HRS 2017: True incidence of atrial fibrillation is underestimated

James Reiffel

A substantial number of patients at risk for atrial fibrillation (AF) may remain undetected using conventional monitoring techniques, a late-breaking clinical trial session at the Heart Rhythm Society’s 38th Annual Scientific Sessions (10‒13 May, Chicago, USA) heard. Investigators for REVEAL AF, which aimed to determine the incidence of previously undiagnosed AF in a high-risk population using an insertable cardiac monitor (ICM), said further studies are needed.

James Reiffel, from Columbia University in New York, USA, said: “The average population chosen in this trial represents a very common group of patients encountered in clinical practice. Our results may have important public health implications regarding AF screening and treatment in high-risk patients, but we recognize further trials are warranted. Two prior trials, both smaller and shorter had similar findings. When you pool them together, it is about 900 patients. I do not think this is the result of one small random study.”

He outlined the rationale behind the prospective single-arm open label multicentre trial: “Stroke may be the first manifestation of AF—about 20% of patients with AF-related strokes have no prior history of AF, and that is a big problem. Strokes from AF are larger, associated with higher disability, greater mortality, and greater costs, and importantly, the formation of an atrial thromboembolism does not depend upon the presence of other AF symptoms. We should be just as worried about silent AF as we are about symptomatic AF.”

Patients with a CHADS2 score ≥3 (or 2, with an additional AF risk factor) were included. Patients with AF detected on ≥24 hours of screening by external monitoring before ICM insertion were excluded. The primary endpoint was the detection rate of AF lasting at least six seconds at 18 months, with secondary detection rates from 30 days to 30 months, and among CHADS2 subgroups.

In 385 patients (mean age 71.5 years, 52.2% male) followed for a mean of 22.5 months, the AF detection rate at 18 months was 29.3%, with rates at 30 days, six, 12, 24, and 30 months of 6.2%, 20.4%, 27.1%, 33.6%, and 40%, respectively. These patients received continuous monitoring with either the Reveal XT ICM or the Reveal LINQ ICM (predominately the latter) (both from Medtronic). At 18 months, AF incidence was similar among patients with CHADS2 scores of 2 (24.7%), 3 (32.8%), and ≥4 (31.7%) (p=0.23). The median time from device insertion to detection of the first AF episode was 123 (IQR 41-330) days. Oral anticoagulants were prescribed to 56.3% of patients.

Additionally, two sensitivity analyses were performed, one assuming that all prematurely exited patients had AF (18-month incidence rate 41.1%) and one assuming that none had AF (28.2%).

Reiffel said: “Given our observed rate was 29.3%, it is possible, in fact I think likely, that the true incidence would have been higher had all of the patients continued in the trial…But more importantly, over ¾ of the patients with AF detected would have been missed with more conventional monitoring approaches, which use 24 hours to 30 days.”