Machine Learning for Sudden Cardiac Death Prediction in the Atherosclerosis Risk in Communities Study

Introduction Sudden cardiac death (SCD) is a devastating consequence often without antecedent expectation. Current risk stratification methods derived from baseline independently modeled risk factors are insufficient. Novel random forest machine learning (ML) approach incorporating time-dependent variables and complex interactions may improve SCD risk prediction. Methods Atherosclerosis Risk in Communities (ARIC) study participants were followed for adjudicated SCD. ML models were compared to standard Poisson regression models for interval data, an approximation to Cox regression, with stepwise variable selection. Eighty-two time-varying variables (demographics, lifestyle factors, clinical characteristics, biomarkers, etc.) collected at four visits over 12 years (1987-98) were used as candidate predictors. Predictive accuracy was assessed by area under the receiver operating characteristic curve (AUC) through out-of-bag prediction for ML models and 5-fold cross validation for the Poisson regression models. Results Over a median follow-up time of 23.5 years, 583 SCD events occurred among 15,661 ARIC participants (mean age 54 years and 55% women). Compared to different Poisson regression models (AUC at 6-year ranges from 0.77-0.83), the ML model improved prediction (AUC at 6-year 0.89). Top predictors identified by ML model included prior coronary heart disease (CHD), which explained 47.9% of the total phenotypic variance, diabetes mellitus, hypertension, and T wave abnormality in any of leads I, aVL, or V6. Using the top ML predictors to select variables, the Poisson regression model AUC at 6-year was 0.77 suggesting that the non-linear dependencies and interactions captured by ML, are the main reasons for its improved prediction performance. Conclusions Applying novel ML approach with time-varying predictors improves the prediction of SCD. Interactions of dynamic clinical characteristics are important for risk-stratifying SCD in the general population.

Associations Between Atrial Arrhythmias and Brain Amyloid Deposition: The ARIC-PET Study

The aim of this study is to determine if there is an association between atrial arrhythmias and brain amyloid-β (Aβ), measured on florbetapir (FBP) PET. 346 nondemented participants from the Atherosclerosis Risk in Communities study underwent FBP-PET, 185 also wore Zio® XT Patch. The associations between global cortical Aβ (> 1.2 standardized uptake value ratio) and history of atrial fibrillation, zio-defined atrial tachycardia and premature atrial contractions, each, were evaluated. Among nondemented community-dwelling older adults, we did not find an association between atrial arrhythmias and Aβ. Other brain pathology may underlie the association described between atrial arrhythmias and cognition.

Associations between estimated and measured carotid-femoral pulse wave velocity in older Black and White adults: the atherosclerosis risk in communities (ARIC) study

Introduction: Aortic stiffness offers important insight into vascular aging and cardiovascular disease (CVD) risk. The referent measure of aortic stiffness is carotid-femoral pulse wave velocity (cfPWV). cfPWV can be estimated (ePWV) from age and mean arterial pressure. Few studies have directly compared the association of ePWV to measured cfPWV, particularly in non-White adults. Moreover, whether ePWV and cfPWV correlate similarly with CVD risk remains unexplored. Aim: (1) To estimate the strength of the agreement between ePWV and cfPWV in both Black and White older adults; and (2) to compare the associations of ePWV and cfPWV with CVD risk factors and determine whether these associations were consistent across races. Methods and Results: We evaluated 4478 [75.2 (SD 5.0) years] Black and White older adults in the Atherosclerosis Risk in Communities (ARIC) Study. cfPWV was measured using an automated pulse waveform analyzer. ePWV was derived from an equation based on age and mean arterial pressure. Association and agreement between the two measurements were determined using Pearson’s correlation coefficient (r), standard error of estimate (SEE), and Bland-Altman analysis. Associations between traditional risk factors with ePWV and cfPWV were evaluated using linear mixed regression models. We observed weak correlations between ePWV and cfPWV within White adults (r = 0.36) and Black adults (r = 0.31). The mean bias for Bland-Altman analysis was low at -0.17 m/s (95%CI: -0.25 to -0.09). However, the inspection of the Bland-Altman plots indicated systematic bias (P < 0.001), which was consistent across race strata. The SEE, or typical absolute error, was 2.8 m/s suggesting high variability across measures. In models adjusted for sex, prevalent diabetes, the number of prevalent cardiovascular diseases, and medication count, both cfPWV and ePWV were positively associated with heart rate, triglycerides, and fasting glucose, and negatively associated with body mass index (BMI) and smoking status in White adults (P < 0.05). cfPWV and ePWV were not associated with heart rate, triglycerides, and fasting glucose in Black adults, while both measures were negatively associated with BMI in Black adults. Conclusions: Findings suggest a weak association between ePWV and cfPWV in older White and Black adults from ARIC. There were similar weak associations between CVD risk factors with ePWV and cfPWV in White adults with subtle differences in associations in Black adults. One sentence summary: Estimated pulse wave velocity is weakly associated with measured carotid-femoral pulse wave velocity in older Black and White adults in ARIC.

Association of Head Injury With Late-Onset Epilepsy: Results From the Atherosclerosis Risk in Communities Cohort

Background and Objectives:Late-onset epilepsy (LOE; i.e., epilepsy starting in later adulthood) is affects a significant number of individuals. Head injury is also a risk factor for acquired epilepsy, but the degree to which prior head injury may contribute to LOE is less well understood. Our objective was to determine the association between head injury and subsequent development of LOE.Methods:Included were 8,872 participants enrolled in the Atherosclerosis Risk in Communities (ARIC) study with continuous Centers for Medicare Services (CMS) fee-for-service (FFS) coverage (55.1% women, 21.6% black). We identified head injuries through 2018 from linked Medicare FFS claims for inpatient/emergency department care, active surveillance of hospitalizations, and participant self-report. LOE cases through 2018 were identified from linked Medicare FFS claims. We used Cox proportional hazards models to evaluate associations of head injury with LOE, adjusting for demographic, cardiovascular, and lifestyle factors.Results:The adjusted hazard ratio (HR) for developing LOE after a history of head injury was 1.88 (95%CI=1.44-2.43). There was evidence for dose-response associations with greater risk for LOE with increasing number of prior head injuries (HR=1.37, 95%CI=1.01-1.88 for 1 prior head injury and HR=3.55, 95%CI=2.51-5.02 for 2+ prior head injuries, compared to no head injuries) and with more severe head injury (HR=2.53, 95%CI=1.83-3.49 for mild injury and HR=4.90, 95%CI=3.15-7.64 for moderate/severe injury, compared to no head injuries). Associations with LOE were significant for head injuries sustained at older age (age≥67 years: HR=4.01, 95%CI=2.91-5.54), but not for head injuries sustained at younger age (age<67 years: HR=0.98, 95%CI=0.68-1.41).Discussion:Head injury was associated with increased risk of developing LOE, particularly when head injuries were sustained at an older age, and there was evidence for higher risk for LOE after a greater number of prior head injuries and after more severe head injuries.Classification of Evidence:This study provides Class I evidence that an increased risk of late-onset epilepsy is associated with head injury and increases further with multiple and more severe head injuries.

Associations of Serum Magnesium with Brain Morphology and Subclinical Cerebrovascular Disease: The Atherosclerosis Risk in Communities-Neurocognitive Study

Circulating magnesium has been associated with a lower risk of dementia, but the physiologic effects by which magnesium may prevent neurological insults remain unclear. We studied 1466 individuals (mean age 76.2 ± 5.3, 28.8% black, 60.1% female) free of prevalent stroke, with measured serum magnesium and with available MRI scans obtained in 2011–2013, participating in the Atherosclerosis Risk in Communities Neurocognitive Study (ARIC-NCS). Cross-sectional differences in frontal, temporal, parietal, and occipital lobe volume, along with deep grey matter, total brain, and white matter hyperintensity volume across serum magnesium (categorized into quintiles and per standard deviation increases) were assessed using multiple linear regression. We also examined associations of magnesium with the prevalence of cortical, subcortical, and lacunar infarcts using multiple logistic regression. After adjusting for demographics, biomarkers, medications, and cardiometabolic risk factors, higher circulating magnesium was associated with greater total brain volume and frontal, temporal, and parietal lobe volumes (volumes 0.14 to 0.19 standard deviations higher comparing Q5 to Q1). Elevated magnesium was also associated with lower odds of subcortical infarcts (OR (95%CI): 0.44 (0.25, 0.77) comparing Q5 to Q1) and lacunar infarcts (OR (95%CI): 0.40 (0.22, 0.71) comparing Q5 to Q1). Elevated serum magnesium was cross-sectionally associated with greater brain volumes and lower odds of subclinical cerebrovascular disease, suggesting beneficial effects on pathways related to neurodegeneration and cerebrovascular damage. Further exploration through prospective analyses is needed to assess increasing circulating magnesium as a potential neuroprotective intervention.

Obesity Duration, Severity, and Distribution Trajectories and Cardiovascular Disease Risk in the Atherosclerosis Risk in Communities Study

Background Research examining the role of obesity in cardiovascular disease (CVD) often fails to adequately consider heterogeneity in obesity severity, distribution, and duration. Methods and Results We here use multivariate latent class mixed models in the biracial Atherosclerosis Risk in Communities study (N=14 514; mean age=54 years; 55% female) to associate obesity subclasses (derived from body mass index, waist circumference, self‐reported weight at age 25, tricep skinfold, and calf circumference across up to four triennial visits) with total mortality, incident CVD, and CVD risk factors. We identified four obesity subclasses, summarized by their body mass index and waist circumference slope as decline (4.1%), stable/slow decline (67.8%), moderate increase (24.6%), and rapid increase (3.6%) subclasses. Compared with participants in the stable/slow decline subclass, the decline subclass was associated with elevated mortality (hazard ratio [HR] 1.45, 95% CI 1.31, 1.60, P <0.0001) and with heart failure (HR 1.41, 95% CI 1.22, 1.63, P <0.0001), stroke (HR 1.53, 95% CI 1.22, 1.92, P =0.0002), and coronary heart disease (HR 1.36, 95% CI 1.14, 1.63, P =0.0008), adjusting for baseline body mass index and CVD risk factor profile. The moderate increase latent class was not associated with any significant differences in CVD risk as compared to the stable/slow decline latent class and was associated with a lower overall risk of mortality (HR 0.85, 95% CI 0.80, 0.90, P <0.0001), despite higher body mass index at baseline. The rapid increase latent class was associated with a higher risk of heart failure versus the stable/slow decline latent class (HR 1.34, 95% CI 1.10, 1.62, P =0.004). Conclusions Consideration of heterogeneity and longitudinal changes in obesity measures is needed in clinical care for a more precision‐oriented view of CVD risk.

Association of Physical Activity with the Incidence of Atrial Fibrillation in Persons >65 Years Old: The Atherosclerosis Risk In Communities (ARIC) Study

ABSTRACTBackgroundThough moderate levels of physical activity (PA) seem to reduce the risk of atrial fibrillation (AF), the association of PA with AF in the elderly remains unclear.MethodsWe studied 5,166 participants of the Atherosclerosis Risk in Communities (ARIC) cohort that took part in visit 5 (2011-2013), were free of AF and had complete information on all variables. Self-reported PA was evaluated with a validated questionnaire and weekly minutes of leisure-time moderate to vigorous physical activity (MVPA) were calculated and categorized using the 2018 Physical Activity Guidelines for Americans (no activity [0 min/week], low [>0-<150 min/week], adequate [150-<300 min/week], high [≥300 min/week]). Incident AF between the visit 5 and the end of 2019 was ascertained from hospital discharges and death certificates. Cox models were used to calculate hazard ratios (HR) and 95% confidence intervals (CI) for AF by levels of physical activity adjusting for potential confounders.ResultsThe mean (SD) age for the sample was 75 (5) years; 59% were female and 22% were Black. During a mean (SD) follow-up time of 6.3 (2.0) years, 703 AF events were identified. The association of MVPA with AF incidence showed a U-shaped relationship. Compared to those not engaging in MVPA, individuals with low MVPA had a 23% lower hazard of AF (HR= 0.77; 95% CI: 0.61, 0.96), while those with adequate MVPA had a 14% lower hazard (HR 0.86; 95% CI: 0.69, 1.06). High levels of MVPA were not associated with AF risk (HR 0.97; 95% CI: 0.78, 1.20). There was no evidence of heterogeneity when stratified by race and sex.ConclusionThis study suggests that being involved in low to moderate levels of MVPA was associated with a reduced hazard of AF. There was no evidence of increased risk of AF in those with higher levels of MVPA.