stansfeld S, Patel C, North F, Head J, White I, Brunner E, Feeney A. Health inequalities among British civil servants: the Whitehall II study. Lancet. 1991;337:1387–1393. doi: 10.1016/0140-6736(91)93068-k PubMed

Sudlow C, Gallacher J, Allen N, Beral V, Burton P, Danesh J, Downey P, Elliott P, Green J, Landray M, et al. . UK biobank: an open access resource for identifying the causes of a wide range of complex diseases of middle and old age. PLoS Med. 2015;12:e1001779. doi: 10.1371/journal.pmed.1001779 PubMed PMC

Schmidt MI, Duncan BB, Mill JG, Lotufo PA, Chor D, Barreto SM, Aquino EM, Passos VM, Matos SM, Molina MCB, et al. . Cohort profile: Longitudinal Study of Adult Health (ELSA-Brasil). Int J Epidemiol. 2015;44:68–75. doi: 10.1093/ije/dyu027 PubMed PMC

Cardoso CS, Sabino EC, Oliveira CDL, de Oliveira LC, Ferreira AM, Cunha-Neto E, Bierrenbach AL, Ferreira JE, Haikal DSA, Reingold CL, et al. . Longitudinal study of patients with chronic Chagas cardiomyopathy in Brazil (SaMi-Trop project): a cohort profile. BMJ Open. 2016;6:e011181. doi: 10.1136/bmjopen-2016-011181 PubMed PMC

Chollet F. Keras. 2015. https://keras.io

Abadi M, Agarwal A, Barham P, Brevdo E, Chen Z, Citro C, Corrado GS, Davis A, Dean J, Devin M, et al. . TensorFlow: large-scale machine learning on heterogeneous distributed systems. 2016. doi: 10.48550/arXiv.1603.04467

Syakur M, Khusnul Khotimah B, Rohman E. Integration K-means clustering method and elbow method for identification of the best customer profile cluster. IOP Conf Ser Mater Sci Eng. 2018;336:012017. doi: 10.1088/1757-899X/336/1/012017

Hnatkova K, Andršová I, Novotný T, Britton A, Shipley M, Vandenberk B, Sprenkeler DJ, Junttila J, Reichlin T, Schlögl S, et al. . QRS micro-fragmentation as a mortality predictor. Eur Heart J. 2022;43:4177–4191. doi: 10.1093/eurheartj/ehac085 PubMed PMC

Bai W, Suzuki H, Huang J, Francis C, Wang S, Tarroni G, Guitton F, Aung N, Fung K, Petersen SE, et al. . A population-based phenome-wide association study of cardiac and aortic structure and function. Nat Med. 2020;26:1654–1662. doi: 10.1038/s41591-020-1009-y PubMed PMC

Selvaraju RR, Cogswell M, Das A, Vedantam R, Parikh D, Batra D. Grad-CAM: visual explanations from deep networks via gradient-based localization. Int J Comput Vis. 2019;128:336–359. doi: 10.1007/s11263-019-01228-7

Cardoso CS, Ribeiro ALP, Oliveira CDL, Oliveira LC, Ferreira AM, Bierrenbach AL, Silva JLP, Colosimo EA, Ferreira JE, Lee TH, et al. . Beneficial effects of benznidazole in Chagas disease: NIH SaMi-Trop cohort study. PLoS NeglTrop Dis. 2018;12:e0006814. doi: 10.1371/journal.pntd.0006814 PubMed PMC

Chambers JC, Zhao J, Terracciano CM, Bezzina CR, Zhang W, Kaba R, Navaratnarajah M, Lotlikar A, Sehmi JS, Kooner MK, et al. . Genetic variation in SCN10A influences cardiac conduction. Nat Genet. 2010;42:149–152. doi: 10.1038/ng.516 PubMed

Holm H, Gudbjartsson DF, Arnar DO, Thorleifsson G, Thorgeirsson G, Stefansdottir H, Gudjonsson SA, Jonasdottir A, Mathiesen EB, Njølstad I, et al. . Several common variants modulate heart rate, PR interval and QRS duration. Nat Genet. 2010;42:117–122. doi: 10.1038/ng.511 PubMed

McNair WP, Ku L, Taylor MR, Fain PR, Dao D, Wolfel E, Mestroni L; Familial Cardiomyopathy Registry Research Group. SCN5A mutation associated with dilated cardiomyopathy, conduction disorder, and arrhythmia. Circulation. 2004;110:2163–2167. doi: 10.1161/01.CIR.0000144458.58660.BB PubMed

Yang KC, Rutledge CA, Mao M, Bakhshi FR, Xie A, Liu H, Bonini MG, Patel HH, Minshall RD, Dudley SC, Jr. Caveolin-1 modulates cardiac gap junction homeostasis and arrhythmogenecity by regulating cSrc tyrosine kinase. Circ Arrhythm Electrophysiol. 2014;7:701–710. doi: 10.1161/CIRCEP.113.001394 PubMed PMC

Méndez-Giráldez R, Gogarten SM, Below JE, Yao J, Seyerle AA, Highland HM, Kooperberg C, Soliman EZ, Rotter JI, Kerr KF, et al. . GWAS of the electrocardiographic QT interval in Hispanics/Latinos generalizes previously identified loci and identifies population-specific signals. Sci Rep. 2017;7:17075. doi: 10.1038/s41598-017-17136-0 PubMed PMC

Ntalla I, Weng LC, Cartwright JH, Hall AW, Sveinbjornsson G, Tucker NR, Choi SH, Chaffin MD, Roselli C, Barnes MR, et al. . Multi-ancestry GWAS of the electrocardiographic PR interval identifies 202 loci underlying cardiac conduction. Nat Commun. 2020;11:2542. doi: 10.1038/s41467-020-15706-x PubMed PMC

Stewart S, Playford D, Scalia GM, Currie P, Celermajer DS, Prior D, Codde J, Strange G; NEDA Investigators. Ejection fraction and mortality: a nationwide register-based cohort study of 499 153 women and men. Eur J Heart Fail. 2021;23:406–416. doi: 10.1002/ejhf.2047 PubMed

Surkova E, Muraru D, Genovese D, Aruta P, Palermo C, Badano LP. Relative prognostic importance of left and right ventricular ejection fraction in patients with cardiac diseases. J Am Soc Echocardiogr. 2019;32:1407–1415.e3. doi: 10.1016/j.echo.2019.06.009 PubMed

Kalam K, Otahal P, Marwick TH. Prognostic implications of global LV dysfunction: a systematic review and meta-analysis of global longitudinal strain and ejection fraction. Heart. 2014;100:1673–1680. doi: 10.1136/heartjnl-2014-305538 PubMed

Sillesen H, Muntendam P, Adourian A, Entrekin R, Garcia M, Falk E, Fuster V. Carotid plaque burden as a measure of subclinical atherosclerosis: comparison with other tests for subclinical arterial disease in the high risk plaque bioimage study. JACC Cardiovasc Imaging. 2012;5:681–689. doi: 10.1016/j.jcmg.2012.03.013 PubMed

Münzel T, Hahad O, Gori T, Hollmann S, Arnold N, Prochaska JH, Schulz A, Beutel M, Pfeiffer N, Schmidtmann I, et al. . Heart rate, mortality, and the relation with clinical and subclinical cardiovascular diseases: results from the Gutenberg Health Study. Clin Res Cardiol. 2019;108:1313–1323. doi: 10.1007/s00392-019-01466-2 PubMed PMC

Kalahasti V, Nambi V, Martin DO, Lam CT, Yamada D, Wilkoff BL, Niebauer MJ, Jaeger FJ, Tchou PJ, Chung MK. QRS duration and prediction of mortality in patients undergoing risk stratification for ventricular arrhythmias. Am J Cardiol. 2003;92:798–803. doi: 10.1016/s0002-9149(03)00886-5 PubMed

Zhang Y, Post WS, Blasco-Colmenares E, Dalal D, Tomaselli GF, Guallar E. Electrocardiographic QT interval and mortality: a meta-analysis. Epidemiology. 2011;22:660–670. doi: 10.1097/EDE.0b013e318225768b PubMed PMC

Aro AL, Anttonen O, Kerola T, Junttila MJ, Tikkanen JT, Rissanen HA, Reunanen A, Huikuri HV. Prognostic significance of prolonged PR interval in the general population. Eur Heart J. 2013;35:123–129. doi: 10.1093/eurheartj/eht176 PubMed

Das MK, Zipes DP. Fragmented QRS: a predictor of mortality and sudden cardiac death. Heart Rhythm. 2009;6:S8–14. doi: 10.1016/j.hrthm.2008.10.019 PubMed

Prenner SB, Shah SJ, Goldberger JJ, Sauer AJ. Repolarization heterogeneity: beyond the QT interval. J Am Heart Assoc. 2016;5:e003607. doi: 10.1161/JAHA.116.003607 PubMed PMC

Hu D, Barajas-Martínez H, Pfeiffer R, Dezi F, Pfeiffer J, Buch T, Betzenhauser Matthew J, Belardinelli L, Kahlig Kristopher M, Rajamani S, et al. . Mutations in SCN10A are responsible for a large fraction of cases of Brugada syndrome. J Am Coll Cardiol. 2014;64:66–79. doi: 10.1016/j.jacc.2014.04.032 PubMed PMC

Rook MB, Bezzina Alshinawi C, Groenewegen WA, van Gelder IC, van Ginneken ACG, Jongsma HJ, Mannens MMAM, Wilde AAM. Human SCN5A gene mutations alter cardiac sodium channel kinetics and are associated with the Brugada syndrome. Cardiovasc Res. 1999;44:507–517. doi: 10.1016/s0008-6363(99)00350-8 PubMed

Hesse M, Kondo CS, Clark RB, Su L, Allen FL, Geary-Joo CT, Kunnathu S, Severson DL, Nygren A, Giles WR, et al. . Dilated cardiomyopathy is associated with reduced expression of the cardiac sodium channel Scn5a. Cardiovasc Res. 2007;75:498–509. doi: 10.1016/j.cardiores.2007.04.009 PubMed

Marcsa B, Dénes R, Vörös K, Rácz G, Sasvári-Székely M, Rónai Z, Törő K, Keszler G. A common polymorphism of the human cardiac sodium channel alpha subunit (SCN5A) gene is associated with sudden cardiac death in chronic ischemic heart disease. PLoS One. 2015;10:e0132137. doi: 10.1371/journal.pone.0132137 PubMed PMC

Engelman JA, Zhang XL, Galbiati F, Lisanti MP. Chromosomal localization, genomic organization, and developmental expression of the murine caveolin gene family (Cav-1, -2, and -3). Cav-1 and Cav-2 genes map to a known tumor suppressor locus (6-A2/7q31). FEBS Lett. 1998;429:330–336. doi: 10.1016/s0014-5793(98)00619-x PubMed

Ellinor PT, Lunetta KL, Albert CM, Glazer NL, Ritchie MD, Smith AV, Arking DE, Müller-Nurasyid M, Krijthe BP, Lubitz SA, et al. . Meta-analysis identifies six new susceptibility loci for atrial fibrillation. Nat Genet. 2012;44:670–675. doi: 10.1038/ng.2261 PubMed PMC

Su ZJ, Hahn CN, Goodall GJ, Reck NM, Leske AF, Davy A, Kremmidiotis G, Vadas MA, Gamble JR. A vascular cell-restricted RhoGAP, p73RhoGAP, is a key regulator of angiogenesis. Proc Natl Acad Sci USA. 2004;101:12212–12217. doi: 10.1073/pnas.0404631101 PubMed PMC

Young WJ, Lahrouchi N, Isaacs A, Duong T, Foco L, Ahmed F, Brody JA, Salman R, Noordam R, Benjamins JW, et al. . Genetic analyses of the electrocardiographic QT interval and its components identify additional loci and pathways. Nat Commun. 2022;13:5144. doi: 10.1038/s41467-022-32821-z PubMed PMC

Movassagh M, Choy MK, Goddard M, Bennett MR, Down TA, Foo RS. Differential DNA methylation correlates with differential expression of angiogenic factors in human heart failure. PLoS One. 2010;5:e8564. doi: 10.1371/journal.pone.0008564 PubMed PMC

Sun W, Kalmady SV, Sepehrvand N, Salimi A, Nademi Y, Bainey K, Ezekowitz JA, Greiner R, Hindle A, McAlister FA, et al. . Towards artificial intelligence-based learning health system for population-level mortality prediction using electrocardiograms. npj Digit Med. 2023;6:21. doi: 10.1038/s41746-023-00765-3 PubMed PMC

Fry A, Littlejohns TJ, Sudlow C, Doherty N, Adamska L, Sprosen T, Collins R, Allen NE. Comparison of sociodemographic and health-related characteristics of UK biobank participants with those of the general population. Am J Epidemiol. 2017;186:1026–1034. doi: 10.1093/aje/kwx246 PubMed PMC

Ardissino M, McCracken C, Bard A, Antoniades C, Neubauer S, Harvey NC, Petersen SE, Raisi-Estabragh Z. Pericardial adiposity is independently linked to adverse cardiovascular phenotypes: a CMR study of 42 598 UK Biobank participants. Eur Heart J Cardiovasc Imaging. 2022;23:1471–1481. doi: 10.1093/ehjci/jeac101 PubMed PMC

Hughes JW, Tooley J, Torres Soto J, Ostropolets A, Poterucha T, Christensen MK, Yuan N, Ehlert B, Kaur D, Kang G, et al. . A deep learning-based electrocardiogram risk score for long term cardiovascular death and disease. npj Digit Med. 2023;6:169. doi: 10.1038/s41746-023-00916-6 PubMed PMC

Suchard MA, Schuemie MJ, Krumholz HM, You SC, Chen R, Pratt N, Reich CG, Duke J, Madigan D, Hripcsak G, et al. . Comprehensive comparative effectiveness and safety of first-line antihypertensive drug classes: a systematic, multinational, large-scale analysis. Lancet. 2019;394:1816–1826. doi: 10.1016/S0140-6736(19)32317-7 PubMed PMC

Pedregosa F, Varoquaux G, Gramfort A, Michel V, Thirion B, Grisel O, Blondel M, Prettenhofer P, Weiss R, Dubourg V. Scikit-learn: machine learning in python. J Mach Learn Res. 2011;12:2825–2830. doi: 10.6009/jjrt.2023-2266

Stensrud MJ, Hernan MA. Why test for proportional hazards? JAMA. 2020;323:1401–1402. doi: 10.1001/jama.2020.1267 PubMed

Wu P, Gifford A, Meng X, Li X, Campbell H, Varley T, Zhao J, Carroll R, Bastarache L, Denny JC, et al. . Mapping ICD-10 and ICD-10-CM codes to phecodes: workflow development and initial evaluation. JMIR Med Inform. 2019;7:e14325. doi: 10.2196/14325 PubMed PMC

Meyer HV, Dawes TJW, Serrani M, Bai W, Tokarczuk P, Cai J, de Marvao A, Henry A, Lumbers RT, Gierten J, et al. . Genetic and functional insights into the fractal structure of the heart. Nature. 2020;584:589–594. doi: 10.1038/s41586-020-2635-8 PubMed PMC

Yang J, Lee SH, Goddard ME, Visscher PM. GCTA: a tool for genome-wide complex trait analysis. Am J Hum Genet. 2011;88:76–82. doi: 10.1016/j.ajhg.2010.11.011 PubMed PMC