Daily Cardiology Research Analysis
Analyzed 204 papers and selected 3 impactful papers.
Summary
Global validation across 6.4 million individuals shows PREVENT and SCORE2 cardiovascular risk equations perform well across regions and settings, with new scaling factors enabling 1–9 year risk estimation. AI-ECG accurately detects left ventricular systolic dysfunction in a resource-limited Kenyan setting, supporting scalable screening. In inoperable chronic thromboembolic pulmonary hypertension, balloon pulmonary angioplasty was associated with substantially improved 8-year survival, including with partial treatment.
Research Themes
- Global validation and transportability of cardiovascular risk prediction models
- AI-enabled diagnostics for heart failure risk in resource-limited settings
- Long-term outcomes of interventional therapy in pulmonary vascular disease
Selected Articles
1. Multinational validation of the PREVENT and SCORE2 cardiovascular risk equations across 6.4 million individuals.
Across 44 cohorts and 18 trials (n≈6.42M for PREVENT; n≈5.44M for SCORE2), both risk equations demonstrated strong discrimination and calibration across regions. Newly provided scaling factors enable 1–9 year risk estimation, facilitating shorter-horizon prediction for research and trial enrollment.
Impact: This work provides rare, large-scale, multinational external validation of widely used CVD risk equations and introduces practical scaling to shorter horizons, increasing transportability and implementation potential.
Clinical Implications: Clinicians and health systems can more confidently apply PREVENT or SCORE2 across diverse populations, including integration into EHRs, and use 1–9 year predictions for targeted prevention and trial screening.
Key Findings
- Both PREVENT and SCORE2 showed similar discrimination and calibration across North America, Europe, Asia/other regions, and multi-regional trials.
- Over a mean 5.1 years, 293,737 PREVENT total CVD and 258,086 SCORE2 CVD events were observed among 6.42M and 5.44M individuals, respectively.
- New scaling factors allow PREVENT-based prediction over 1–9 years, enabling shorter-term risk estimation for research and trial enrollment.
Methodological Strengths
- Extremely large, multi-regional dataset (44 cohorts, 18 RCTs) with consistent evaluation of discrimination and calibration
- Introduces validated scaling to 1–9 year horizons to enhance practical applicability
Limitations
- Heterogeneity in outcome definitions and predictor sets across cohorts and regions
- Observational validation; not designed to test treatment decision thresholds prospectively
Future Directions: Prospective impact studies testing clinical outcomes from model-guided prevention across regions; calibration refinement in underrepresented ancestries and health systems.
The American Heart Association's PREVENT equations estimate risk of total cardiovascular disease (CVD), atherosclerotic CVD (ASCVD), and heart failure (HF) to guide lipid and blood pressure-lowering therapy in people ages 30 to 79 years in the United States. The SCORE2 risk algorithm is used to estimate CVD risk for similar purposes in people ages 40 and older in Europe. Neither set of equations has been comprehensively validated in global observational cohorts and randomized trials. Here, in 44 observational cohorts and 18 randomized trials, we assessed discrimination and calibration of the two risk algorithms across geographical regions (North America, Europe, Asia/other, multi-region trials). We also created scaling factors for risk prediction over 1-9 years using the PREVENT equations, enabling shorter-term risk prediction for research purposes or to facilitate clinical trial enrolment. Over 5.1 years of mean follow-up, 293,737 PREVENT total CVD events (fatal and non-fatal ASCVD or HF) and 258,086 SCORE2 CVD events (myocardial infarction, stroke, or cardiovascular death) were observed among 6,422,714 and 5,437,384 individuals, respectively. Despite differences in CVD outcome definitions, target populations and predictor variables, overall discrimination and calibration were similar for both equations, with generally good performance across regions, including in multi-regional randomized trials. These findings lend support for adoption of PREVENT or SCORE2 for cardiovascular risk stratification across diverse settings.
2. Artificial Intelligence Electrocardiogram and Left Ventricular Systolic Dysfunction in Kenya.
In 1,444 Kenyan adults across 8 outpatient facilities, an AI-ECG algorithm detected LV systolic dysfunction (LVEF <40%) with 95.6% sensitivity, 79.4% specificity, and AUC 0.96 compared to echocardiography. The negative predictive value was 99.1%, supporting AI-ECG as a scalable screening tool in resource-limited settings.
Impact: Demonstrates high diagnostic accuracy of AI-ECG for LV systolic dysfunction in a real-world, resource-limited context, addressing a key access gap to echocardiography.
Clinical Implications: AI-ECG can be deployed to screen for LV dysfunction and triage patients for echocardiography, potentially enabling earlier initiation of guideline-directed medical therapy for heart failure.
Key Findings
- AI-ECG identified LVSD with sensitivity 95.6%, specificity 79.4%, and AUC 0.96 against echocardiographic gold standard.
- Negative predictive value was 99.1%, supporting its role as an effective rule-out test.
- Performance was consistent across prespecified cardiovascular risk strata (AUC 0.96–0.98).
Methodological Strengths
- Multisite, real-world deployment across 8 facilities with paired echocardiography confirmation
- Robust diagnostic metrics with narrow confidence intervals and stratified performance analyses
Limitations
- Cross-sectional design; no longitudinal outcomes to assess downstream clinical impact
- Positive predictive value was modest, implying need for confirmatory echocardiography
Future Directions: Prospective implementation trials to measure impact on time to diagnosis, HF therapy uptake, and outcomes; evaluation of workflow integration and cost-effectiveness.
IMPORTANCE: Early detection of risk of heart failure with reduced ejection fraction remains challenging in resource-limited settings due to limited access to echocardiography. Artificial intelligence electrocardiogram (AI-ECG) algorithms have demonstrated promise for identifying left ventricular systolic dysfunction (LVSD), but their feasibility in resource-constrained settings remains unknown. OBJECTIVE: To determine the frequency of patients in Kenya with a high probability of LVSD by AI-ECG and assess AI-ECG algorithm performance against the gold standard of echocardiography. DESIGN, SETTING, AND PARTICIPANTS: This was a cross-sectional study with enrollment from June to December 2024. Participants underwent baseline assessment and 12-lead ECG, and a subset completed echocardiography within 7 days. The echocardiography subset included participants from 3 prespecified risk strata: those with prior cardiovascular disease, those at high cardiovascular risk (Framingham Risk Score [FRS] ≥10%), and those at low risk (FRS <10%). The study took place at 8 outpatient health care facilities across Kenya. A total of 1444 patients 18 years and older seeking routine care were enrolled and completed paired echocardiogram. Exclusion criteria included inability to provide informed consent. EXPOSURE: Risk of LVSD was identified using a validated convolutional neural network AI-ECG algorithm (AiTiALVSD). MAIN OUTCOMES AND MEASURES: Key outcomes were the diagnostic performance (sensitivity, specificity, and positive and negative predictive values) of the AI-ECG algorithm for detecting LVSD (LVEF <40%) when confirmed on echocardiography. RESULTS: Among 1444 participants (mean [SD] age, 59.0 [16.7] years; 907 [62.8%] female; 1118 [77.4%] at high risk), LVSD was identified in 204 (14.1%). The AI-ECG algorithm had a sensitivity of 95.6% (95% CI, 91.8-97.7), specificity of 79.4% (95% CI, 77.0-81.5), positive predictive value of 43.2% (95% CI, 38.7-47.9), negative predictive value of 99.1% (95% CI, 98.3-99.5), and area under the receiver operating characteristic curve (AUC) of 0.96 (95% CI, 0.95-0.97). Performance remained consistent across cardiovascular risk strata (AUC, 0.96-0.98). CONCLUSIONS AND RELEVANCE: In this study, the AI-ECG algorithm demonstrated the potential clinical utility for screening of LVSD risk with high sensitivity and negative predictive value and may be particularly scalable in a resource-limited setting.
3. Long-Term Survival of Balloon Pulmonary Angioplasty for Inoperable Chronic Thromboembolic Pulmonary Hypertension: A Multicenter Study.
In inoperable CTEPH, BPA was associated with markedly improved 8-year survival (86.7% vs 57.8%; HR 0.20), with benefits observed even among partial BPA sessions (HR 0.38). Results underscore potential mortality reduction but require confirmation in RCTs.
Impact: Provides long-term survival evidence for BPA in a population without surgical options, informing practice and trial design in a rare, high-mortality disease.
Clinical Implications: For inoperable CTEPH, referral to BPA-capable centers should be prioritized; even partial revascularization may confer survival benefits when full completion is not feasible.
Key Findings
- 8-year survival was 86.7% with BPA vs 57.8% without BPA (P<0.001).
- BPA was associated with reduced all-cause mortality (HR 0.20; 95% CI 0.12–0.32).
- Partial BPA also improved survival vs non-BPA (HR 0.38; 95% CI 0.22–0.70).
Methodological Strengths
- Multicenter design with long median follow-up (6.0 years) and survival analyses
- Comparison with contemporaneous non-BPA cohort and subgroup analysis (full vs partial BPA)
Limitations
- Retrospective, nonrandomized design with potential selection bias and residual confounding
- Generalizability limited to centers with BPA expertise
Future Directions: Randomized controlled trials comparing BPA strategies and optimized medical therapy; studies defining patient selection, session intensity, and partial vs full revascularization thresholds.
BACKGROUND: Balloon pulmonary angioplasty (BPA) is recommended for inoperable chronic thromboembolic pulmonary hypertension (CTEPH). OBJECTIVES: The aim of this study was to evaluate the long-term survival benefit of BPA for inoperable CTEPH, especially partial BPA sessions. METHODS: In this multicenter cohort study, 232 patients undergoing BPA (the BPA group) and 70 patients refusing the BPA procedure (the non-BPA group) were enrolled. The BPA group was further divided into the full-BPA group (129 patients) and the partial-BPA group (103 patients). The primary outcome was all-cause mortality. RESULTS: During a median follow-up time of 6.0 years (Q1-Q3: 3.7-7.3), 17 and 26 patients in the BPA and non-BPA groups died, contributing to 8-year survival rates of 86.7% (95% CI: 80.1%-93.8%) and 57.8% (95% CI: 45.9%-72.8%) in the BPA and non-BPA groups, respectively (P < 0.001, log-rank test). BPA was associated with significantly reduced all-cause mortality in inoperable CTEPH patients (HR: 0.20; 95% CI: 0.12-0.32; P < 0.001). In secondary analysis, the 8-year survival rates were 97.1% (95% CI: 93.8%-99.9%) and 70.0% (95% CI: 55.8%-87.8%) in the full-BPA and partial-BPA groups, respectively, both better than the non-BPA group (P < 0.001, log-rank test). Compared with the non-BPA group, partial BPA was associated with significantly reduced all-cause mortality in inoperable CTEPH patients (HR: 0.38; 95% CI: 0.22-0.70; P = 0.001). CONCLUSIONS: BPA tended to be associated with a reduced risk for all-cause mortality in patients with inoperable CTEPH, even those undergoing partial BPA sessions. These findings are preliminary and must be confirmed in randomized controlled trials.