Carotid artery stiffness mechanisms, heart failure events, and atrial fibrillation in MESA: the Multi-Ethnic Study of Atherosclerosis

Abstract

Arterial stiffness can be separated into two main mechanisms: 1) load-dependent stiffening from higher blood pressure and 2) structural stiffening due to remodeling of the vessel wall. The relationship of stiffness mechanisms with heart failure (HF) and atrial fibrillation (AF) is unknown. MESA (multi-ethnic study of atherosclerosis) participants with baseline carotid ultrasound images were included in this study (HF n = 6,278; AF n = 5,292). Carotid pulse wave velocity (cPWV) was calculated from B-mode carotid ultrasound to represent total stiffness. Structural stiffness was calculated by adjusting cPWV to a 120/80 mmHg blood pressure with participant-specific models. Load-dependent stiffness was the difference between total and structural stiffness. Associations with incident heart failure events and atrial fibrillation diagnosis were assessed with adjusted Cox hazard models. Four hundred-seven HF events and 1,157 AF diagnoses occurred during a median 17.7 and 16.8 years of follow-up. The associations of carotid artery stiffness mechanisms with HF events were: total cPWV adjusted HR per 1 SD 1.09 [0.98-1.22], P = 0.11; structural cPWV adjusted HR 1.06 [0.94-1.18], P = 0.33; and load-dependent PWV adjusted HR 1.23 [1.05-1.44] per 1 m/s, P = 0.009. The associations of carotid artery stiffness mechanisms with AF diagnoses were: total cPWV adjusted HR 1.11 (1.04-1.20), P = 0.004; structural cPWV adjusted HR 1.10 [1.02-1.16], P = 0.017; load-dependent cPWV adjusted HR 1.12 [1.02-1.23], P = 0.020. Both structural and load-dependent cPWV were associated with the development of AF, and load-dependent cPWV was associated with HF events. These findings indicate that load-dependent cPWV may be a potential treatment target to reduce the incidence of both HF and AF.NEW & NOTEWORTHY We evaluated associations between novel components of arterial stiffness: 1) load-dependent stiffening from higher blood pressure and 2) structural stiffening due to remodeling of the vessel wall and their associations with incident heart failure (n = 6,278) and atrial fibrillation (n = 5,292) over ∼17 years of follow-up. We found that both baseline structural and load-dependent stiffness were associated with the development of atrial fibrillation and load-dependent stiffness was associated with heart failure events.

Keywords: arterial stiffness; atrial fibrillation; epidemiology; heart failure.

Figure 1:

Graphical representation of methods used to differentiate structural and load-dependent stiffness. A, Total stiffness is calculated from measured pressure and diameter. B, A participant-specific model of arterial mechanics is fitted. C, Structural stiffness is calculated from pressure and diameter adjusted to 120/80 mm Hg. D, Load-dependent stiffness is calculated as the difference of total and structural stiffness. Representative results are shown for 2 participants (one in red, one in black) who had similar total carotid pulse wave velocity (cPWV) (6.4 vs 6.3 m/s) but via different mechanisms. One participant (red lines and text) had higher structural cPWV (7.6 vs 6.0 m/s), whereas the other participant (black lines and font) had higher load-dependent cPWV (0.4 vs −1.3 m/s). This figure is reproduced from (10).

Figure 2:

Unadjusted Kaplan-Meier curves showing event free survival curves for participants with low structural and low load-dependent stiffness (low-both), low structural and high load-dependent stiffness, high structural and low load-dependent stiffness, and high structural and high load-dependent stiffness (high both). 75^th percentile values were used to separate participants with low and high stiffness values. A. Heart failure events, and B. Atrial fibrillation diagnosis. Note that the Kaplan-Meier curves are not adjusted for any co-variates.

Figure 3:

Hazard ratios (HRs) for the association of carotid artery stiffness parameters with all heart failure events and atrial fibrillation diagnosis. HRs calculated from Cox hazard models adjusted for: age, sex, race, study site, education, BMI, diabetes, hypertension, any antihypertensive medication use, RAAS inhibitor use, total cholesterol, HDL cholesterol, lipid lower medication use, smoking status, mean arterial pressure, and pulse pressure. PWV – pulse wave velocity, Struct. – structural, Load-Dep. – load-dependent.