Cardiovascular disease has an extremely high incidence and prevalence worldwide, and accounts for a high percentage of overall mortality, particularly in western societies1.

Cardiovascular mortality is determined mainly by clinical conditions that involve the arterial system, and hence research into this system has become of prime importance. It is highly likely that a better understanding of the pathophysiology of arterial disease will have a significant positive impact in terms of public health.

Various methods have been developed to assess the arterial system using different approaches, but their usefulness in clinical practice is heavily dependent on operator- and technology-related variables.

Measurement of pulse wave velocity (PWV) is a rigorous method of assessing arterial stiffness2, with a solid experimental basis and with important contributions from physics3,4. From a clinical standpoint, several studies have shown that aortic PWV is an independent cardiovascular risk marker in various clinical contexts5–15. The 2007 joint guidelines for the management of arterial hypertension of the European Society of Hypertension (ESH) and the European Society of Cardiology (ESC) recognize this parameter as a marker of subclinical target organ damage16.

Despite the clinical importance of this measure of aortic stiffness, reference values for PWV have yet to be defined. The ESH/ESC guidelines suggest a cut-off of 12 m/s for aortic PWV to define vascular damage16, but this lacks an experimental foundation since only one prospective study in a highly selected population (patients with renal failure) suggests this value to define cardiovascular risk. At the same time, the nature of PWV is not conducive to an operational approach to defining normality, given that cardiovascular risk is continuous and that this parameter is highly age-dependent, as all studies have shown in this context, and so a given PWV value will have a different clinical significance depending on the age-group studied. We have therefore always considered that a statistical approach would be more suiTable for defining normal PWV values, with reference values adjusted for age and gender, an opinion recently adopted by an important research group in this area17.

The aim of the EDIVA project was to develop an approach to assessing global cardiovascular risk by quantifying individual risk using this methodology in a large population. It resulted in a prospective, multicenter, observational study involving 2200 individuals (of whom 668 were healthy), mean follow-up currently two years, with the aim of defining reference criteria for the Portuguese population (the scope of the present study), as well as improving understanding of the value of PWV measurement in predicting major adverse cardiovascular events (MACE)18.

Arterial stiffness is an important area in cardiovascular research, particularly because of its value for risk stratification in various clinical situations. Carotid-femoral pulse wave velocity is without doubt the best indicator of aortic stiffness2, and is unequivocally related to cardiovascular mortality and morbidity in patients with diabetes, hypertension or renal failure, in the elderly, and in the general population5–15. Over the last two decades, our group has focused on this subject, investigating the importance of PWV in conditions such as hypertension22–25 and atherosclerotic disease25, as well as its pathophysiological26–28 and therapeutic29,30 aspects. We have now embarked upon a more ambitious project, which sets out to study prospectively the importance of PWV as a cardiovascular risk marker in what is an unusually large population in this area of research. The inclusion of a large group of low cardiovascular risk individuals posed the greatest challenge but is at the same time the project's most innovative aspect. Although well aware that there is no universally accepted definition of normality, we were able to select a group that is highly representative of the healthy Portuguese population, and to use their characteristics as an objective reference for PWV assessment. We therefore opted for a statistical definition of normality, based on the 95th percentile of PWV, adjusted for age and gender, which is an intuitive and simple measure for use in clinical practice. We were thus able to establish different PWV cut-offs indexed to age and gender, which constitute the reference values for the methodology used to assess arterial stiffness in the study population.

Willum-Hansen et al.14 and Shokawa et al.13 proposed different but useful methods for the clinical interpretation of PWV in the general population, assessing cardiovascular risk in terms of increasing levels of arterial stiffness or defining risk according to PWV thresholds, the latter being well documented in other studies in different clinical contexts.

The present study was designed to explore an alternative method of interpreting PWV values in clinical practice. This was based on recognition of the importance of age as the main factor affecting arterial stiffness. The same PWV value will reflect different pathophysiological contexts in a young or elderly individual, and should therefore be interpreted differently from a clinical standpoint.

Assessment of cardiovascular risk based on the reference values proposed has been shown to have high discriminatory power18, with an adjusted hazard ratio (HR) of 5.252 for the primary endpoint (MACE), and approximate adjusted HRs for cerebrovascular (HR=5.142, confidence interval [CI]: 1.92-17.80) and coronary (HR=3.681, CI: 1.22-11.16) risk. Subgroup analysis provided data that reinforce the validity of the methodology used: the HR for each 1 m/s increment in PWV was 1.748 (CI: 1.470-2.078) for those aged under 50, higher than in the 50–60 age-group (HR=1.514, CI: 1.187-1.933) and in those aged over 60 (HR=1.182, CI: 0.902-1.849). At the same time, the discriminatory power of PWV tended to be stronger in individuals without hypertension or diabetes. While this finding should be interpreted with caution in view of the small number of events recorded to date, they should not be disregarded given the possible clinical implications. If this trend means that PWV measurement is particularly useful in primary prevention, by indicating very early vascular involvement prior to any clinical manifestation of diseases such as hypertension, diabetes or atherosclerosis, then assessment of this parameter based on levels of normality will be of considerable importance.

Comparison of our data with those from joint research in various European centers17 shows a tendency for higher PWV values in the Portuguese population for all the age-groups under analysis. This may be related to specific ethnic characteristics, although certain factors well-documented in the Portuguese population that are known modulators of arterial stiffness, such as high salt intake, cannot be discounted31. Nevertheless, methodological differences may affect linear comparison of PWV values. Our study was based on PWV measurement using a strict and consistent methodology, all assessments being performed by the same operator. In the study referred to above17, PWV measurement was performed using a variety of methods and by a large number of operators, and it is questionable whether the conversion algorithms used would have succeeded in harmonizing the results from the different centers involved. At the same time, the rigor of our definition of normal values is considerably greater than that proposed by this research group.

Another aim achieved in the present study, which has important practical implications, is the establishment of a measure for clinical evaluation of increasing PWV in serial assessments. Studies have demonstrated that patients showing progressive deterioration in PWV present a higher risk for cardiovascular events, but up to now there has been no indication of the percentage of annual PWV increase that exceeds what would be expected physiologically due to aging and that should thus be considered abnormal. The proposed Figure of 5% may prove to be an additional tool for the dynamic evaluation of patients' general state of arterial distensibility, although the implications for cardiovascular risk stratification require confirmation.

However, the project has other aims, in particular to contribute to raising awareness of the concept of arterial distensibility, as well as to demonstrate its suitability for day-to-day clinical practice. In order to achieve these aims, considerable effort has gone into training and into convincing physicians that, if they can obtain PWV assessment for their patients, it should be included in their overall therapeutic approach.

In conclusion, analysis of aortic PWV based on the 95th percentile for age and gender appears to be a simple, intuitive way to broaden use of the concept of arterial stiffness and adapt it to day-to-day clinical practice. At the same time, the establishment of a measure to assess PWV variation over time is crucial to the adoption of this parameter in integrated follow-up programs for patients at cardiovascular risk. This sub-analysis of the EDIVA project, based on the overall results of that epidemiological study, suggests a method of estimating arterial distensibility in clinical practice that represents a significant contribution to improving assessment of cardiovascular risk.

The authors have no conflicts of interest to declare.