Dataset Information

Neural networks applied to 12-lead electrocardiograms predict body mass index, visceral adiposity and concurrent cardiometabolic ill-health.

ABSTRACT:

Background

Obesity is associated with electrophysiological remodeling, which manifests as detectable changes on the surface electrocardiogram (ECG).

Objective

To develop neural networks (NN) to predict body mass index (BMI) from ECGs and test the hypothesis that discrepancies between NN-predicted BMI and measured BMI are indicative of underlying adiposity and/or concurrent cardiometabolic ill-health.

Methods

NN models were developed using 36,856 12-lead resting ECGs from the UK Biobank. Two architectures were developed for continuous and categorical BMI estimation (normal weight [BMI <25 kg/m²] vs overweight/obese [BMI ≥25 kg/m²]). Models for male and female participants were trained and tested separately. For each sex, data were randomly divided into 4 folds, and models were evaluated in a leave-1-fold-out manner.

Results

ECGs were available for 17,807 male and 19,049 female participants (mean ages: 61 ± 7 and 63 ± 8 years; mean BMI 26 ± 5 kg/m² and 27 ± 4 kg/m², respectively). NN models detected overweight/obese individuals with average accuracies of 75% and 73% for male and female subjects, respectively. The magnitudes of difference between NN-predicted BMI and actual BMI were significantly correlated with visceral adipose tissue volumes. Concurrent hypertension, diabetes, dyslipidemia, and/or coronary heart disease explained false-positive classifications (ie, calculated BMI <25 kg/m² misclassified as ≥25 kg/m² by NN model, P < .001).

Conclusion

NN models applied to 12-lead ECGs predict BMI with a reasonable degree of accuracy. Discrepancies between NN-predicted and calculated BMI may be indicative of underlying visceral adiposity and concomitant cardiometabolic perturbation, which could be used to identify individuals at risk of cardiometabolic disease.

SUBMITTER: Li X

PROVIDER: S-EPMC8669785 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:OBJECTIVES:The visceral adiposity index (VAI), an indirect marker of visceral adipose tissue, serves as a model associated with cardiometabolic risk, but has limitations regarding the Asian population. We sought to develop a new VAI (NVAI) for the Korean population and compare it to VAI for prediction of atherosclerotic cardiovascular disease (ASCVD) risk and development of major cardiovascular diseases (CVD) and stroke. METHODS:Patients (969) who underwent visceral fat area measurement were analyzed. After exclusion, 539 patients (142 men, 397 women) were randomly divided into internal (n = 374) and external validation (n = 165) data set. The NVAI was developed using univariate and multivariate logistic regression with backward selection of predictors. Receiver operating characteristic (ROC) curve analysis and comparison of the area under the curve (AUC) verified the better predictor of ASCVD risk score. Additionally, nationwide population-based cross-sectional survey data (Korean National Health and Nutrition Examination Survey [KNHANES] 2008-2010, n = 29,235) was used to validate the NVAI's ability to predict ASCVD risk and major CVD and stroke. RESULTS:The NVAI better reflected visceral fat area in internal and external data sets, with AUCs of 0.911 (95% confidence interval [CI]: 0.882-0.940) and 0.879 (95% CI: 0.828-0.931), respectively. NVAI better discriminated for ASCVD risk (AUC = 0.892, 95% CI: 0.846-0.938) compared to VAI (0.559, 95% CI: 0.439-0.679). The NVAI also better predicted MI or angina, and stroke with AUCs of 0.771 (95% CI: 0.752-0.789), and 0.812 (95% CI: 0.794-0.830), respectively, compared with waist circumference (WC), body mass index (BMI), TG to HDL ratio, and VAI via KNHANES, in a statistically significant manner. CONCLUSIONS:The NVAI has advantages as a predictor of visceral obesity and is significantly associated with ASCVD risks and development of major CVD and stroke in the Korean population. The NVAI could be a screening tool for improved risk estimation related to visceral obesity.

Project description:Abstract Objectives The Visceral Adiposity Index (VAI) was developed to estimate adipose tissue dysfunction as well as visceral fat distribution. In an ancillary analysis of the CROSSROADS Study (clinicaltrials.gov #NCT00955903), this study investigated the effects of exercise with and without intentional energy restriction on VAI changes among older adults with obesity. Methods Participants (n = 163, 37.4% male, 70.3 ± 4.7 years) were randomized to the following groups: exercise only, exercise + nutrient-dense weight maintenance diet, and exercise + nutrient-dense energy restriction of 500 kcal/d. Visceral adiposity determined by magnetic resonance imaging (MRI) and dual-energy X-ray absorptiometry (DXA), anthropometrics, and cardiometabolic biomarkers were assessed at baseline and 12 months. Waist circumference, body mass index, triglycerides, and high-density lipoprotein (HDL) cholesterol were used to compute VAI at both time points. Data were analyzed utilizing Spearman's correlations and general linear models to determine differences among groups following adjustment for covariates. Results Among all participants, significant correlations were observed between VAI and glucose, insulin, tumor necrosis factor-alpha (TNF-α), and the adiponectin: leptin ratio at baseline (r = 0.343, P < 0.001; r = 0.363, P < 0.001; r = 0.200, P = 0.011; r = −0.246, P = 0.002, respectively). DXA and MRI measures of visceral adiposity were not significantly correlated with VAI. Upon completion of the 12-month study period, there were significant differences in VAI among groups [F (2,137) = 5.957, P = 0.003)]. Post hoc analysis revealed that VAI was significantly lower in the exercise + nutrient-dense energy restriction group compared to the exercise only group (P = 0.002). Conclusions Results suggest that VAI may be an adequate measure of adipose tissue dysfunction, yet its ability to estimate levels of visceral adiposity warrants further investigation to determine clinical utility. Nonetheless, results suggest that an exercise + nutrient-dense energy restriction intervention may reduce VAI among older adults with obesity. Funding Sources NIH.