Daily Cardiology Research Analysis
Three impactful cardiology studies stood out today: two implementation- and AI-focused diagnostics and one pragmatic systems-level randomized trial. A wrist-worn AI device matched Holter accuracy for atrial fibrillation over 28 days, an AI echocardiography model improved HFpEF classification and prognostic stratification, and semiautomated pharmacist referrals substantially increased evidence-based statin prescribing in high-risk patients.
Summary
Three impactful cardiology studies stood out today: two implementation- and AI-focused diagnostics and one pragmatic systems-level randomized trial. A wrist-worn AI device matched Holter accuracy for atrial fibrillation over 28 days, an AI echocardiography model improved HFpEF classification and prognostic stratification, and semiautomated pharmacist referrals substantially increased evidence-based statin prescribing in high-risk patients.
Research Themes
- AI-enabled cardiovascular diagnostics
- Implementation science to close prevention gaps
- Digital health and remote rhythm monitoring
Selected Articles
1. Encouraging Pharmacist Referrals for Evidence-Based Statin Initiation: Two Cluster Randomized Clinical Trials.
Across two pragmatic cluster RCTs (n=1412 and n=1950), a visit-based interruptive EHR alert did not significantly increase statin prescribing versus usual care, whereas an asynchronous semiautomated pharmacist referral strategy increased prescribing by 16.4 percentage points (31.6% vs 15.2%). This scalable workflow leverages pharmacists to close preventive care gaps in high-risk patients.
Impact: Pragmatic, system-level randomization demonstrates an immediately implementable strategy to increase evidence-based statin use, a cornerstone of cardiovascular prevention.
Clinical Implications: Health systems should consider implementing asynchronous semiautomated pharmacist referrals for lipid management to increase appropriate statin prescribing in high-risk patients, rather than relying on interruptive alerts.
Key Findings
- Visit-based interruptive EHR alerts did not significantly increase statin prescribing compared with usual care (15.6% vs 11.6%).
- Semiautomated pharmacist referral increased statin prescribing by 16.4 percentage points versus usual care (31.6% vs 15.2%).
- Trials enrolled 1412 and 1950 patients across 12 practices with mean 10-year ASCVD risk 17.9%.
Methodological Strengths
- Two pragmatic cluster RCTs across multiple primary care practices
- Clear primary outcome with prespecified analysis and large samples
Limitations
- Single health system; generalizability may be limited
- Outcome focused on prescribing, not downstream clinical events
Future Directions: Evaluate long-term adherence and clinical outcomes, and test scalability across diverse health systems and payers.
IMPORTANCE: Despite statins' benefit in preventing major adverse cardiovascular events, most patients with an indication for statin therapy are not appropriately treated. Clinicians' limited time and lack of systematic efforts to address preventive care likely contribute to gaps in statin prescribing. OBJECTIVE: To determine the effect on statin prescribing of 2 interventions to refer appropriate patients to a pharmacist for lipid management. DESIGN, SETTING, AND PARTICIPANTS: These 2 pragmatic cluster randomized clinical trials were conducted among 12 total primary care practices in a community health system. Trial 1 was a delayed-intervention design of a visit-based intervention with randomization at the clinician level in a single clinic, and trial 2 was a parallel-arm trial of an asynchronous intervention with randomization at the clinic level in 11 clinics. Patients who were assigned to a primary care clinician at a participating practice, had an indication for a high-intensity or moderate-intensity statin, and were either not prescribed a statin or prescribed an inappropriately low statin dose were eligible for inclusion.
2. External validation of artificial intelligence for detection of heart failure with preserved ejection fraction.
In a matched cohort (n=496), an AI echocardiography model (EchoGo Heart Failure v2) showed similar discrimination to H2FPEF/HFA-PEFF but reduced intermediate classifications and improved net classification and clinical decision impact. Positive AI classification identified patients with a two-fold higher risk of death or HF hospitalization, adding prognostic value.
Impact: Addresses a major diagnostic gap in HFpEF using validated AI on routine echocardiograms, with demonstrated decision-support and prognostic utility.
Clinical Implications: Integrating AI-based echocardiography analysis into HFpEF pathways can reduce indeterminate cases, prioritize further testing, and identify patients at higher risk for closer management.
Key Findings
- AI HFpEF model achieved similar discrimination and calibration to H2FPEF/HFA-PEFF but with fewer intermediate classifications.
- Integration of AI with existing scores improved correct management decisions.
- AI-positive patients had approximately two-fold higher risk of mortality or HF hospitalization.
Methodological Strengths
- External validation with matched controls and multiple performance metrics
- Assessment of clinical utility and prognostic associations
Limitations
- Single-model evaluation; generalizability to diverse scanners and populations requires further study
- Observational design; no randomized clinical impact assessment
Future Directions: Prospective multi-center impact studies, integration with workflows, and calibration across devices/vendors to ensure equitable performance.
Artificial intelligence (AI) models to identify heart failure (HF) with preserved ejection fraction (HFpEF) based on deep-learning of echocardiograms could help address under-recognition in clinical practice, but they require extensive validation, particularly in representative and complex clinical cohorts for which they could provide most value. In this study enrolling patients with HFpEF (cases; n = 240), and age, sex, and year of echocardiogram matched controls (n = 256), we compare the diagnostic performance (discrimination, calibration, classification, and clinical utility) and prognostic associations (mortality and HF hospitalization) between an updated AI HFpEF model (EchoGo Heart Failure v2) and existing clinical scores (H2FPEF and HFA-PEFF). The AI HFpEF model and H2FPEF score demonstrate similar discrimination and calibration, but classification is higher with AI than H2FPEF and HFA-PEFF, attributable to fewer intermediate scores, due to discordant multivariable inputs.
3. Ambulatory atrial fibrillation detection and quantification by wristworn AI device compared to standard holter monitoring.
In a 150-participant study, a wrist-worn PPG+ECG AI device achieved ≥95% sensitivity and ≥98% specificity for AF detection and 99% correlation for AF burden versus 24-h Holter. Extending monitoring to 28 days nearly doubled AF detection (14.7% to 26.7%), highlighting the value of comfortable, longer-term wearable monitoring.
Impact: Demonstrates high diagnostic accuracy and substantial yield gain with prolonged monitoring using an accessible wearable, with direct implications for AF screening and secondary prevention.
Clinical Implications: Wrist-worn AI-enabled devices can augment or extend Holter monitoring to improve AF detection and burden quantification, informing anticoagulation and rhythm management decisions.
Key Findings
- Sensitivity ≥95% and specificity ≥98% for AF detection compared to 24-hour Holter.
- AF burden showed 99% correlation between the wearable and Holter.
- AF prevalence increased from 14.7% to 26.7% with 28-day wearable monitoring.
Methodological Strengths
- Head-to-head performance versus gold-standard Holter
- Extended 28-day monitoring to assess real-world detection yield
Limitations
- Single-device study with modest sample size
- Clinical outcomes not assessed; potential selection bias
Future Directions: Assess clinical outcomes and cost-effectiveness of wearable-guided AF detection and management; validate across diverse populations and device ecosystems.
Timely detection of atrial fibrillation (AF) is crucial for the prevention of serious consequences such as stroke and heart failure, yet it remains challenging due to its often asymptomatic or paroxysmal nature. Wearable devices with artificial intelligence algorithms offer promising solutions. AF detection by the CardioWatch 287-2 (CW2), a wrist-worn photoplethysmography (PPG) and single-lead ECG device, was compared to 24-h Holter. Patient compliance, AF prevalence and AF burden were evaluated for 27 additional days. Data from 150 participants (mean age 64 ± 12 SD; 41% female) were analysed. The CW2's PPG and single-lead ECG algorithms achieved a specificity ≥98% and sensitivity ≥95% for AF detection, and 99% correlation for AF burden, compared to 24-h Holter. AF prevalence increased from 14.7% (24-h Holter) to 26.7% (28-day CW2).