Use of pulse wave analysis to measure arterial stiffness in old age.

SIR—We read with interest Fantin’s cross-sectional study comparing the influence of age on arterial stiffness using the technique of applanation tonometry with pulse wave analysis (PWA) [1]. Their study is particularly welcome since the literature contains few direct comparisons of different markers of arterial stiffness [2]. The overall PWA completion rate was 66% (302/458) for carotid tonometry and 99% (453/458) for radial tonometry. This is consistent with previous research that has found carotid tonometry technically more demanding, less acceptable to patients and less repeatable than radial artery tonometry [3]. This might explain why the fitted straight-line for female carotid augmentation pressure (AG) appears inconsistent with the rest of the data [1]. The relationship between radial AG and age was best described by a simple straight line. The relationship between radial augmentation index (AIx) and age was best explained by a quadratic curve with a plateau at around 60 years for women and 80 years for men. We have also used SphygmoCor (with radial artery tonometry) in a preliminary comparative study of arterial stiffness in acute stroke patients, individuals with equivalent cardiovascular risk and healthy volunteers (N = 59, 73% aged over 55 years; LREC Ref:05/S0801/45). While our sample size is insufficient to permit a useful comparison by sex (and we share with Fantin the problem of a heterogeneous study population) we also observed similar age-related patterns (see Figure 1). Both AIx and AG had an initial linear relationship with age, before levelling off (at 55 years for AIx and 75 years for AG) and finally dipping at the extremes of older age. We suspect this final dip reflects a ‘healthy survivor effect’ (a bias to which cross-sectional data is particularly prone) rather than a true physiological reversal of arterial ageing. Our analysis employs both simple linear regression and non-parametric ‘local linear regression smoothing’ (SPSS v15). The latter approach is useful for exploring arterial stiffness across age-bands without having to make assumptions about the overall distribution of the observations [4]. Since AIx varies with heart rate [5], it is also shown standardised to 75 bpm. It would have been useful to see the results of similar analyses in Fantin’s study. While our findings appear comparable with Fantin’s much larger study, we think it is premature to conclude that AG is a better guide to arterial ageing than AIx for two main reasons [1]. Firstly, pooling data from several cross-sectional surveys creates a heterogeneous study population, which restricts generalisability and potentially compounds any individual study biases. Secondly, as the accompanying editorial suggests, the clinical utility of AG and AIx will depend upon their relative prognostic ability to predict future cardiovascular events [2]. Nevertheless, Fantin’s findings justify further research into the role of PWA indices in cardiovascular risk stratification. Figure 1.