Publications

BACKGROUND: Low left ventricular ejection fraction (LEF) can progress undiagnosed. Artificial intelligence-based electrocardiogram (ECG-AI) screening may provide a scalable means to detect LEF.

OBJECTIVES: The purpose of this study was to validate a complete ECG-AI software as a medical device for LEF detection.

METHODS: Four geographically diverse sites in the United States identified patients with both ECGs and transthoracic echocardiograms performed within 30 days of each other in clinical practice. Data were electronically extracted to specific guidelines and transmitted to the coordinating center for analysis.

RESULTS: Records of 16,000 subjects were extracted, resulting in an evaluable set of 13,960 subjects (mean age 66 years; 52% male). The device demonstrated excellent discrimination (AUROC: 0.92 [95% CI: 0.91-0.93]) and was 84.5% (95% CI: 82.2%-86.6%) sensitive and 83.6% (95% CI: 82.9%-84.2%) specific for LEF. The overall prevalence of LEF in the study data set was 7.9%, with LEF among 1.6% of the ECG-AI negative and 30.5% of ECG-AI positive subjects, contributing to positive and negative predictive values of 30.5% (95% CI: 28.8%-32.1%) and 98.4% (95% CI: 98.2%-98.7%), respectively.

CONCLUSIONS: External validation studies such as this one provide a rigorous framework to validate an algorithm's performance. This study demonstrated the algorithm's strong diagnostic accuracy over a geographically diverse, independent set of patients. In this generally unselected population, the algorithm produced a test negative result in 78% of the cases, suggesting potential utility as a rule-out strategy to defer echocardiography when other clinical findings are absent.

OBJECTIVES: To examine the determinants of unmet transcatheter aortic valve replacement (TAVR) needs and their impact on patient survival among Medicare beneficiaries with aortic stenosis.

METHODS: We developed a county-level mismatch score measuring the gap between actual TAVR procedures performed and expected need based on population differences. Counties were classified as metropolitan, semiurban, or rural. Factors associated with larger mismatches were identified, and mortality rates among aortic stenosis (AS) patients were examined in relation to mismatch scores. We analyzed Medicare data from 2016 to 2022 across 3129 US counties. The mismatch score was developed to account for population differences and county urbanicity classification. Statistical analyses identified factors associated with TAVR mismatch and its relationship to mortality outcomes.

RESULTS: We found substantial geographic variation in TAVR delivery. Counties with higher TAVR mismatch scores showed associations with fewer TAVR-providing hospitals, less market concentration, higher AS prevalence, and lower household incomes. Counties with greater gaps between needed and actual TAVR procedures were also associated with higher mortality rates. This relationship between mismatch and mortality was particularly strong in semiurban counties.

CONCLUSIONS: Our findings identify associations between TAVR access gaps and patient outcomes, as well as factors linked to these access patterns. Counties with higher TAVR mismatch scores showed correlations with healthcare capacity constraints, geographic location, and socioeconomic factors. These associations suggest that mismatches may be addressed through targeted approaches based on local needs to improve care delivery for patients with AS in regions currently experiencing access challenges.

BACKGROUND: Sepsis remains a leading cause of mortality, and optimizing treatment is challenging due to patient heterogeneity. Identification of cardiac phenotypes may inform precision medicine approaches and guide resuscitation. We performed a clustering analysis of patients with sepsis using echocardiographic data without using any a priori definitions of cardiac dysfunction or outcomes to establish the subgroups.

METHODS: This was a retrospective cohort study of patients admitted to the hospital with sepsis at a single academic center. Patients were identified using sepsis-related ICD codes, and those who had echocardiogram performed within 14 days of admission underwent chart review to ensure sepsis-3 criteria were met. Those with preexisting heart disease were excluded. Clustering by echocardiographic variables was performed using latent profile analysis. Clinical features such as patient characteristics, laboratory studies, sepsis source, and outcomes were compared across the clusters.

RESULTS: There were 2,071 patients included in the analysis. Our cluster analysis yielded five phenotypes: cluster 1, elevated mean E/e' 24.5 (SD 9.6); cluster 2, reduced ejection fraction, mean 33.1% (SD 10.6), and cardiac index 2.6 L/min/m2 (SD 0.9); cluster 3, right ventricular dilation with right ventricular basal diameter 4.5 cm (SD 0.9) and elevated tricuspid regurgitation gradient 60.0 mmHg (SD 13.5); cluster 4, hyperdynamic with mean left ventricular ejection fraction 75% (SD10.9) and mean cardiac index 6.6 L/min/m2 (SD 2.6); and lastly cluster 5, normal echocardiographic parameters. Group 3 had the highest mortality compared to the normal group (36.9% vs. 19.6%, p = 0.002), with an odds ratio of 2.3 (95%CI 1.4-3.9).

CONCLUSIONS: Using an unsupervised clustering analysis, we identified five phenotypes of cardiac function in sepsis based on commonly recorded echocardiographic data: diastolic dysfunction, left ventricular systolic dysfunction with low cardiac index, right ventricular dilation with elevated tricuspid regurgitation gradient, hyperdynamic cardiac function, and normal. The right ventricular dilation group had the highest mortality. Future research should explore mechanisms and potential treatment implications for these groups.

IMPORTANCE: Chronic kidney disease (CKD) is common and often coexists with cardiometabolic risk factors and cardiovascular disease (CVD).

OBJECTIVES: To evaluate CKD prevalence and awareness among US adults overall and in those with cardiometabolic risk factors or CVD.

DESIGN, SETTING, AND PARTICIPANTS: This serial cross-sectional study was conducted among US adults aged 20 years or older participating in the National Health and Nutrition Examination Survey between 2011 and March 2020.

MAIN OUTCOMES AND MEASURES: The primary outcomes were prevalence of CKD, defined as estimated glomerular filtration rate less than 60 mL/min/1.73 m2 or urine albumin to creatinine ratio of 30 mg/g or greater, and awareness, based on self-report of a "yes" response to the question "Ever told you had weak/failing kidneys?" among all US adults and those with cardiometabolic conditions (hypertension, diabetes, hyperlipidemia, obesity) or CVD. Survey-weighted logistic regression models were also fit to determine temporal changes in prevalence over the study period.

RESULTS: This cross-sectional study included 24 646 adults (weighted mean age, 49 years; 48.4% female), including 20 224 adults with cardiometabolic risk factors or CVD. The overall prevalence of CKD among US adults was 14.6% (95% CI, 14.0%-15.3%), and only 12.3% (95% CI, 11.1%-13.5%) were aware of "weak/failing" kidneys. Among adults with cardiometabolic risk factors or CVD, CKD prevalence was 16.7% (95% CI, 16.0%-17.4%). Awareness of "weak/failing" kidneys was low in this population-only 13.2% (95% CI, 11.9%-14.4%) were aware of their diagnosis over the study period, and the largest awareness gaps occurred among those aged 20 to 64 years, women, and Hispanic adults. Although awareness among adults with CKD and cardiometabolic conditions increased modestly, from 11.5% (95% CI, 8.5%-14.5%) in 2011-2012 to 15.1% (95% CI, 13.1%-17.2%) in 2017 through March 2020 (P = .02), these gains were concentrated among older adults aged 65 years or older (10.8%; 95% CI, 6.9%-14.6% to 17.7%; 95% CI, 14.2%-21.3%), men (9.7%; 95% CI, 5.6%-13.8% to 18.4%; 95% CI, 15.5%-21.4%), and non-Hispanic White adults (10.8%; 95% CI, 6.1%-15.5% to 16.3%; 95% CI, 13.4%-19.2%). No significant improvements in awareness were observed among younger adults aged 20 to 64 years, women, or Black and Hispanic adults.

CONCLUSIONS AND RELEVANCE: In this nationally representative study, CKD affected 1 in 6 US adults with cardiometabolic conditions, and only a minority of respondents were aware of "weak/failing" kidneys. These findings underscore a significant opportunity to promote awareness and optimal management of CKD.