Measuring arterial function in diabetes

The prevalence of diabetes is increasing and the vascular complications of metabolic diseases, particularly diabetes, are arguably the greatest worldwide threat to longevity and disability-free living in the medium-term future. Diabetic vascular disease has unique characteristics, including a propensity to affect large and small calibre arteries. Both have major clinical sequelae. The former predisposes to atherosclerosis and major cardiovascular events, including myocardial infarction and stroke. The microvascular effects are equally serious and include blindness due to diabetic retinopathy and renal failure. Thus, if any disease deserves a detailed analysis of effects at every level of the arterial tree, it is diabetes. Large artery properties and the influences of a range of physiological and pathological factors have become an area of intense investigation recently. This reflects improving evidence that they may be important determinants of outcome. Another factor is the availability of a range of techniques and devices that allow easy measurement in the clinic of one or more of many variables related to the stiffness of large or medium-sized arteries. The fact that there is a confusing array of devices, each producing results pertaining to different aspects of vascular structure and function, should not discourage clinicians in the light of the emerging consensus that arterial properties are both important and quantifiable. Equally, the desire for a simple measurement should not obscure the fact that different techniques provide different information. Apparently conflicting results may often be resolvable by considering what raw data was obtained and how it was manipulated. The report by Lacy et al. [1] in this issue of the journal may be a case in point. Diabetic subjects had stiffer large arteries according to pulse wave velocity measurements but were not different from non-diabetic subjects when measurements of either peripheral or central augmentation index were compared. Pulse pressure is strongly influenced by large artery stiffness and this accounts for the well-known widening of pulse pressure with age. Longitudinal studies have shown that pulse pressure is a good predictor of outcome. There is also good evidence that pulse wave velocity, another relatively uncomplicated measurement, is both a good indicator of changes in arterial properties and a predictor of long-term outcome [2]. β-index is measured by two-dimensional imaging of an artery and calculating the ratio of the natural logarithm of pulse pressure to normalized systolic expansion. β-index is a regional measurement that can be performed at different levels of the arterial tree. Other measures that reflect arterial properties are more global, and also involve mathematical modelling and/or computations of the original data obtained by applanation tonometry, sphygmomanometry and, in some cases, Doppler velocimetry. These include augmentation index, pulse wave velocity, systemic arterial compliance, arterial impedance and a range of other variables. In some cases, measurements made on peripheral arteries are extrapolated to the central arterial circulation using transfer functions. Where it can be easily performed (e.g. the carotid or radial artery), applanation tonometry is generally a satisfactory substitute for invasive measurements for determining the shape of the arterial pulse wave at that site. Subsequently, calibration is required to determine the pressure at different points along the arterial pulse wave tracing. A sphygmomanometric measurement of diastolic and mean pressure at the level of the brachial artery is usually employed. The aortic pressure and augmentation index calculated using the SphygmoCor® device (AtCor Medical Pty. Ltd, West Ryde, New South Wales, Australia) requires application of a general transfer factor. Irrespective of the question of the validity of the general transfer function assumption, this and other global measures of arterial properties are influenced by the structure and function of the arterial system at multiple levels. There are examples of influences on arterial stiffness that have effects at one order of artery size but not at others. Hormone replacement therapy lowers aorto-femoral pulse wave velocity, showing that not everything with favourable effects on arterial stiffness has long-term benefits on cardiovascular outcomes. Coronary plaque burden measured by angiogram was found to be related to pulse wave velocity between the carotid and radial arteries, but not to augmentation index or pulse pressure [3,4]. The study by Lacy et al. [1] involved type I and type II diabetic subjects, as well as non-diabetic controls. The population was also heterogeneous in that many patients were taking blood pressure- and lipid-lowering drugs known to affect measurements of large artery function [5,6]. The diabetic subjects had a higher average blood pressure, and arterial stiffness is inherently blood pressure dependent. Pulse wave velocity from the carotid to the femoral artery was increased, consistent with diabetic subjects having a higher pulse pressure and increased arterial stiffness. However, the augmentation index was similar in the diabetic and control groups. There can be little doubt that arterial function differs between diabetic and non-diabetic populations [7,8]. The likely explanation for why this was not apparent in measurements of augmentation index in the study performed by Lacy et al. [1] is due to the underlying processes influencing this parameter. It was not related to the use of a general transfer factor to derive central from peripheral pressures, although it has been suggested that the appropriate transfer factor may differ in diabetic subjects [9]. In the present study, the same lack of difference in diabetics was observed in measurements of the augmentation index made by applanation of radial and carotid arteries. This finding remained even after correction for heart rate, an important determinant of augmentation index [10]. Augmentation index is determined by the relative magnitudes of the forward and reflected pressure waves and the timing of the arrival of the reflected wave relative to the forward wave. Pulse wave reflection occurs at all levels of the arterial circulation but the predominant or average site is at the level of the primary and secondary branches of the aorta. The timing of pulse wave reflection is determined by the distance to the reflection site and is thus influenced by aortic length (height), cardiac cycle time (heart rate and ejection time) and pulse wave velocity. The study by Lacy et al. [1] demonstrates clear differences in wave reflection patterns between patients with diabetes and controls. These differences appear to interact, such that central augmentation index did not differ between groups. This finding highlights the danger of assessing a single parameter in isolation. Parameters of wave reflection may be useful surrogates of arterial stiffness, but only when they are considered in context. Augmentation index has some credentials as a variable predicting outcome. This has been shown in patients with renal failure in a study by London et al. [11], who found that a 10% increase in augmentation index was associated with a relative risk increase in both all cause and cardiovascular mortality [11]. Previous comparisons between diabetic and non-diabetic subjects have been conflicting. Many studies have shown increased arterial stiffness in the circulation of both type I and type II diabetics, consistent with the pulse wave velocity findings in the study by Lacy et al. [12,13]. However, in some studies, increased arterial stiffness was demonstrable only in those patients with complications, or in some arteries in diabetic subjects but not in others [14,15]. There have also been reports of a high augmentation index in diabetic patients [16]. In summary, diabetic patients have a unique pattern of disordered arterial function. In large arteries, glycation end products have a role in increasing arterial stiffness, and these arteries are susceptible to atherosclerosis. This mechanism likely contributed to increased pulse wave velocity in the diabetic study population reported by Lacy et al. [1]. However, diabetes also alters small artery function and structure, leading to microvascular complications. Augmentation index is strongly influenced by wave reflection. Lacy et al. [1] suggest that the small artery consequences of diabetes are associated with dissipation of the pulse wave, which reduces wave reflection. The effects of large and small artery changes on augmentation index cancel out in their study, with no overall difference between diabetic and non-diabetic groups. Until the exact mechanisms of how arterial properties influence the outcome of disease are clarified, it is important to look for coherence in the results obtained using multiple techniques and variables. Clearly, no single measurement can define the effects of a disease state or treatment on the whole arterial system.