Pilot Evaluation of a Novel Artificial Intelligence (AI) Heart Rate Variability (HRV)-Guided Risk Stratification for Chest Pain in the Emergency Department: A Randomized Controlled Trial

Background: Heart rate variability (HRV) analysis powered by artificial intelligence (AI) offers a rapid, non-invasive, and objective approach for acute coronary syndrome (ACS) risk stratification in the emergency department (ED). The objective of this study was to evaluate the feasibility and impact of aiTriage™, an AI HRV-guided tool for chest pain triage, compared with standard care. Methods: In this single-blinded randomized controlled trial, 560 ED patients with suspected ACS underwent 5- minute ECG monitoring for HRV analysis, which generated a 0-100 risk score and triage recommendations (high, medium, or low risk). Patients were randomized to standard care (control) or an HRV-guided protocol (intervention). Physicians in the control group were blinded to HRV results. Results: Of 426 analysed patients (mean age 54 ± 13 years, 35% female, 16.2% prior MI), the HRV-guided protocol reduced hospital admissions (50.2% vs 61.1%; risk difference -10.9 percentage points, 95% CI: -20.1 to -1.8) and serial cardiac enzyme testing (32.1% vs 41.7%; risk difference -9.6 percentage points, 95% CI: -18.4 to -0.9) compared with standard care. Among discharged patients, the median ED length of stay was 20 minutes shorter in the intervention group (95% CI: -45 minutes to 3 minutes). The overall 30-day MACE rate was 9.5%, with no events among discharged patients. Conclusion: A rapid AI HRV-guided risk stratification tool was feasible to deploy, and has potential to reduce serial cardiac enzyme testing, ED LOS and hospital admissions. An adequately powered RCT is needed to confirm these findings and assess clinical safety. This trial is registered at ClinicalTrials.gov (NCT07074808).