Introduction
Atrial fibrillation (AF) is the most common cardiac arrhythmia, with a prevalence of 2% in a
general Western population, rising to 22% in patients with heart failure (HF). AF can occur in
paroxysms or be present continuously. HF with AF is associated with an increased risk of
stroke. Strokes due to AF that occur in HF patients are more severe than those that occur in
patients without HF. Oral anticoagulation therapy (OAT) reduces the risk of stroke in patients
with AF, although with an increased risk of major bleeding. While the relationship between
persistent AF and stroke events in HF patients is well described, there is less clarity about the
stroke risk from paroxysmal forms of AF. Paroxysmal AF is commonly described with reference
to an ‘AF burden’, but how large a burden of paroxysmal AF warrants therapy with OAT is not
known. Remote monitoring of Cardiac Implantable Electronic Devices (CIEDs) can provide a
continuous assessment of cardiac rhythm. This offers a new method of describing the burden of
paroxysmal AF that occurs in patients with HF.
Methods
Remote monitoring data downloads from patients with CIEDs and HF were collected from
patients at 2 UK hospitals.
169 patients were analysed in the first centre. Both AF episode data and clinical data were
collected. The data from this centre was used to determine the feasibility of the technique, and
the burden of AF was described. AF episodes were then collated into ‘AF days’ and the incidence
of significant AF days was described with reference to 5 previously used definitions of a
significant AF day in previously published work. 327 patients were analysed from a second UK
centre with data collected in order to validate the incidence of AF observed in the first centre.
Results
In patients from Centre 1, 3,426 episode of AF were detected during remote follow-up in 115
patients over 446 patient years. Paroxysmal AF was detected in 63% of patients who were
presumed to have sinus rhythm at the time of implant. The median total AF burden in this group
was small (187 seconds; 16 seconds – 7.2 days), but 32% of patients with pAF had a total
burden of greater than 1 day. The burden of AF was described in several ways, and exploratory
12
analyses of AF burden was also carried out. The burden of AF was compared to that in the 5
previously published studies of remotely detected AF. One definition (‘≥ 30 seconds of AF in a 24 hour period’) was found to be met by all patients who then also met other definitions, and
patients appeared to fulfil this definition of a clinically significant burden of AF sooner. In centre
2, 7,148 episodes of AF occurred in 236 patients over a follow period of 805 patient years. The
burden of AF appeared broadly similar to that observed in the first centre, and the comparison
with the previously published data was also similar.
Conclusions
AF occurs frequently in patients with HF in whom dual chamber CIEDs have been implanted.
Many patients spend only a short amount of time in AF, and the clinical significance of this is
unclear. A definition of ‘at least 30 seconds in a 24 hour period’ was met by all patients who met
other published definitions of a significant burden of AF and was met sooner than the other
definitions of significance. Further work is needed to determine if this burden of AF could
predict longer burdens and to correlate pAF burden with stroke risk.