Long term cardiac monitoring with implanted device frequently reveals previously undiagnosed AF
Incidence of Previously Undiagnosed Atrial Fibrillation Using Insertable Cardiac Monitors in a High-Risk Population - The REVEAL AF studyLiterature - Reiffel JA, Verma A, Kowey PR, et al. - JAMA Cardiology 2017; 2(10):1120-27
- 326 patients had an ICM inserted and reached the 18-month follow-up timepoint; they were 71.5 (SD 9.9) years old on average and 77.1% were 65 years or older.
- 128 participants had AF of 6 min or more, with an AF detection rate of 29.3% (95%CI: 24.4-33.8) at 18 months.
- AF detection rates at 30 days and 6, 12, 24, and 30 months were 6.2% (95%CI: 3.8-8.6), 20.4% (95%CI: 16.2-24.3), 27.1% (95%CI: 22.5-31.5), 33.6% (95%CI: 28.3-38.6%), and 40.0% (95% CI: 33.6-45.8%), respectively.
- Of the 128 participants with AF of at least 6 min, 113 (88.3%) had AF of 30 min or more, 97 (75.8%) had AF or 1 hour or more, and 53 (41.4%) had AF of 6 hours or more in one day. 13 patients (10.2%) had at least 1 episode of 24 hours or longer.
- The median time (IQ range) from ICM insertion to first AF detection was 123 (41-330) days.
- The AF detection rate at 18 months was higher in patients with palpitations at baseline compared to those without (35.3%, 95%CI: 29.0-42.6 vs 23.0%, 95%CI: 17.5-29.8, P=0.02).
- No significant difference in AF incidence between CHADS2 scores of 2, 3 or 4 or more was detected at 18 months.
- Among patients who met the primary outcome of AF of 6 min or more, 72 (56.3%) were prescribed OAC therapy and 19 (14.8%) were prescribed rhythm control.
The incidence of AF of 6 min or more in previously undiagnosed AF, detected by ICM, was 29.3% at 18 months and 40.0% at 30 months in a high-risk population. When using external devices, monitoring is typically shorter. Considering the median time to AF in this study, most patients would not have been identified with those methods. AF incidence was still rising at 30 months, thus the ideal monitoring duration is unclear. These results may impact AF screening and stroke prevention in this population.
In his commentary , Healey emphasizes that AF is not only the most common and potent risk factor for stroke, but more importantly, it is the risk factor that, when treated, can prevent stroke to the greatest extent. He discusses the incidence of AF in light of other trial results and although inclusion criteria are different, they all suggest that undetected AF is extremely common in a high-risk patient population. The ASSERT-II study found that an increasing age was associated with an increased rate of AF detection. Since the age in the REVEAL AF and ASSERT-II studies was higher (mean age was ≥10 years higher) compared to the CRYSTAL AF study, this might explain the higher rate of AF detection in the first 2 studies. The ASSERT-II study enrolled patients with evidence of left atrial enlargement and found that increasing left atrial size was associated with an increased AF risk, which might explain the higher incidence of AF in this study compared to that in the REVEAL AF trial. Finally, he mentions that the REVEAL AF study was unique in enrolling patients with a variety of nondiagnostic factors. Although Healy congratulates the authors on the completion of this large, carefully designed study, he raises the point that the study does not give us insights in the stroke risk and the effectiveness of anticoagulation therapy, due to the sample size and the number of patients with detected AF who used anticoagulation (56%). He concludes that ‘Over the next 3 to 4 years, subgroup analyses, economic evaluations, and randomized clinical trials will help determine if this insight can be translated into a cost-effective stroke prevention strategy for high-risk individuals.’