Implications of brief episodes of atrial fibrillation (micro-AF)

Abstract

INTRODUCTION: Atrial fibrillation (AF) is the most common clinically relevant arrhythmia. It is diagnosed using an electrocardiogram (ECG) and defined as an irregular heart rhythm without p-waves lasting for at least 30 seconds. This 30-second time criterion is based solely on expert consensus. AF is associated with a five-fold increase in stroke risk but with use of stroke preventive oral anticoagulation treatment, the risk can be reduced by at least two thirds. Connections between supraventricular activity, AF and stroke have been identified in register studies. Still, there are no studies evaluating the risks associated with episodes of AF-like activity with duration shorter than 30 seconds, termed micro-AF. One can assume that micro-AF is associated with AF, but there are no guidelines for clinicians on how to investigate or treat individuals with these findings.

The aims of this thesis are: 1) to evaluate if individuals with micro-AF have a higher risk of subsequent AF and thus would benefit from extended screening and follow-up; 2) to study the prevalence of micro-AF in an elderly population; 3) to postulate the time span from micro-AF to AF; 4) to assess if micro-AF burden predict development of AF; 5) to identify which investigation is the most suitable to detect AF in an elderly, high-risk population by comparing intermittent ECG and continuous event-recording; 6) finally, the thesis aspires to gain a larger perspective by comparing the risks associated with micro-AF to the risk increase in individuals with other types of excessive supraventricular activity.

METHODS AND RESULTS: The participants in study I were identified from STROKESTOP I, a population-based mass-screening study for AF. In STROKESTOP I individuals aged 75- and 76-years recorded intermittent ECGs for 30 seconds, twice daily for two weeks. Participants free from AF but in whom micro-AF was detected were followed-up. In study I micro-AF was defined as abrupt onset episodes of irregular heart rhythm, ≥4 consecutive supraventricular beats and absence of p-waves lasting for less than 30 seconds. Participants free from both micro-AF and AF acted as a control group. After 2.3 years both groups were invited to repeat AF screening using intermittent ECG for 30 seconds twice daily in parallel with continuous event-recording for two weeks. AF was found in 50% (n=27/54) of participants in the micro-AF group compared to 10% (n=5/48) in the control group, p<0.001. One hundred percent of the AF cases detected during repeat screening were found by the continuous event-recorder and of those, intermittent ECG detected 40%. In study II, participants were identified from the STROKESTOP II study – similarly to STROKESTOP I, a mass-screening study for AF. In contrast only participants with N-terminal pro b-type natriuretic peptides (NT-proBNP) levels ≥125 ng/L were asked to perform intermittent ECG four times daily for two weeks. Participants free from AF but with micro-AF during intermittent recordings were invited to additional extended screening using continuous event-recording for two weeks. A tachycardia criterion was added to the micro-AF definition, and the duration was prolonged to ≥5 consecutive beats (the same definition was also used in study IV). A control group was screened using both ECG modalities simultaneously in STROKESTOP II. Continuous event-recording detected AF in 13% (n=26/196) in the micro-AF group and 3% (n=7/250) in the control group, p <0.001. The group performing intermittent ECGs and continuous event-recording in parallel during STROKESTOP II were also invited to participate in study III. The participants were asked to fill out a questionnaire comparing ease of use and compliance to the two ECG modalities; the intermittent ECG, a Zenicor II device and the continuous event-recording, an R-Test 4 device from Novacor. Continuous event-recording detected new AF in 6% (n=15/269) and intermittent ECG in 2% (n=5/269), p=0.002. Both devices were well tolerated, but intermittent ECG was graded “very easy to use” whereas continuous event-recording was graded “easy to use”, p<0.001. Study I included parts of the participants with micro-AF in STROKESTOP I. In study IV an automated algorithm was used to identify all participants with micro-AF and other supraventricular arrhythmias in the STROKESTOP I database. All ECGs identified by the algorithm were also manually assessed. Participants were followed up using three-year data from National Swedish Health registers. Supraventricular tachycardias were associated with an increased risk of AF compared to excessive supraventricular ectopic beats. Participants with supraventricular tachycardias with AF characteristics, micro-AF, n= 97 (1,6%), were shown to have a higher risk of AF than participants with other supraventricular arrhythmias (hazard ratio 4,3; 95% confidence interval 2,7-6,8). They also had an increased risk of death (hazard ratio 2,0; 95% confidence interval 1,1-3,8).

CONCLUSION: Extensive supraventricular ectopy, including frequent isolated supraventricular ectopic beats and supraventricular tachycardias, is common and associated with AF development in elderly people. Individuals with supraventricular tachycardias with AF characteristics – termed micro-AF – showed the highest risk of a future AF as well as an increased risk of death. Micro-AF also seems to be associated with an increased risk of already existing undetected AF. The risk of having AF detected by screening seems to increase with time in individuals with micro-AF, indicating a possibly progressive disease in the atria. We found no evidence that micro-AF burden affects AF development. Extended and repeat screening could be recommended for the elderly population with micro-AF as detection of AF in the majority of cases would lead to initiation of stroke preventive oral anticoagulant treatment. Intermittent ECG is well tolerated as a screening method in an elderly population, but continuous event-recording for two weeks detects three-times as many new cases of AF and is therefore a preferable screening method.