Predicting survival in atrial fibrillation: results from SAGE-AF.
10.26599/1671-5411.2025.03.004
- Author:
David C PARISH
1
;
Catarina I KIEFE
2
;
Jordy MEHAWEJ
3
;
Edith Mensah OTABIL
3
;
Carly N BENIEK
3
;
Francis C DANE
4
Author Information
1. Department of Medicine, Mercer University School of Medicine, Macon, GA, USA.
2. Department of Population and Quantitative Health Sciences, University of Massachusetts Chan Medical School, Worcester, MA, USA.
3. Department of Medicine, University of Massachusetts Chan Medical School, Worcester, MA, USA.
4. Department of Psychology, Radford University, Roanoke, VA, USA.
- Publication Type:Journal Article
- From:
Journal of Geriatric Cardiology
2025;22(3):344-350
- CountryChina
- Language:English
-
Abstract:
BACKGROUND:Using Systematic Assessment of Geriatric Elements in Atrial Fibrillation (SAGE-AF) data, determine how well the rich mix of demographic, clinical history, geriatric assessments, and clinically adjudicated events can predict two-year survival.
METHODS:Subjects were recruited from participating outpatient practices if they had non-valvular AF, were 65 or over with CHA2DS2-VASc scores of at least 2, and were candidates for anticoagulation. Demographics, clinical history, and geriatric qualities of life were assessed by interview and medical records review using standardized protocols and repeated at one and two years. Events identified were abstracted and submitted for adjudication using standard definitions of events and categories. Non-mortality event categories included hospitalizations (cardiovascular, bleeding, other), bleeding (major, clinically relevant non-major, minor), and seven major adverse cardiovascular events.
RESULTS:The 1245 subjects experienced 1960 events, primarily hospitalizations (935) and/or bleeding (817); 114 subjects (9.2%) died during two years of follow-up. Events initially abstracted to more than one category (172) were combined, resulting in 1788 unique incidents. Most subjects had zero or one event (69%) and fewer than 7% had more than 3 types. Most variables were significant in bivariate analysis. Using multiple logistic regression with two-year survival as the outcome variable, the best-fit model included event number and type, number of unique incidents, and number of bleeding events (R2 = 0.511, C = 93.1) with sensitivity = 97.9% and specificity = 44.7%.
CONCLUSIONS:Two-year survival was high. This model, if validated, could have major implications for treatment of patients with AF. Patients in the large group with no or one event are at very low risk of death (under 2%). The small group with high risk for further complications, including death, deserve reassessment to determine if this trajectory can be altered.
