TrialsgrambetaRCTs are beautiful

ARISE (AI-ECG)

2024

Artificial Intelligence-Powered Rapid Identification of ST-Elevation Myocardial Infarction via Electrocardiogram (ARISE)

In 43,234 patients presenting to the emergency department or admitted to hospital, AI-ECG interpretation with an automated STEMI alert to physicians reduced door-to-balloon time for STEMI compared with standard care.

Result

Primary endpoint: Door-to-balloon time in emergency-department patients (primary endpoint not stated in the available summary)

Median door-to-balloon time for ED patients 82.0 vs 96.0 min (P<0.001); cardiac death 85 vs 116 (OR 0.73; P=0.029). The AHA 2023 presentation reported 43,994 randomised, 124 confirmed STEMI and ECG-to-catheterisation-laboratory time 43.3 vs 52.3 min (P=0.003).

Original publication
Subtopic
Acute diagnosis & risk alerts
Design
Pragmatic single-centre RCT, Taiwan
Relevance
Important: one of several pillars.
Source
Lin C, et al. NEJM AI. 2024;1(7):AIoa2400190.doi 10.1056/AIoa2400190

Discussion & critique

The first RCT of an AI STEMI alert, showing that automatic flagging of the ECG can shorten the time to reperfusion in an unselected emergency and inpatient population. The numbers here stem from a published summary rather than the primary abstract, and the conference presentation reported different figures (43,994 randomised and different time metrics), so all ARISE numbers need verification against the primary publication. Single-centre and open-label, with confirmed STEMI in only a small fraction of those randomised.