Summary. Introduction

Clin Physiol Funct Imaging (2008) doi: 10.1111/j.1475-097X.2008.00816.x 1 Arterial compliance and endothelium-dependent vasodilation are independently related to coronary risk in the elderly: the Prospective Investigation of the Vasculature in Uppsala Seniors (PIVUS) study Lars Lind Department of Medicine, Uppsala University Hospital, Uppsala, Sweden Summary Correspondence Lars Lind, MD, Department of Medicine, University Hospital, 751 85 Uppsala, Sweden E-mail: lars.lind@medsci.uu.se Accepted for publication Received 15 February 2008; accepted 30 April 2008 Key words artery; cardiovascular; compliance; elderly; endothelium; risk; vasodilation Background: Measurements of both arterial compliance and endothelium-dependent vasodilation have previously been related to coronary risk factors, but not in the same study. In the Prospective Investigation of the Vasculature in Uppsala Seniors (PIVUS) study, we studied the interplay between arterial compliance and endothelium-dependent vasodilation on coronary risk. Methods: In the population-based PIVUS study (1016 subjects aged 70 years), arterial compliance was determined by ultrasound in the carotid artery, by pulse wave analysis (augmentation index) and the stroke volume to pulse pressure ratio by echocardiography, while endothelium-dependent vasodilation (EDV) was assessed by the invasive forearm technique with acetylcholine, brachial artery ultrasound [flow-mediated dilatation (FMD)] and pulse wave analysis with terbutaline provocation [change in reflection index (RI)]. Results: Factor analysis disclosed three major factors. The first factor was reflecting the three arterial compliance methods, the second factor was reflecting EDV and the change in RI, while the third factor mainly was reflecting FMD. All these three factors were independently related to the Framingham risk score in multiple regression analysis (P<0Æ0001, P = 0Æ0002 and P = 0Æ0046, respectively). Conclusions: In conclusion, both arterial compliance and endothelium-dependent vasodilation were independently related to the Framingham risk score, suggesting that it is worthwhile to evaluate the parallel use of these two vascular characteristics in a prospective fashion. Introduction Two types of functional properties of the arteries have been commonly evaluated in the recent years: arterial compliance and endothelium-dependent vasodilation. Three different techniques are widely used to evaluate arterial compliance in humans: pulse wave velocity in the aorta, analysis of pulse wave refection and the distensibility of an arterial segment evaluated by ultrasound (Oliver & Webb, 2003). All three techniques have been shown to predict future cardiovascular events in different populations (Blacher et al., 1998; Guerin et al., 2001; Laurent et al., 2001; London et al., 2001). Also another measure of arterial compliance, the stroke volume to pulse pressure ratio (SV PP ratio), predicts cardiovascular events in prospective studies (Lind et al., 2004). Two different techniques are widely used to evaluate endothelium-dependent vasodilation in humans: the invasive forearm technique using intra-brachial infusion of acetylcholine and the ultrasound-based non-invasive evaluation of flowmediated dilatation (FMD) in the brachial artery. Both techniques predict future cardiovascular events (Perticone et al., 2001; Gokce et al., 2003). A third new technique evaluates the reduction in pulse wave reflection following a beta-2 agonist challenge (Hayward et al., 2002; Wilkinson et al., 2002). Few studies have measured both arterial compliance and vasoreactivity in the same study. In one such study, Nair et al. (2005) evaluated arterial compliance with pulse wave analysis and vasoreactivity with FMD in 122 subjects with peripheral artery disease (PAD) or risk factors for PAD. In that study, arterial compliance was more closely related to coronary risk than FMD. 1

2 Vascular function and coronary risk, L. Lind In the Prospective Investigation of the Vasculature in Uppsala Seniors (PIVUS) study, three methods to evaluate endotheliumdependent vasodilation and three methods to evaluate arterial compliance were investigated in more than 1000 subjects aged 70 years living in the community of Uppsala, Sweden. We have previously reported that indices of endothelium-dependent vasodilation were related to coronary risk (Lind et al., 2005), as assessed by the Framingham risk score. We have also recently shown different measurements of arterial compliance to be related to coronary risk (Lind et al., 2006), and that arterial compliance could influence measurements of FMD (Wilson et al., 1998). In the present investigation, we studied the interplay between measures of arterial compliance and endothelium-dependent vasodilation and their relation to coronary risk, with the hypothesis that these two characteristics of vascular function independently were related to Framingham risk score. Material and methods This section has previously been given in detail together with basic characteristics and vasodilatory tests of the study sample (Lind et al., 2005, 2006). Subjects Eligible were all subjects aged 70 years living in the community of Uppsala, Sweden. The subjects were randomly chosen from the register of community living. One thousand and sixteen subjects participated giving a participation rate of 50Æ1%. Approximately 10% of the cohort reported a history of coronary heart disease, 4% reported stroke and 9% diabetes mellitus. Almost half the cohort reported any cardiovascular medication (45%), with antihypertensive medication being the most prevalent (32%). Fifteen per cent reported use of statins, while insulin and oral antiglycemic drugs were reported in 2% and 6%, respectively (see Nair et al., 2005 for further details). The study was approved by the Ethics Committee of the University of Uppsala. All subjects were investigated in the morning after an overnight fast. No medication or smoking was allowed after midnight. An arterial cannula was inserted in the brachial artery for blood sampling and later regional infusions of vasodilators. Blood pressure was measured by a calibrated mercury sphygmomanometer. From these data, the Framingham risk score was calculated (Wilson et al., 1998). (EDV) and 5 and 10 lg min )1 for SNP (Nitropress, Abbot, UK) to evaluate endothelium-independent vasodilation (EIDV). Endothelium-dependent vasodilation was defined as FBF during infusion of 50 lg min )1 of acetylcholine minus resting FBF divided by resting FBF. EIDV was defined as FBF during infusion of 10 lg min )1 of SNP minus resting FBF divided by resting FBF. The brachial artery ultrasound technique The brachial artery was assessed by external B-mode ultrasound imaging 2 3 cm above the elbow (Acuson XP128 with a 10-MHz linear transducer; Acuson, Mountain View, CA, USA). An increase in blood flow was induced by inflation of a pneumatic cuff placed around the forearm to a pressure at least 50 mmhg above systolic blood pressure for 5 min. FMD was defined as the maximal brachial artery diameter recorded between 30 and 90 s following cuff release minus diameter at rest divided by the diameter at rest. Pulse wave analysis for arterial compliance and vasoreactivity The pulse wave in the radial artery was captured by aplanation tonometry( Sphygmocor; Pulse Wave Medical Ltd, Sydney, Australia). Based on transfer functions, aortic systolic and diastolic blood pressures were calculated. Central pulse pressure (PP) was defined as central systolic minus central diastolic blood pressure. Aortic augmentation index is calculated as the ratio between amplitude of the first reflected wave divided by the amplitude of the first systolic peak. The relative height of the first diastolic reflected wave (the reflection index, RI) was used for the evaluation of vasoreactivity. After a baseline recording, terbutaline (0Æ25 mg given subcutaneously) was given and a re-evaluation was performed after 15 and 20 min. The maximal change of the RI following terbutaline in relation to baseline was given. Carotid artery compliance The diameter of the common carotid artery (CCA) of the right side 1 2 cm proximal of the bifurcation was measured at its maximal diameter in systole and the minimal diameter in diastole. The distensibility of the CCA was calculated as the change in diameter maximum to minimum in relation to the minimal diameter in diastole divided by the central PP obtained by pulse wave analysis. The invasive forearm technique Forearm blood flow (FBF) was measured by venous occlusion plethysmography (Elektromedicin, Kullavik, Sweden). After evaluation of resting FBF, local intra-arterial drug infusions were given during 5 min for each dose. The infused dosages were 25 and 50 lg min )1 for acetylcholine (Clin-Alpha, Basel, Switzerland) to evaluate endothelium-dependent vasodilation Stroke volume to pulse pressure ratio Echocardiography was performed (2Æ5 MHz transducer; Acuson XP124) and using Teichholz formula ejection fraction (EF) and stroke volume (SV) were calculated. The SV PP ratio was calculated as SV divided by central PP (achieved by pulse wave analysis).

Vascular function and coronary risk, L. Lind 3 Statistics Non-normally distributed variables were transformed to achieve a normal distribution. Relationships between pairs of variables were evaluated by PearsonÕs correlation coefficient. Principal component analysis was used to generate factors and factor scores. Multiple regression analysis was applied to relate several independent variables to a dependent variable. Two-tailed significance values were given with P<0Æ05 regarded as significant. The statistical program package STATVIEW (SAS Inc., Cary, NC, USA) was used. Results Basic characteristics of the study population and the investigated indices of arterial compliance and endothelium-dependent vasodilation are given in Table 1. When univariate correlations were calculated between indices of measurements of arterial compliance and those of endothelium-dependent vasodilation, the only relationship that reached statistical significance was that between carotid artery distensibility and FMD (r =0Æ11, P<0Æ01, see Table 2 for further details). In factor analysis with the six indices of arterial compliance and endothelium-dependent vasodilation, three factors with an eigenvalue >1Æ0 were created. The first factor was mainly reflecting the three arterial compliance methods, the second factor was mainly reflecting EDV and the change in RI, while the third factor mainly was reflecting FMD (Table 3). When the factor scores for the three factors mentioned above were related to the Framingham risk score in univariate analysis, all three factors were correlated with coronary risk (r = 0Æ30, P <0Æ0001, r =0Æ14, P =0Æ0021 and r =0Æ14, P =0Æ0021, respectively). Also in a multiple regression analysis with the Framingham risk score as the dependent variable, all three factors were independently related to coronary risk (P<0Æ0001, P = 0Æ0002 and P = 0Æ0046, respectively, see Table 4 for details). Discussion The main finding in the present study was that arterial compliance and endothelium-dependent vasodilation were independently related to coronary risk. Although arterial compliance was more closely related to coronary risk than endothelium-dependent vasodilation, this finding indicates a Table 1 Basic characteristics, major cardiovascular risk factors and measures of endothelium-dependent vasodilation and arterial compliance in the total sample. Total sample n 1016 Females (%) 50Æ2 Height (cm) 169 ± 9Æ1 Weight (kg) 77 ± 14 Waist circumference (cm) 91 ± 12 BMI (kg m )2 ) 27Æ0 ±4Æ3 Waist hip ratio 0Æ90 ± 0Æ075 SBP (mmhg) 150 ± 23 DBP (mmhg) 79 ± 10 Heart rate (beats min )1 ) 62 ± 8Æ7 Serum cholesterol (mmol l )1 ) 5Æ4 ±1Æ0 LDL-cholesterol (mmol l )1 ) 3Æ3 ±0Æ88 HDL-cholesterol (mmol l )1 ) 1Æ5 ±0Æ42 Serum triglycerides (mmol l )1 ) 1Æ3 ±0Æ60 Fasting blood glucose (mmol l )1 ) 5Æ3 ±1Æ6 Current smoking (%) 11 EDV (%) 459 (199 909) EIDV (%) 328 (149 629) FMD (%) 4Æ4 (0Æ0 9Æ7) Change in RI (%) )31 ± 14 CCA distensibility (% mmhg )1 ) 0Æ086 (0Æ04 0Æ15) Stroke volume pulse pressure (ml mmhg )1 ) 1Æ3 (0Æ80 2Æ3) Aorta AIx (%) 151 ± 17 Means are given ± SD or as median and 10th and 90th percentiles in parentheses. SBP, systolic blood pressure; DBP, diastolic blood pressure; BMI, body mass index; EDV, endothelium-dependent vasodilation (invasive forearm technique); EIDV, endothelium-independent vasodilation (invasive forearm technique); FMD, flow-mediated dilatation; AIx, augmentation index; RI, reflectance index; CCA, common carotid artery. Table 2 Relationships between three indices of arterial compliance and three indices of endothelium-dependent vasodilation. PearsonÕs correlation coefficient is given. CCA distensibility SV PP ratio AIx EDV )0Æ02 )0Æ06 0Æ06 FMD 0Æ11** )0Æ02 0Æ01 Change in RI )0Æ05 0Æ06 0Æ03 EDV, endothelium-dependent vasodilation (invasive forearm technique); FMD, flow-mediated dilatation; AIx, augmentation index; RI, reflectance index; CCA, common carotid artery; SV PP, stroke volume to pulse pressure. **P<0Æ01. Table 3 Factor analysis of three indices of arterial compliance and three indices of endothelium-dependent vasodilation. Eigenvalues for the factors are given in the first row and factor loadings for the six vascular function indices are given in the following rows. Factor 1 Factor 2 Factor 3 Eigenvalue 1Æ64 1Æ10 1Æ01 CCA distensibility 0Æ68 )0Æ29 0Æ06 SV PP ratio 0Æ81 0Æ07 )0Æ13 AIx )0Æ73 0Æ006 0Æ08 Change in RI 0Æ08 0Æ74 0Æ23 EDV )0Æ17 )0Æ57 )0Æ32 FMD 0Æ04 )0Æ37 0Æ90 EDV, endothelium-dependent vasodilation (invasive forearm technique); FMD, flow-mediated dilatation; AIx, augmentation index; RI, reflectance index; CCA, common carotid artery; SV PP, stroke volume to pulse pressure.

4 Vascular function and coronary risk, L. Lind Table 4 Multiple regression analysis with the Framingham risk score as the dependent variable and the factor scores for the three factors as three independent variables. Partial correlation coefficient t-value P-value Factor 1 )0Æ31 7Æ1 <0Æ0001 Factor 2 0Æ16 3Æ7 0Æ0002 Factor 3 )0Æ12 2Æ8 0Æ0046 value for measurements of both arterial compliance and endothelium-dependent vasodilation in the assessment of coronary risk. Although some investigators have shown a relationship between distensibility in an arterial segment and FMD (Witte et al., 2005), no other close relationships between arterial compliance and endothelium-dependent vasodilation were seen in the present study, neither when using univariate linear regression, nor applying factor analysis. In the present study, we chose to relate the calculated factors rather than the original indices of arterial compliance and endothelium-dependent vasodilation to the Framingham risk score. As we have previously shown the three indices of arterial compliance to be inter-related (Lind et al., 2006) and also the two indices reflecting endothelium-dependent vasodilation in resistance arteries to be correlated (Lind et al., 2005), the factor analysis approach calculated three factors not being inter-related out of six indices. Problems with relationships between the independent variables in the multiple regression model could be avoided with this approach. The present finding that arterial compliance is more closely related with coronary risk than FMD is in agreement with a recent study using pulse wave analysis to assess arterial compliance in large and small vessels (Nair et al., 2005). In that study, only a non-significant tendency for a relationship between FMD and coronary risk was seen. The larger sample size in the present study might be one reason for that we could detect an independent relationship between FMD and coronary risk. It is generally accepted that FMD and acetylcholine-mediated vasodilation in the forearm are markers of endothelial NO release. However, stiff arteries have a limitation to dilate that might reduce the vasodilatory action of NO. On the other hand, it is generally believed that structural components in the arterial wall, such as increased collagen and reduced elastin, together with atherosclerotic lesions are the main determinants of arterial compliance. However, it has also been shown that arterial compliance includes a more dynamic NO-dependent component (Kinlay et al., 2001; Wilkinson et al., 2002). Thus, while measures of endothelium-dependent vasodilation mainly are NO dependent and arterial compliance mainly is dependent on structural elements, they have been shown to share some common features. However, no strong relationships were seen between these two vascular characteristics and therefore they can both be related to coronary risk independently of each other. Lack of NO might lead to reduced coronary vasodilation in situation when the myocardium needs to increase coronary blood flow because of an increased metabolic demand. Stiff arteries, on the other hand, would increase cardiac work because of increased afterload. This would increase the metabolic demand and also would induce hypertrophy of the left ventricle, further increasing metabolic demand. So, defects in both endothelium-dependent vasodilation and arterial compliance would act in concert to promote coronary events. The present sample is limited to Caucasians aged 70 years. So, caution should be made to draw conclusions to other ethnic and age groups. Another limitation is that pulse wave velocity was not evaluated. The cohort also has a low prevalence of cardiovascular diseases. We have also published several papers from this cohort (see http://www.medsci.uu.se/pivus/pivus.htm for details) with different hypothesis. Thus, some of these hypotheses might therefore produce false-positive findings by chance and these data must therefore be taken with caution until reproduced by others. In this study, the Framingham score was used as a surrogate for hard coronary events. It will take several years before we have gathered enough coronary events in a prospective fashion to make a meaningful prospective evaluation. Before that, we cannot say if testing with measures of both compliance and vasodilation adds some predictive power to the easier collected Framingham score. However, the fact that both arterial compliance and vasoreactivity were related to the Framingham score in an independent fashion indicates that the present findings are worthwhile to explore in a prospective fashion. In conclusion, both arterial compliance and endotheliumdependent vasodilation were independently related to the Framingham risk score, suggesting that it is worthwhile to evaluate the parallel use of these two vascular characteristics in a prospective fashion. Acknowledgement The outstanding work at the endothelium laboratory performed by Nilla Fors, Jan Hall, Kerstin Marttala and Anna Stenborg is highly acknowledged. References Blacher J, Pannier B, Guerin AP, Marchais SJ, Safar ME, London GM. Carotid arterial stiffness as a predictor of cardiovascular and all-cause mortality in end-stage renal disease. Hypertension (1998); 32: 570 574. Gokce N, Keaney JF Jr, Hunter LM, Watkins MT, Nedeljkovic ZS, Menzoian JO, Vita JA. Predictive value of noninvasively determined endothelial dysfunction for long-term cardiovascular events in patients with peripheral vascular disease. J Am Coll Cardiol (2003); 41: 1769 1775. Guerin AP, Blacher J, Pannier B, Marchais SJ, Safar ME, London GM. Impact of aortic stiffness attenuation on survival of patients in endstage renal failure. Circulation (2001); 103: 987 992.

Vascular function and coronary risk, L. Lind 5 Hayward CS, Kraidly M, Webb CN, Collins P. Assessment of endothelial function using peripheral waveform analysis. A clinical application. J Am Coll Cardiol (2002); 40: 521 528. Kinlay S, Creager MA, Fukumoto M, Hikita H, Fang JC, Selwyn AP, Ganz P. Endothelium-derived nitric oxide regulates arterial elasticity in human arteries in vivo. Hypertension (2001); 38: 1049 1053. Laurent S, Boutouyrie P, Asmar R, Gautier I, Laloux B, Guize L, Ducimetiere P, Benetos A. Aortic stiffness is an independent predictor of all-cause and cardiovascular mortality in hypertensive patients. Hypertension (2001); 37: 1236 1241. Lind L, Andren B, Sundstrom J. The stroke volume pulse pressure ratio predicts coronary heart disease mortality in a population of elderly men. J Hypertens (2004); 22: 899 905. Lind L, Fors N, Hall J, Marttala K, Stenborg A. A comparison of three different methods to evaluate endothelium-dependent vasodilation in the elderly. The Prospective Investigation of the Vasculature in Uppsala Seniors (PIVUS) study. Arterioscler Thromb Vasc Biol (2005); 25: 2368 2375. Lind L, Fors N, Hall J, Marttala K, Stenborg A. A comparison of three different methods to evaluate arterial compliance in the elderly. The Prospective Investigation of the Vasculature in Uppsala Seniors (PIVUS) study. J Hypertens (2006); 24: 1075 1082. London GM, Blacher J, Pannier B, Guerin AP, Marchais SJ, Safar ME. Arterial wave reflections and survival in end-stage renal failure. Hypertension (2001); 38: 434 438. Nair N, Oka RK, Waring LD, Umoh EM, Taylor CB, Cooke JP. Vascular compliance versus flow-mediated vasodilation: correlation with cardiovascular risk factors. Vasc Med (2005); 10: 275 283. Oliver JJ, Webb DJ. Noninvasive assessment of arterial stiffness and risk of atherosclerotic events. Arterioscler Thromb Vasc Biol (2003); 23: 554 566. Perticone F, Ceravolo R, Pujia A, Ventura G, Iacopino S, Scozzafava A, Ferraro A, Chello M, Mastroroberto P, Verdecchia P, Schillaci G. Prognostic significance of endothelial dysfunction in hypertensive patients. Circulation (2001); 104: 191 196. Wilkinson IB, Hall IR, MacCallum H, Meckenzie IS, McEniery CM, van der Arend BJ, Shu YE, MacKay LS, Webb DJ, Cockcroft JR. Pulse wave analysis. Clinical evaluation of a noninvasive, widely applicable method for assessing endothelial function. Arterioscler Thromb Vasc Biol (2002a); 22: 147 152. Wilkinson IB, Qasem A, McEniery CM, Webb DJ, Avolio AP, Cockcroft JR. Nitric oxide regulates local arterial distensibility in vivo. Circulation (2002b); 105: 213 217. Wilson PWF, D Agostino RB, Levy D, Belanger AM, Silbershatz H, Kannel WB. Prediction of coronary heart disease using risk factor categories. Circulation (1998); 97: 1837 1847. Witte DR, van der Graaf Y, Grobbee DE, Bots ML, SMART Study Group. Measurement of flow-mediated dilatation of the brachial artery is affected by local elastic vessel wall properties in high-risk patients. Atherosclerosis (2005); 182: 323 330.