Aortic Pulse Wave Velocity: An Independent Marker of Cardiovascular Risk

Michel E. Safar, MD, Olivier Henry, MD, Sylvie Meaume, MD


Am J Geriatr Cardiol. 2002;11(5) 

In This Article

Aortic PWV as a Marker of CV Risk

In recent years, epidemiologic investigations have been performed in hypertensive subjects in the presence or absence of end-stage renal disease (ESRD).

In a cohort of 241 subjects with ESRD with an average follow-up of 11 years,[5] logistic regression and Cox analysis identified two dominant predictors of CV and all-cause mortality: PWV, and age at inclusion. Hemoglobin and low diastolic blood pressure were also predictive but to a much smaller extent. After adjustment for confounding variables including, in particular, age, left ventricular hypertrophy, and blood pressure, the odds ratio for PWV (>12.00 m/s vs. <9.4 m/s) was 5.4 (95% confidence interval (CI), 2.4-11.9) for all-cause mortality, and 5.9 (95% CI, 2.3-15.5) for CV mortality (Figure 2). These results provided the first evidence that, in patients with ESRD undergoing hemodialysis, increased aortic stiffness is a major and independent predictor of CV and all-cause mortality. Furthermore, in those patients with ESRD, similar results were obtained when carotid incremental elastic modulus was used in place of PWV as an index of the stiffness of carotid arterial wall material.[6]

Subjects with end-stage renal disease: probabilities of overall survival (A), and event-free survival cardiovascular (CV) mortality (B), in study population according to level of pulse wave velocity (PWV) divided into tertiles. Comparisons between survival curves were highly significant (c2=47.04 for CV mortality and 67.23 for overall mortality; p<0.0001 for both). Numbers in italics represent individuals at each point according to tertile of PWV. Reproduced with permission from Circulation. 1999;99:2434-2439.[5]

In subjects with essential hypertension, the situation is more complex because longitudinal studies on arterial stiffness as a CV risk factor are lacking. However, calculation of CV risk using Framingham equations[7] can partially resolve this problem. This was done in a study including 710 subjects with hypertension, in which the odds ratio of being in a high-risk group, based on the presence or absence of various CV risk factors, was assessed.[8] The risk of any CV complication consistently increased in parallel with the increase of the single measurement of PWV (Figure 3). Furthermore, at any given value of age, aortic PWV was the best theoretical predictor of CV mortality, and even more potent than age, blood pressure, and cardiac mass. The odds ratio of being in the group at high risk of CV mortality (>5% for 10 years) for patients with PWV >13.5 m/s was 7.1 (95% CI, 4.5-11.3). This study provided the first evidence that a single measurement of aortic PWV constitutes a potent indicator of CV risk in hypertensive patients. These findings are in agreement with recent longitudinal studies which showed that the ratio between stroke volume and pulse pressure, an indirect marker of arterial stiffness, and aortic PWV, a direct marker, were independent predictors of CV risk.[9,10]

Relationship between 10-year cardiovascular disease risk and aortic pulse wave velocity in subjects with essential hypertension. Reproduced with permission from Hypertension. 1999.33:1111-1117.[8]

It is important to note that, in hypertensive subjects with or without ESRD, PWV was a predictor of CV risk independent of blood pressure level and/or cardiac hypertrophy. In all these studies, the odds ratio was higher for PWV than for the two other parameters. When blood pressure and PWV were both involved in logistic regressions, it was constantly a low (and not a high) diastolic blood pressure that predicted CV risk.[8,11]

In summary, these findings suggest that increased aortic PWV is a strong independent predictor of CV risk, regardless of whether this mechanical factor plays a causative role in CV risk or merely serves as a marker of vascular disease already present. Clearly the former hypothesis needs to be tested using specific intervention studies, as recently published.[12]