Pulse pressure and inflammatory markers

With the increase in epidemiological information available from large clinical trials and longitudinal cohorts, such as the Framingham Heart Study, together with improvements in methods of non-invasive measurement, biomedical instrumentation and biochemical assays, there is an emerging paradigm shift on how cardiovascular risk can be best assessed for the management of cardiovascular disease (CVD). The shift is from specific parameters having intrinsic value in risk prediction to their being an integral part of a continuous spectrum of inter-related risk factors. Thus, with the general move from overall risk prediction to risk stratification, new parameters have emerged and have been added to the list of traditional factors for the better diagnosis and management of CVD [1]. In the search for more specific markers of cardiovascular risk, factors related to endothelial function, plaque formation, thrombosis and inflammatory processes have been considered. There is now considerable evidence to associate increased levels of blood pressure, arterial stiffness and inflammation with prediction of increased cardiovascular disease [2–4]. Specifically, pulse pressure has been shown to be a strong predictor of CVD in later age (age > 50 years) [5,6]. Recent evidence shows that elevated levels of inflammation markers, in particular C-reactive protein (CRP), are associated with increased cardiovascular risk [7] and that it is the inflammatory process itself that leads to the development of CVD [8]. For a given stroke volume, pulse pressure is determined essentially by arterial stiffness and peripheral wave reflection [9]. Arterial stiffness, or its surrogate measure, pulse wave velocity, has been shown to be an independent predictor of cardiovascular and all cause mortality in elderly hypertensive subjects and in subjects with end-stage renal disease [10,11]. Reduced aortic compliance has also been associated with prediction of ischaemic threshold in coronary artery disease [12]. Hence, given the common association of haemodynamic and inflammatory parameters with CVD, there has been increasing interest in possible links between the two. This has been investigated by analysis of data from very large population groups. Abramson et al. [13] analysed data from 9867 subjects from the Third National Health and Nutrition Examination Survey (NHANES III, n = 33 994 subjects; 1988–94). This study established an association between pulse pressure and CRP in healthy adults in the USA, independent of systolic and diastolic pressure. In this issue of the journal, Amar et al. [14] report a similar association in a much smaller French population of 891 healthy subjects from the MONICA study (Monitoring Trends and Determinants in Cardiovascular Disease, 1997–99). However, their study also showed that this association is independent of other parameters of arterial structure and stiffness such as carotid intima–media thickness and the presence of atherosclerotic plaques and aortic pulse wave velocity, respectively. One of the features of these investigations is the substantial statistical treatment required to establish associations between variables and, given the large numbers, relatively weak correlations become statistically significant. However, in the absence of plausible mechanisms, it would be difficult to establish causal relationships, even though strong associations are found in large studies. A study by Chae et al. [15] in 508 healthy men suggests some direct association between levels of blood pressure and two specific inflammatory markers, soluble intercellular adhesion molecule-1 (sICAM-1) and interleukin-6. In contrast, the study by Amar et al. [14] found no association between pulse pressure and sICAM-1. Furthermore, while data obtained by Chae et al. [15] suggests a possible underlying mechanism for the association between blood pressure and atherosclerosis, Amar et al. [14] found that the association between pulse pressure and CRP was independent of atherosclerosis (measured locally as carotid plaques) and pulse wave velocity (measured globally across the aortic trunk). There are potential methodological and conceptual limitations to non-invasive measurements of arterial structural and functional parameters and then relating them to basic mechanisms of inflammation and to overall cardiovascular risk. The localized nature of plaque formation and the irregular distribution along the aortic trunk would make it difficult to associate changes in aortic pulse wave velocity and atherosclerosis. This is because the surrogate association of pulse wave velocity with arterial stiffness is primarily determined by medial changes in the arterial wall [9], especially if the measurements are made over large distances (normally carotid-femoral artery). To obtain significant changes in pulse wave velocity due to atherosclerotic processes, there would have to be gross generalized systemic changes. In studies performed in primates, Farrar et al. [16,17] showed that significant changes in aortic pulse wave velocity were found only when monkeys were fed a high cholesterol diet for 3 years, thus causing a generalized uniform diffusion of atherosclerotic plaques throughout the aorta, resulting in generalized increased wall thickness. In addition, atherosclerotic animals also showed an increase in pulse pressure [16]. Local changes in pulse wave velocity (measured over 5 cm) due to modulations of endothelial mediators such as nitric oxide [18] and endothelin-1 [19] have been shown in the sheep iliac artery (essentially a muscular artery), but it remains to be seen whether similar changes can be reliably detected in aortic wave velocity (measured over some 60–80 cm), the essential component of stiffness that is a major determinant of pulse pressure. In population studies where significant differences exist in the severity of atherosclerosis between American and Japanese subjects [20], there was no marked difference in aortic distensibility measured post-mortem in both groups [21]. Similar findings were also observed in our own studies in Chinese populations where serum cholesterol was lower than that in Western populations (and presumably the prevalence of atherosclerosis), but aortic pulse wave velocity was markedly higher [22] Studies such as those conducted by Amar et al. [14] establish a range of correlations between a wide spectrum of parameters subdivided into inflammatory, arterial, haemodynamic, haemotological and clinical variables. However, in relation to pulse pressure and inflammation, to date, it is not yet clear whether the increase in pulse pressure causes mechanical changes to the endothelial cells, promoting a cascade of mechanisms associated with inflammation, and thus detected by an elevated level of inflammatory markers, or whether it is the inflammation that then leads to increases in pulse pressure. Some studies performed in vitro have established mechanical correlates with flow and endothelial function. In peripheral beds, there is a tendency for increased pulse pressure to be associated with reversal of flow during diastole [23]. Flow reversal, in association with oscillatory sheer, has been shown to increase expression of adhesion molecule in human endothelial cells in culture [24]. In intact carotid arteries of rabbits [25], it was shown that increased pulse pressure impairs endothelial relaxation induced by acetylcholine, and the inhibition of relaxation is related to generation of oxygen radicals. Because the study by Amar et al. [14] found no interaction between pulse pressure and sICAM-1, the authors suggest that the in-vitro association of pulse pressure and adhesian molecules in endothelial cells in culture may not be detected in some population-based studies. The increasing interest in cardiovascular risk evaluation and stratification has initiated the convergence of the various fields in the physical and life sciences to produce relevant knowledge for the reliable assessment and treatment of clinical problems. This is manifest by recent investigations such as that by Amar et al. [14] and other similar studies. Population data analysis indicates a strong association between a physical variable (pulse pressure) that can be measured non-invasively and biochemical markers of inflammation as robust predictors of cardiovascular risk [26,27]. The challenge is to establish plausible mechanisms so as to move from associations to causal relationships. Of course, with advances in molecular biology, genetic sciences and bio-informatics, this will provide a basis for more accurate and precise risk stratification in groups with differences in their disease profiles and their response to physical and pharmacological interventions, and also the same benefits for the elderly. The search for accurate risk stratification with solid causal relationships will become a significant driving force for much of the research effort in addressing the increasing and changing cardiovascular burden in developing societies and in the ageing global population.