Frequency of Atrial Fibrillation in VITAL-AF (2024)

Atrial fibrillation (AF) is a common and morbid arrhythmia. Ascertaining AF frequency is challenging because AF may be paroxysmal and diagnostic codes can be inaccurate. Rigorous contemporary estimates of AF prevalence and incidence are needed to inform interventions to reduce the public health burden of AF. We sought to quantify clinically diagnosed AF prevalence and incidence among control patients in VITAL-AF (Screening for Atrial Fibrillation Among Older Patients in Primary Care Clinics),¹ a cluster-randomized trial of AF screening among individuals 65 years or older within 16 Massachusetts General Hospital primary care practices.

This cohort study is a secondary analysis of the VITAL-AF trial and included 17 238 eligible patients from 8 control practices, excluding 2.4% who were screened, who were enrolled from July 31, 2018, to October 8, 2019. Atrial fibrillation was ascertained in a staged manner facilitated by the electronic health record (EHR). First, AF events occurring before study entry and meeting a high-specificity algorithm for prevalent AF (positive predictive value, 98.4%^1,2) were classified as prevalent. Second, additional candidate AF events (including all incident events) were identified using 1 or more International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10) codes for AF or atrial flutter or 12-lead electrocardiogram with AF or atrial flutter diagnosis and manually adjudicated by 2 research nurses with consensus resolution of discrepant adjudications and cardiologist resolution of unresolved discrepancies. We assessed AF prevalence at study entry and AF incidence rates (per 1000 person-years [PY]) over 1-year follow-up, stratified by demographic characteristics and comorbidities. Race was classified using self-reported categories and included as a stratification variable given prior associations with diagnosed AF.³ Informed consent was provided by participants prior to screening in the original trial. Protocols were approved by the Mass General Brigham Institutional Review Board. Further methods are reported in the eMethods in Supplement 1.

A total of 17 238 individuals (mean [SD] age, 74.6 [7.1] years; 9659 [56.0%] female; 14 376 [83.4%] White) were enrolled; 2278 (13.2%) had prevalent AF. Among 14 960 individuals without prevalent AF, the AF incidence rate was 23.7 per 1000 PY (95% CI, 21.0-26.7). Both AF prevalence (6.4% at 65-69 years to 28.5% at ≥85 years) and incidence (14.2 per 1000 PY at 65-69 years to 50.8 per 1000 PY at ≥85 years) increased with age. Incidence was greater in men (33.4 per 1000 PY) vs women (16.8 per 1000 PY). Prevalence and incidence are given in Table 1.

This cohort study provides contemporary estimates of clinically diagnosed AF prevalence and incidence in primary care practices of a large academic medical center. New AF incidence was substantial, and nearly one-third of individuals 85 years or older had prevalent AF. In contrast to previous estimates,^3-5 our AF diagnoses were facilitated by linkage to complete EHR data, then reviewed manually for confirmation.

Previous AF frequency estimates vary substantially based on population and AF ascertainment method^3,4 (Table 2). Prospective research cohorts have generally applied stringent AF definitions but are relatively small and noncontemporary.⁴ In contrast, claims-based studies³ rely on billing and diagnosis codes, which may be less accurate. We leverage a unique resource of manually adjudicated AF diagnoses within a large contemporary primary care population.

We report higher AF prevalence compared with community-based cohort studies^4,5 or claims-based data.³ We estimate higher AF incidence compared with research cohorts but lower AF incidence compared with claims-based estimates. Observed differences may relate to greater comorbidity predisposing individuals to AF or to contrasting ascertainment methods, where our staged approach using EHRs and manual adjudication offers greater sensitivity for past AF diagnoses while retaining high specificity for incident AF. Overall, our findings suggest higher AF rates in primary care populations vs community-based research cohorts. A limitation is that our sample may possess characteristics affecting generalizability, although more than 70% of US individuals 65 years or older receive regular primary care.⁶ This cohort study’s findings may prove useful to studies and programs focusing on AF prevention, diagnosis, and management.

Accepted for Publication: December 22, 2022.

Published: February 13, 2023. doi:10.1001/jamanetworkopen.2022.55838

Open Access: This is an open access article distributed under the terms of the CC-BY-NC-ND License. © 2023 Khurshid S et al. JAMA Network Open.

Corresponding Author: Steven A. Lubitz, MD, MPH, Demoulas Center for Cardiac Arrhythmias, Massachusetts General Hospital, 55 Fruit St, Boston, MA 02114 (slubitz@mgh.harvard.edu).

Author Contributions: Dr Lubitz had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Drs Khurshid and Ashburner served as co–first authors, each with equal contribution to the manuscript.

Concept and design: All authors.

Acquisition, analysis, or interpretation of data: Ashburner, Atlas, Singer, Lubitz.

Drafting of the manuscript: Khurshid, Ashburner.

Critical revision of the manuscript for important intellectual content: Ashburner, Ellinor, McManus, Atlas, Singer, Lubitz.

Obtained funding: McManus, Lubitz.

Administrative, technical, or material support: Khurshid, Ashburner, Ellinor, McManus.

Supervision: Ashburner, McManus, Atlas, Singer, Lubitz.

Conflict of Interest Disclosures: Dr Ellinor reported receiving personal fees from Bayer AG and Novartis and grants from IBM and Bayer AG during the conduct of the study. Dr McManus reported receiving personal fees from Fitbit Inc during the conduct of the study and serving on the scientific boards of Bristol Myers Squibb and Pfizer, serving on the data safety and monitoring boards of Avania and Boston Biomedical, serving as a reviewer for Venturewell, serving as an editor for the Heart Rhythm Society, receiving grants from Boehringer Ingelheim and the National Institutes of Health, and receiving nonfinancial support from Apple Computers. Dr Atlas reported receiving grants and personal fees from Bristol Myers Squibb, personal fees from Pfizer and Fitbit Inc, and grants from the American Heart Association outside the submitted work. Dr Singer reported receiving personal fees from Bristol Myers Squibb, Fitbit Inc, Medtronic, and Pfizer outside the submitted work. Dr Lubitz reported receiving personal fees from Bristol Myers Squibb, Pfizer, Invitae, and Blackstone Life Sciences, receiving grants from Boehringer Ingelheim, Fitbit Inc, Medtronic, Premier, and IBM, and being employed by the Novartis Institutes for BioMedical Research Employment as of July 18, 2022, outside the submitted work. No other disclosures were reported.

Funding/Support: This investigator-initiated study was funded by the Bristol Myers Squibb/Pfizer Alliance. Dr Lubitz was supported by grant 1R01HL139731 from the National Institutes of Health (NIH) and grant 18SFRN34250007 from the American Heart Association (AHA) during this study. Dr Atlas is supported by grant 18SFRN34250007 from the AMA. Dr Singer was supported by the Eliot B. and Edith C. Shoolman fund of the Massachusetts General Hospital. Dr Ashburner is supported by grant K01HL148506 from the NIH and grant 18SFRN34250007 from the AHA. Dr McManus is supported by grants R01HL137734, R01HL137794, R01HL141434, R01HL36660, R21AG060529, U54HL143541, and U01HL14682 from the NIH. Dr Ellinor is supported by grants 2R01HL092577 and K24HL105780 from the NIH and grant 18SFRN34110082 from the AMA.

Role of the Funder/Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation or approval of the manuscript; and decision to submit the manuscript for publication. The funding source reviewed the manuscript before submission.

Data Sharing Statement: See Supplement 2.