RAPID-AF: a novel method to detect paroxysmal atrial fibrillation in primary care and post-stroke patients with no history of AF.

Atrial fibrillation (AF) increases the risk of stroke more than 5 times. With an increasing incidence with age, AF is recognized as a major public health burden in the Western population. Since stroke risk can be effectively reduced by medication once AF has been diagnosed, early detection of the arrhythmia is warranted. Detection of AF, however, is complicated by its often paroxysmal (pAF) and asymptomatic nature. Effective pAF detection therefore requires multiple-day continuous heart rhythm (Holter) monitoring in high risk groups. This is a major burden on patients.

Previous studies with the Stroke Risk Analysis (SRA) algorithm, which assesses pAF risk from electrocardiographic anomalies in short (1-24h) snippets of Holter data, have shown that the algorithm is able to predict with reasonable sensitivity and specificity the outcome of multiple-day Holter monitoring. The RAPID-AF study will be the first study to perform a large-scale validation of the SRA algorithm in both primary care and post-stroke patients.

When sufficiently validated, the SRA algorithm could serve as a triage test for further Holter monitoring. The SRA algorithm thus has potential to reduce screening burden to several hours in those patients who score ‘low risk of pAF’, where they would otherwise have had to undergo a complete 14-day Holter.

Stroke Risk Analysis (SRA) (Apoplex Medical, Pirmasens, Germany) is a Holter based algorithm that assesses the risk of underlying paroxysmal atrial fibrillation (pAF) from electrocardiographic anomalies in 1-24h snippets of Holter data while the patient is still in sinus rhythm, with two weeks continuous Holter monitoring as reference standard.

We expect that SRA has both sufficient negative predictive power (sensitivity) to select patients who don’t need any further evaluation for AF and sufficient positive predictive power (specificity) to select patients who are eligible for further long term ECG monitoring.

- Day 0: start of continuous Holter monitoring;

- Day 14: end of Holter monitoring; assessment of diagnosis 'pAF yes or no' from Holter monitoring and relaying outcome to treating physician (neurologist in AMC cohort, general practitioner in D2AF cohort);

- Day 14+: performing SRA algorithm on snippets of first 1, 2, 6, 12 & 24h of 14-day recording (blinded to the final outcome of the 14-day recording, i.e. the reference standard); assessment of patient satisfaction through questionnaire.

- All included patients undergo 14-day continuous Holter monitoring, either as standard of care after ischemic stroke or TIA (post-stroke cohort) or as per inclusion into the intervention arm of the D2AF study (primary care cohort).

- We perform the SRA algorithm on snippets of the first 1, 2, 6, 12 and 24 hours of the 14-day continuous Holter monitoring. For each snippet, the algorithm provides an outcome expressed as either of 3 categories: (1) sinus rhythm (SR) with low risk of pAF; (2) SR with high risk of pAF, and; (3) present AF.

- We compare the outcome of the SRA algorithm with the reference standard, i.e. the final outcome of conventional 14-day continuous Holter monitoring.

- We assess patient satisfaction with undergoing 14-day Holter by means of a questionnaire.