Dataset Information

Evaluation of Risk Prediction Models of Atrial Fibrillation (from the Multi-Ethnic Study of Atherosclerosis [MESA]).

ABSTRACT: Atrial fibrillation (AF) is prevalent and strongly associated with higher cardiovascular disease (CVD) risk. Machine learning is increasingly used to identify novel predictors of CVD risk, but prediction improvements beyond established risk scores are uncertain. We evaluated improvements in predicting 5-year AF risk when adding novel candidate variables identified by machine learning to the CHARGE-AF Enriched score, which includes age, race/ethnicity, height, weight, systolic and diastolic blood pressure, current smoking, use of antihypertensive medication, diabetes, and NT-proBNP. We included 3,534 participants (mean age, 61.3 years; 52.0% female) with complete data from the prospective Multi-Ethnic Study of Atherosclerosis. Incident AF was defined based on study electrocardiograms and hospital discharge diagnosis ICD-9 codes, supplemented by Medicare claims. Prediction performance was evaluated using Cox regression and a parsimonious model was selected using LASSO. Within 5 years of baseline, 124 participants had incident AF. Compared with the CHARGE-AF Enriched model (c-statistic, 0.804), variables identified by machine learning, including biomarkers, cardiac magnetic resonance imaging variables, electrocardiogram variables, and subclinical CVD variables, did not significantly improve prediction. A 23-item score derived by machine learning achieved a c-statistic of 0.806, whereas a parsimonious model including the clinical risk factors age, weight, current smoking, NT-proBNP, coronary artery calcium score, and cardiac troponin-T achieved a c-statistic of 0.802. This analysis confirms that the CHARGE-AF Enriched model and a parsimonious 6-item model performed similarly to a more extensive model derived by machine learning. In conclusion, these simple models remain the gold standard for risk prediction of AF, although addition of the coronary artery calcium score should be considered.

SUBMITTER: Bundy JD

PROVIDER: S-EPMC6911821 | biostudies-literature | 2020 Jan

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Evaluation of Risk Prediction Models of Atrial Fibrillation (from the Multi-Ethnic Study of Atherosclerosis [MESA]).

Bundy Joshua D JD Heckbert Susan R SR Chen Lin Y LY Lloyd-Jones Donald M DM Greenland Philip P

The American journal of cardiology 20191010 1

Atrial fibrillation (AF) is prevalent and strongly associated with higher cardiovascular disease (CVD) risk. Machine learning is increasingly used to identify novel predictors of CVD risk, but prediction improvements beyond established risk scores are uncertain. We evaluated improvements in predicting 5-year AF risk when adding novel candidate variables identified by machine learning to the CHARGE-AF Enriched score, which includes age, race/ethnicity, height, weight, systolic and diastolic blood ...[more]

PMID: 31706453

Similar Datasets

Project description:Background Left atrial (LA) and left ventricular (LV) structural and functional parameters have independent prognostic values as predictors of atrial fibrillation (AF). Purpose To investigate the prognostic value of a left atrioventricular coupling index (LACI) and average annualized change in LACI (hereafter, ΔLACI) measured by cardiac MRI to predict incident AF in a population-based sample from the Multi-Ethnic Study of Atherosclerosis (MESA). Materials and Methods In a secondary analysis of the prospective MESA, 1911 study participants without clinically recognized AF and cardiovascular disease at baseline had LACI assessed with cardiac MRI at baseline (examination 1, 2000-2002) and 10 years later (examination 5, 2010-2012). LACI was defined as the ratio of LA to LV end-diastolic volumes. Univariable and multivariable Cox proportional hazard models were used to evaluate the associations of LACI and average ΔLACI with incident AF. Results Among the 1911 participants (mean age, 59 years ± 9 [standard deviation]; 907 men), 87 incident AF events occurred over 3.9 years ± 0.9 after the second imaging (examination 5). After adjustment for traditional risk factors, greater LACI and ΔLACI were independently associated with AF (hazard ratio, 1.69 [95% CI: 1.46, 1.96] and 1.71 [95% CI: 1.50, 1.94], respectively; both P < .001). Adjusted models for LACI and ΔLACI showed improvement in model discrimination compared with currently used AF risk score (Cohort for Heart and Aging Research in Genomic Epidemiology-Atrial Fibrillation, or CHARGE-AF, score) model (area under receiver operating characteristic curve [AUC], 0.78 vs 0.74; and AUC, 0.80 vs 0.74, respectively; both P < .001); and to the final model including individual LA or LV parameters for predicting AF incidence (AUC, 0.78 vs 0.76; and AUC, 0.80 vs 0.78, respectively; both P < .001). Conclusion Atrioventricular coupling (left atrioventricular coupling index [LACI]) and coupling change (annual change in LACI) were strong predictors for atrial fibrillation (AF) in a multiethnic population. Both had incremental prognostic value for predicting AF over traditional risk factors, and superior discrimination compared with the Cohort for Heart and Aging Research in Genomic Epidemiology-Atrial Fibrillation, or CHARGE-AF, score and to individual left atrial or left ventricular parameters. © RSNA, 2022 Online supplemental material is available for this article. See also the editorial by Leiner in this issue.

Project description:BACKGROUND:While studies of the left atrium (LA) have demonstrated associations between volumes and emptying fraction with atrial fibrillation (AF), the contribution of right atrial (RA) abnormalities to incident AF remains poorly understood. OBJECTIVES:Assess the association between RA structure and function with incident AF using feature-tracking cardiovascular magnetic resonance (CMR). METHODS:This is a prospective cohort study of all participants in the Multi-Ethnic Study of Atherosclerosis with baseline CMR, sinus rhythm, and free of clinical cardiovascular disease at study initiation. RA volume, strain, and emptying fraction in participants with incident AF (n = 368) were compared against AF-free (n = 2779). Cox proportional-hazards models assessed association between variables. RESULTS:Participants were aged 60 ± 10 yrs., 55% female, and followed an average 11.2 years. Individuals developing AF had higher baseline RA maximum volume index (mean ± standard deviation [SD]: 24 ± 9 vs 22 ± 8 mL/m2, p = 0.002) and minimum volume index (13 ± 7 vs 12 ± 6 mL/m2, p < 0.001), and lower baseline RA emptying fraction (45 ± 15% vs 47 ± 15%, p = 0.02), peak global strain (34 ± 17% vs 36 ± 19%, p < 0.001), and peak free-wall strain (40 ± 23% vs 42 ± 26%, p = 0.049) compared with the AF-free population. After adjusting for traditional cardiovascular risk factors and LA volume and function, we found RA maximum volume index (hazards ratio [HR]: 1.13 per SD, p = 0.041) and minimum volume index (HR: 1.12 per SD, p = 0.037) were independently associated with incident AF. CONCLUSIONS:In a large multiethnic population, higher RA volume indices were independently associated with incident AF after adjustment for conventional cardiovascular risk factors and LA parameters. It is unclear if this predictive value persists when additional adjustment is made for ventricular parameters.

Dataset Information

Evaluation of Risk Prediction Models of Atrial Fibrillation (from the Multi-Ethnic Study of Atherosclerosis [MESA]).

Publications

Evaluation of Risk Prediction Models of Atrial Fibrillation (from the Multi-Ethnic Study of Atherosclerosis [MESA]).

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets