Abstract 3163: Central Pulse Pressure As A Robust Predictor Of First Atrial Fibrillation: Study Of Atrial Fibrillation In High Risk Elderly (SAFFIHRE)
Background: Data regarding the relationship between arterial stiffness and atrial fibrillation (AF) is sparse. We compared the utility of peripheral pulse pressure (P-PP) versus central pulse pressure (C-PP) in prediction of first AF.
Methods: SAFFIHRE is an ongoing propsective NIH-funded study of the pathophysiology of AF in persons aged ≥65 years with ≥ 2 other risk factors for AF (hypertension, coronary artery disease (CAD), heart failure (HF), diabetes (DM)), and without any history of AF or stroke at baseline. Clinical, echocardiographic, electrocardiographic (ECG), radial tonometry and other assessments are performed annually. Baseline C-PP was assessed by radial tonometry. Cox regression models were used for assessment of C-PP and P-PP in prediction of first AF.
Results: Of 800 enrolled participants (66% men), mean age 73.6±5.6 yrs, ECG-verified AF developed in 30 subjects (4%) after a mean of 1.5±1.10 years follow-up. The prevalence of hypertension (97%), DM (54%), and CAD (63%) in this cohort reflects their high risk status. Incidence rates of AF at year 3 were 4.3%, 11.3%, and 13.4% for tertiles of P-PP, and 3.8%, 12.1%, and 13.4% for C-PP (P for trend, <0.0001 for both). In age and sex adjusted models, both P-PP and C-PP were each highly predictive of AF (P<0.0001). In separate models also adjusting for body mass index, smoking, hypertension, DM, CAD, HF, systolic blood pressure (SBP), indexed LA volume (LAVi), LV ejection fraction (LVEF), and indexed LV mass, C-PP remained predictive of first AF (P<0.0001) (Table⇓), but not P-PP (P=0.37). The global model chi square and C-statistic were both greater for the model containing C-PP as compared to that containing P-PP (model chi square, 47.40 versus 37.37; C statistic, 0.81 versus 0.78).
Conclusion: In this high risk elderly population, C-PP was stronger than P-PP as an independent predictor of AF. It dominated the model for prediction of first AF, eclipsing well-established clinical and echocardiographic risk factors.