Arrhythmias & electrophysiologyDiagnostic or prognostic study
Comprehensive Evaluation of Rhythm Monitoring Strategies in Screening for Atrial Fibrillation
- Diagnostic accuracy study
- simulated screening strategies vs implantable loop recorder
Authors' conclusion
In screening for AF among participants with stroke risk factors, the diagnostic yield increased with duration, dispersion, and number of screenings, although all strategies had low yield compared with the implantable loop recorder. The sensitivity was higher among participants who were older, were male, or had higher NT-proBNP.
Study context
- Population
- Individuals with stroke risk factors but without atrial fibrillation (AF), recruited from the general population
- Participants
- 590 individuals
- Comparison
- Simulated AF screening strategies vs implantable loop recorder
- Outcome
- Sensitivity and NPV for AF detection (new-onset AF ≥6 minutes)
Performance
As reported in the abstract
Sensitivity (NPV) for AF detection vs implantable loop recorder
| Test or model | Sensitivity | NPV |
|---|---|---|
| Single 10-second ECG | 1.5% | 66% |
| Twice-daily 30-second ECGs during 14 days | 8.3% | 67% |
| 24-hour monitoring | 11% | 68% |
| 48-hour monitoring | 13% | 68% |
| 72-hour monitoring | 15% | 69% |
| 7-day monitoring | 21% | 70% |
| 30-day monitoring | 34% | 74% |
Further findings
AF detection with repeated or dispersed screening
- AF detection improved with subsequent screenings or same duration spread over several periods vs a single period (eg, three 24-hour vs one 72-hour monitoring)
- P < 0.0001
- P<0.0001 for all comparisons
Sensitivity by subgroup
- Sensitivity consistently higher with age ≥75 years, male sex, CHADS2 score >2, NT-proBNP ≥40 pmol/L, and underlying ≥24-hour AF episodes vs shorter AF
- P < 0.0001
- P<0.0001 for all screening strategies
Original publication
Comprehensive Evaluation of Rhythm Monitoring Strategies in Screening for Atrial Fibrillation. Circulation. Published 2 March 2020. doi:10.1161/circulationaha.119.044407
Summarised from the published abstract; every number is checked against it