[Skip to Content]
Access to paid content on this site is currently suspended due to excessive activity being detected from your IP address 18.207.98.97. Please contact the publisher to request reinstatement.
[Skip to Content Landing]
Original Investigation
January 2018

Association of of Atrial Fibrillation Clinical Phenotypes With Treatment Patterns and Outcomes: A Multicenter Registry Study

Author Affiliations
  • 1Duke Clinical Research Institute, Duke University Medical Center, Durham, North Carolina
  • 2Section Editor, JAMA Cardiology
  • 3Harvard Medical School, Boston, Massachusetts
  • 4Massachusetts General Hospital, Boston
  • 5Department of Medicine, Yale University School of Medicine, New Haven, Connecticut
  • 6University of Colorado, Denver
  • 7Department of Medicine, University of California, Los Angeles
  • 8Department of Medicine, Mayo Clinic College of Medicine, Rochester, Minnesota
  • 9Lankenau Institute for Medical Research, Wynnewood, Philadelphia, Pennsylvania
  • 10College of Physicians and Surgeons, Columbia University, New York, New York
  • 11School of Medicine, Penn State University, Hershey, Pennsylvania
  • 12Department of Cardiovascular Research, St. Luke's Mid America Heart Institute, Kansas City, Missouri
  • 13University of Utah, Salt Lake City
JAMA Cardiol. 2018;3(1):54-63. doi:10.1001/jamacardio.2017.4665
Key Points

Question  How different are groupings derived from cluster analysis from traditional classifications of patients with atrial fibrillation?

Findings  This cluster analysis identified 4 statistically driven clusters: (1) a cluster with low comorbid burden; (2) a younger/behavioral disorder cluster; (3) a cluster who resembled patients with tachycardia-bradycardia and had implanted devices; and (4) a cluster defined by atherosclerotic comorbidities. These groupings are not driven by conventional classifications and have different risks for adverse clinical outcomes.

Meaning  Cluster analysis highlighted the considerable heterogeneity among patients with atrial fibrillation; cluster classification may identify differences in phenotypes, variations of clinical practice, and potential targets for improving the quality of care.

Abstract

Importance  Atrial fibrillation (AF) is usually classified on the basis of the disease subtype. However, this characterization does not capture the full heterogeneity of AF, and a data-driven cluster analysis reveals different possible classifications of patients.

Objective  To characterize patients with AF based on a cluster analysis and to evaluate the association between these phenotypes, treatment, and clinical outcomes.

Design, Setting, and Participants  This cluster analysis used data from an observational cohort that included 9749 patients with AF who had been admitted to 174 US sites participating in the Outcomes Registry for Better Informed Treatment of Atrial Fibrillation (ORBIT-AF) registry. Data analysis was completed from January 2017 to October 2017.

Exposure  Patients with diagnosed AF who were included in the registry.

Main Outcomes and Measures  Composite of major adverse cardiovascular or neurological events and major bleeding, as defined by the International Society of Thrombosis and Hemostasis criteria.

Results  Of 9749 total patients, 4150 (42.6%) were female; 8719 (89.4%) were white and 477 (4.9%) were African American. A cluster analysis was performed using 60 baseline clinical characteristics, and it classified patients with AF into 4 statistically driven clusters: (1) those with considerably lower rates of risk factors and comorbidities than all other clusters (n = 4673); (2) those with AF at younger ages and/or with comorbid behavioral disorders (n = 963); (3) those with AF who had similarities to patients with tachycardia-brachycardia and had device implantation owing to sinus node dysfunction (n = 1651); and (4) those with AF and prior coronary artery disease, myocardial infarction, and/or atherosclerotic comorbidities (n = 2462). Conventional classifications, such as AF subtype and left atrial size, did not drive cluster formation. Compared with the low comorbidity AF cluster, adjusted risks of major adverse cardiovascular or neurological events were significantly higher in the other 3 clusters (behavioral comorbidity cluster: hazard ratio [HR], 1.49; 95% CI, 1.10-2.00; device implantation cluster: HR, 1.39; 95% CI, 1.15-1.68; and atherosclerotic comorbidity cluster: HR, 1.59; 95% CI, 1.31-1.92). For major bleeding, adjusted risks were higher in the behavioral disorder comorbidity cluster (HR, 1.35; 95% CI, 1.05-1.73), those with device implantation (HR, 1.24; 95% CI, 1.05-1.47), and those with atherosclerotic comorbidities (HR, 1.13; 95% CI, 0.96-1.33) compared with the low comorbidity cluster. The same clusters were identified in an external validation in the ORBIT AF II registry.

Conclusions and Relevance  Cluster analysis identified 4 clinically relevant phenotypes of AF that each have distinct associations with clinical outcomes, underscoring the heterogeneity of AF and importance of comorbidities and substrates.

×