Multinational Federated Learning Approach to Train ECG and Echocardiogram Models for Hypertrophic Cardiomyopathy Detection.

BACKGROUND: Novel targeted treatments increase the need for prompt hypertrophic cardiomyopathy (HCM) detection. However, its low prevalence (0.5%) and resemblance to common diseases present challenges that may benefit from automated machine learning-based approaches. We aimed to develop machine learning models to detect HCM and to differentiate it from other cardiac conditions using ECGs and echocardiograms, with robust generalizability across multiple cohorts. METHODS: Single-institution HCM ECG models were trained and validated on external data. Multi-institution models for ECG and echocardiogram were trained on data from 3 academic medical centers in the United States and Japan using a federated learning approach, which enables training on distributed data without data sharing. Models were validated on held-out test sets for each institution and from a fourth academic medical center and were further evaluated for discrimination of HCM from aortic stenosis, hypertension, and cardiac amyloidosis. Last, automated detection was compared with manual interpretation by 3 cardiologists on a data set with a realistic HCM prevalence. RESULTS: We identified 74 376 ECGs for 56 129 patients and 8392 echocardiograms for 6825 patients at the 4 academic medical centers. Although ECG models trained on data from each institution displayed excellent discrimination of HCM on internal test data (C statistics, 0.88-0.93), the generalizability was limited, most notably for a model trained in Japan and tested in the United States (C statistic, 0.79-0.82). When trained in a federated manner, discrimination of HCM was excellent across all institutions (C statistics, 0.90-0.96 and 0.90-0.96 for ECG and echocardiogram model, respectively), including for phenotypic subgroups. The models further discriminated HCM from hypertension, aortic stenosis, and cardiac amyloidosis (C statistics, 0.84, 0.83, and 0.88, respectively, for ECG and 0.93, 0.94, 0.85, respectively, for echocardiogram). Analysis of electrocardiography-echocardiography paired data from 11 823 patients from an external institution indicated a higher sensitivity of automated HCM detection at a given positive predictive value compared with cardiologists (0.98 versus 0.81 at a positive predictive value of 0.01 for ECG and 0.78 versus 0.59 at a positive predictive value of 0.24 for echocardiogram). CONCLUSIONS: Federated learning improved the generalizability of models that use ECGs and echocardiograms to detect and differentiate HCM from other causes of hypertrophy compared with training within a single institution.

Pulmonary vascular dysfunction among people aged over 65 years in the community in the Atherosclerosis Risk In Communities (ARIC) Study: A cross-sectional analysis.

BACKGROUND: Heart failure (HF) risk is highest in late life, and impaired pulmonary vascular function is a risk factor for HF development. However, data regarding the contributors to and prognostic importance of pulmonary vascular dysfunction among HF-free elders in the community are limited and largely restricted to pulmonary hypertension. Our objective was to define the prevalence and correlates of abnormal pulmonary pressure, resistance, and compliance and their association with incident HF and HF phenotype (left ventricular [LV] ejection fraction [LVEF] ≥ or < 50%) independent of LV structure and function. METHODS AND FINDINGS: We performed cross-sectional and time-to-event analyses in a prospective epidemiologic cohort study, the Atherosclerosis Risk in Communities study. This is an ongoing, observational study that recruited 15,792 persons aged 45-64 years between 1987 and 1989 (visit 1) from four representative communities in the United States: Minneapolis, Minnesota; Jackson, Mississippi; Hagerstown, Maryland; and Forsyth County, North Carolina. The current analysis included 2,810 individuals aged 66-90 years, free of HF, who underwent echocardiography at the fifth study visit (June 8, 2011, to August 28, 2013) and had measurable tricuspid regurgitation by spectral Doppler. Echocardiography-derived pulmonary artery systolic pressure (PASP), pulmonary vascular resistance (PVR), and pulmonary arterial compliance (PAC) were measured. The main outcome was incident HF after visit 5, and key secondary end points were incident HF with preserved LVEF (HFpEF) and incident HF with reduced LVEF (HFrEF). The mean ± SD age was 76 ± 5 years, 66% were female, and 21% were black. Mean values of PASP, PVR, and PAC were 28 ± 5 mm Hg, 1.7 ± 0.4 Wood unit, and 3.4 ± 1.0 mL/mm Hg, respectively, and were abnormal in 18%, 12%, and 14%, respectively, using limits defined from the 10th and 90th percentile limits in 253 low-risk participants free of cardiovascular disease or risk factors. Left heart dysfunction was associated with abnormal PASP and PAC, whereas a restrictive ventilatory deficit was associated with abnormalities of PASP, PVR, and PAC. PASP, PVR, and PAC were each predictive of incident HF or death (hazard ratio per SD 1.3 [95% CI 1.1-1.4], p < 0.001; 1.1 [1.0-1.2], p = 0.04; 1.2 [1.1-1.4], p = 0.001, respectively) independent of LV measures. Elevated pulmonary pressure was predictive of incident HFpEF (HFpEF: 2.4 [1.4-4.0, p = 0.001]) but not HFrEF (1.4 [0.8-2.5, p = 0.31]). Abnormal PAC predicted HFrEF (HFpEF: 2.0 [1.0-4.0, p = 0.05], HFrEF: 2.8 [1.4-5.5, p = 0.003]), whereas abnormal PVR was not predictive of either (HFpEF: 0.9 [0.4-2.0, p = 0.85], HFrEF: 0.7 [0.3-1.4, p = 0.30],). A greater number of abnormal pulmonary vascular measures was associated with greater risk of incident HF. Major limitations include the use of echo Doppler to estimate pulmonary hemodynamic measures, which may lead to misclassification; inclusions bias related to detectable tricuspid regurgitation, which may limit generalizability of our findings; and survivor bias related to the cohort age, which may result in underestimation of the described associations. CONCLUSIONS: In this study, we observed abnormalities of PASP, PVR, and PAC in 12%-18% of elders in the community. Higher PASP and lower PAC were independently predictive of incident HF. Abnormally high PASP predicted incident HFpEF but not HFrEF. These findings suggest that impairments in pulmonary vascular function may precede clinical HF and that a comprehensive pulmonary hemodynamic evaluation may identify pulmonary vascular phenotypes that differentially predict HF phenotypes.

Subclinical Risk Factors for Heart Failure With Preserved and Reduced Ejection Fraction Among Black Adults.

Importance: Sparse data exist regarding the contributions of subclinical impairments in cardiovascular and noncardiovascular function to incident heart failure (HF) with reduced ejection fraction (HFrEF) and preserved ejection fraction (HFpEF) among Black US residents, limiting understanding of the etiology of HF subtypes. Objectives: To identify subclinical cardiovascular and noncardiovascular risk factors associated with HFrEF and HFpEF in Black US residents. Design, Setting, and Participants: This cohort study used cross-sectional and time-to-event analysis with data from the community-based Jackson Heart Study (JHS), a longitudinal cohort study with baseline data collected from 2000 to 2004 (visit 1) and 10-year follow-up for incident HF. Black US residents from the Jackson, Mississippi, metropolitan area enrolled in JHS; those with prevalent HF, with moderate or greater aortic or mitral valve diseases on visit 1, who died before 2005, and who had missing HF status on follow-up were excluded. The analysis included 4361 participants and was performed between June 2020 to August 2021. Exposures: Quantitative measures of cardiovascular (left ventricular mass index [LVMI], left ventricular ejection fraction [LVEF], left atrial [LA] diameter, and pulse pressure) and noncardiovascular (percent predicted forced expiration volume in 1 second [FEV1 (percent predicted)], estimated glomerular filtration rate (eGFR), waist circumference, and hemoglobin A1c [HbA1c] level) organ function. Main Outcomes and Measures: Incident HF, HFrEF, and HFpEF over 10-year follow-up. Results: The 4361 participants had a mean (SD) age of 54 (13); 2776 (64%) were women; and there were 163 HFpEF and 146 HFrEF events. In multivariable models incorporating measures reflecting each organ system, factors associated with incident HFpEF included greater LA diameter (hazard ratio [HR], 1.23; 95% CI, 1.03-1.47; P = .02), higher pulse pressure (HR, 1.23; 95% CI, 1.05-1.44; P = .009), lower FEV1 (percent predicted) (HR, 1.22; 95% CI, 1.04-1.43; P = .02), lower eGFR (HR, 1.43; 95% CI, 1.19-1.72; P < .001), higher HbA1c level (HR, 1.25; 95% CI, 1.07-1.45; P = .005), and higher waist circumference (HR, 1.41; 95% CI, 1.18-1.69; P < .001). Factors associated with incident HFrEF included greater LVMI (HR, 1.25; 1.07-1.46; P = .005), lower LVEF (HR, 1.65; 95% CI, 1.42-1.91; P < .001), lower FEV1 (percent predicted) (HR, 1.19; 95% CI, 1.00-1.42; P = .047), and lower eGFR (HR, 1.27; 95% CI, 1.04-1.55; P = .02). Conclusions and Relevance: In this community-based cohort study of Black US residents, subclinical impairments in cardiovascular and noncardiovascular organ function were differentially associated with risk of incident HFpEF and HFrEF.

Coronary Artery Disease and Heart Failure With Preserved Ejection Fraction: The ARIC Study.

Background Whether coronary artery disease (CAD) is a significant risk factor for heart failure (HF) with preserved ejection fraction (HFpEF) is unclear. Methods and Results Among 9902 participants in the ARIC (Atherosclerosis Risk in Communities) study, we assessed the association of incident CAD with subsequent incident HFpEF (left ventricular ejection fraction [≥50%]) and HF with reduced ejection fraction (HFrEF; left ventricular ejection fraction <50%) using survival models with time-updated variables. We also assessed the extent to which echocardiographic correlates of prevalent CAD account for the relationship between CAD and incident HFpEF. Over 13-year follow-up, incident CAD developed in 892 participants and 178 subsequently developed HF (86 HFrEF, 71 HFpEF). Incident HFrEF and HFpEF risk were both greatest early after the CAD event. At >1 year post-CAD event, adjusted incidence of HFrEF and HFpEF were similar (7.2 [95% CI, 5.2-10.0] and 6.7 [4.8-9.2] per 1000 person-years, respectively) and CAD remained predictive of both (HFrEF: hazard ratio, 2.76 [95% CI, 1.99-3.84]; HFpEF: 1.85 [1.35-2.54]) after adjusting for demographics and common comorbidities. Among 4779 HF-free participants at Visit 5 (2011-2013), the 490 with prevalent CAD had lower left ventricular ejection fraction and higher left ventricular mass index, E/e', and left atrial volume index (all P<0.01). The association of prevalent CAD with incident HFpEF post-Visit 5 was not significant after adjusting for echocardiographic measures, with the greatest attenuation observed for left ventricular diastolic function. Conclusions CAD is a significant risk factor for incident HFpEF after adjustment for demographics and common comorbidities. This relationship is partially accounted for by echocardiographic alterations, particularly left ventricular diastolic function.

Depressive Symptoms, Cardiac Structure and Function, and Risk of Incident Heart Failure With Preserved Ejection Fraction and Heart Failure With Reduced Ejection Fraction in Late Life.

Background Depressive symptoms are associated with heightened risk of heart failure (HF), but their association with cardiac function and with HF with preserved ejection fraction (HFpEF) and HF with reduced ejection fraction (HFrEF) in late life is unclear. We aimed to determine the prevalence of depression in HFpEF and in HFrEF in late life, and the association of depressive symptoms with cardiac function and incident HFpEF and HFrEF. Methods and Results We studied 6025 participants (age, 75.3±5.1 years; 59% women; 20% Black race) in the ARIC (Atherosclerosis Risk in Communities) study at visit 5 who underwent echocardiography and completed the Center for Epidemiologic Studies Depression Scale questionnaire. Among HF-free participants (n=5086), associations of Center for Epidemiologic Studies Depression Scale score with echocardiography and incident adjudicated HFpEF and HFrEF were assessed using multivariable linear and Cox proportional hazards regression. Prevalent HFpEF, but not HFrEF, was associated with a higher prevalence of depression compared with HF-free participants (P<0.001 and P=0.59, respectively). Among HF-free participants, Center for Epidemiologic Studies Depression Scale score was not associated with cardiac structure and function after adjusting for demographics and comorbidities (all P>0.05). Over 5.5-year follow-up, higher Center for Epidemiologic Studies Depression Scale score was associated with heightened risk of incident HFpEF (hazard ratio [HR] [95% CI], 1.06 [1.04-1.12]; P=0.02), but not HFrEF (HR [95% CI], 1.02 [0.96-1.08]; P=0.54), independent of echocardiographic measures, NT-proBNP (N-terminal pro-B-type natriuretic peptide), troponin, and hs-CRP (high-sensitivity C-reactive protein) (HR [95% CI], 1.06 [1.00-1.12]; P=0.04). Conclusions Worse depressive symptoms predict incident HFpEF in late life, independent of common comorbidities, cardiac structure and function, and prognostic biomarkers. Further studies are necessary to understand the mechanisms linking depression to risk of HFpEF.

Mid- to Late-Life Time-Averaged Cumulative Blood Pressure and Late-Life Cardiac Structure, Function, and Heart Failure.

Limited data exist regarding systolic blood pressure (SBP) through mid- to late-life and late-life cardiac function and heart failure (HF) risk. Among 4578 HF-free participants in the ARIC study (Atherosclerosis Risk in Communities) attending the fifth visit (2011-2013; age 75±5 years), time-averaged cumulative SBP was calculated as the sum of averaged SBPs from adjacent consecutive visits (visits 1-5) indexed to total observation time (24±1 years). Calculations were performed using measured SBPs and also incorporating antihypertensive medication specific effect constants (underlying SBP). Outcomes included comprehensive echocardiography at visit 5 and post-visit 5 incident HF, HF with preserved ejection fraction, and reduced ejection fraction. Higher cumulative SBP was associated with greater left ventricular mass and worse diastolic measures (all P<0.001), associations that were stronger with underlying compared with cumulative SBP (all P<0.05). At 5.6±1.2 years follow-up post-visit 5, higher cumulative measured and underlying SBP were associated with incident HF (hazard ratio per 10 mm Hg for measured: 1.12 [1.01-1.24]; underlying: 1.19 [95% CI, 1.10-1.30]) and HF with preserved ejection fraction (measured: 1.15 [1.00-1.33]; underlying: 1.28 [1.14-1.45]), but not HF with reduced ejection fraction (measured: 1.11 [0.94-1.32]; underlying: 1.11 [0.96-1.24]). Associations with HF and HF with preserved ejection fraction were more robust with cumulative underlying compared with measured SBP (all P<0.05). Time-averaged cumulative SBP in mid to late life is associated with worse cardiac function and risk of incident HF, especially HF with preserved ejection fraction, in late life. These associations were stronger considering underlying as opposed to measured SBP, highlighting the importance of prevention and effective treatment of hypertension to prevent late-life cardiac dysfunction and HF.

Myocardial Infarction in Heart Failure With Preserved Ejection Fraction: Pooled Analysis of 3 Clinical Trials.

OBJECTIVES: The authors investigated the relationship between past or incident myocardial infarction (MI) and cardiovascular (CV) events in heart failure with preserved ejection fraction (HFpEF). BACKGROUND: MI and HFpEF share some common risk factors. The prognostic significance of MI in patients with HFpEF is uncertain. METHODS: The authors pooled data from 3 trials-CHARM Preserved (Candesartan Cilexietil in Heart Failure Assessment of Reduction in Mortality and Morbidity), I-Preserve (Irbesartan in Heart Failure With Preserved Systolic Function), and the Americas region of TOPCAT (Treatment of Preserved Cardiac Function Heart Failure With an Aldosterone Antagonist) (N = 8,916)-and examined whether MI before or following enrollment independently predicted CV death and heart failure (HF) hospitalization. RESULTS: At baseline, 2,668 patients (30%) had history of MI. Prior MI was independently associated with greater risk of CV death (4.7 vs. 3.5 events/100 patient-years [py], adjusted hazard ratio [HR]: 1.42 [95% confidence interval (CI): 1.23 to 1.64]; p < 0.001). Excess sudden death drove this difference (1.9 vs. 1.2 events/100 py, adjusted HR: 1.55 [95% CI: 1.23 to 1.97]; p < 0.001). There was no difference in HF hospitalization (5.9 vs. 5.5 events/100 py, adjusted HR: 1.05, 95% CI: 0.92 to 1.19) or HF death by prior MI. During follow-up, MI occurred in 336 patients (3.8%). Risk of CV death increased 31-fold in the first 30 days after first post-enrollment MI, and remained 58% higher beyond 1 year after MI. Risk of first or recurrent HF hospitalization increased 2.4-fold after MI. CONCLUSIONS: Prior MI in HFpEF is associated with greater CV and sudden death but similar risk of HF outcomes. Patients with HFpEF who experience MI are at high risk of subsequent CV death and HF hospitalization. These data highlight the importance of primary and secondary prevention of MI in patients with HFpEF. (Candesartan Cilexietil in Heart Failure Assessment of Reduction in Mortality and Morbidity [CHARM Preserved]; NCT00634712; Irbesartan in Heart Failure With Preserved Systolic Function [I-Preserve]; NCT00095238; and Treatment of Preserved Cardiac Function Heart Failure With an Aldosterone Antagonist [TOPCAT]; NCT00094302).

Transmeningeal delivery of GABA to control neocortical seizures in rats.

Transmeningeal drug delivery, using an implanted hybrid neuroprosthesis, has been proposed as a novel therapy for intractable focal epilepsy. As part of a systematic effort to identify the optimal compounds and protocols for such a therapy, this study aimed to determine whether transmeningeal gamma-aminobutyric acid (GABA) delivery can terminate and/or prevent neocortical seizures in rats. Rats were chronically implanted with an epidural cup and an adjacent EEG electrode in the right parietal cortex. While the rat was behaving freely, a seizure-inducing concentration of acetylcholine (Ach) was applied into the cup. In a seizure termination study, either artificial cerebrospinal fluid (ACSF) or GABA (0.25, 2.5, 25 or 50mM) was delivered into the exposed neocortical area during an ongoing seizure. In a seizure prevention study, either ACSF or 50mM GABA was delivered into the epidural cup before the application of Ach. Epidural delivery of 50mM GABA completely terminated ongoing Ach-induced EEG seizures and convulsions within 17-437s after its delivery. ACSF and lower concentrations of GABA did not produce this effect, but 25mM GABA reduced seizure severity. However, the used GABA concentration could not prevent the development, or affect the severity, of Ach-induced EEG seizures and convulsions. This study indicates that transmeningeal GABA delivery can be used for terminating neocortical seizures, but to achieve seizure prevention via this route either a more efficient GABA delivery method needs to be developed or other neurotransmitters/pharmaceuticals should be employed for this purpose.

KLF15 and PPARα Cooperate to Regulate Cardiomyocyte Lipid Gene Expression and Oxidation.

The metabolic myocardium is an omnivore and utilizes various carbon substrates to meet its energetic demand. While the adult heart preferentially consumes fatty acids (FAs) over carbohydrates, myocardial fuel plasticity is essential for organismal survival. This metabolic plasticity governing fuel utilization is under robust transcriptional control and studies over the past decade have illuminated members of the nuclear receptor family of factors (e.g., PPARα) as important regulators of myocardial lipid metabolism. However, given the complexity of myocardial metabolism in health and disease, it is likely that other molecular pathways are likely operative and elucidation of such pathways may provide the foundation for novel therapeutic approaches. We previously demonstrated that Kruppel-like factor 15 (KLF15) is an independent regulator of cardiac lipid metabolism thus raising the possibility that KLF15 and PPARα operate in a coordinated fashion to regulate myocardial gene expression requisite for lipid oxidation. In the current study, we show that KLF15 binds to, cooperates with, and is required for the induction of canonical PPARα-mediated gene expression and lipid oxidation in cardiomyocytes. As such, this study establishes a molecular module involving KLF15 and PPARα and provides fundamental insights into the molecular regulation of cardiac lipid metabolism.

Epidural pentobarbital delivery can prevent locally induced neocortical seizures in rats: the prospect of transmeningeal pharmacotherapy for intractable focal epilepsy.

PURPOSE: To determine whether epidural pentobarbital (PB) delivery can prevent and/or terminate neocortical seizures induced by locally administered acetylcholine (Ach) in freely moving rats. METHODS: Rats were implanted permanently with an epidural cup placed over the right parietal cortex with intact dura mater. Epidural screw-electrodes, secured to the cup, recorded local neocortical EEG activity. In the seizure-termination study, Ach was delivered into the epidural cup, and after the development of electrographic and behavioral seizures, the Ach solution was replaced with either PB or artificial cerebrospinal fluid (aCSF; control solution). In the seizure-prevention study, the epidural Ach delivery was preceded by a 10-min exposure of the delivery site to PB or aCSF. Raw EEG recordings, EEG power spectra, and behavioral events were analyzed. RESULTS: Ach-induced EEG seizures associated with convulsions, which were unaffected by epidural aCSF applications, were terminated by epidurally delivered PB within 2-2.5 min. Epidural deliveries of PB before Ach applications completely prevented the development of electrographic and behavioral seizures, whereas similar deliveries of aCSF exerted no influence on the seizure-generating potential of Ach. CONCLUSIONS: This study showed for the first time that epidural AED delivery can prevent, as well as terminate, locally induced neocortical seizures. The findings support the viability of transmeningeal pharmacotherapy for the treatment of intractable neocortical epilepsy.