Arterial stiffening occurs with aging and may be accelerated

Aortic Pulse Wave Velocity Is Associated With the Presence and Quantity of Coronary Artery Calcium A Community-Based Study Iftikhar J. Kullo, Lawrence F. Bielak, Stephen T. Turner, Patrick F. Sheedy II, Patricia A. Peyser Abstract We investigated the relationship of aortic pulse wave velocity (apwv), a measure of central arterial stiffness, with the presence and quantity of coronary artery calcium (CAC) in a community-based sample of adults without prior history of heart attack or stroke (n 401, mean age 59.8 years, 53% men). ECG-gated waveforms of the right carotid and right femoral artery were obtained by applanation tonometry, and apwv was calculated using established methods. CAC was measured noninvasively by electron beam computed tomography, and CAC score was calculated. apwv was significantly correlated with log(cac 1; r 0.41; P 0.0001) and pulse pressure (r 0.47; P 0.0001). Multivariable logistic and linear regression models were used to identify independent predictors of the presence and quantity of CAC, respectively. In multivariable logistic regression analyses, apwv was associated with the presence of CAC (P 0.011) after adjustment for age, male sex, total cholesterol, high-density lipoprotein cholesterol, diabetes, history of smoking, systolic blood pressure, body mass index, and use of hypertension and statin medications. In multivariable linear regression analyses, apwv was significantly associated with log(cac 1) after adjustment for the covariates enumerated above (P 0.0001). apwv remained significantly associated with both the presence and quantity of CAC even after the additional adjustment for diastolic blood pressure. We conclude that apwv is related to subclinical coronary atherosclerosis independent of conventional risk factors (including indices of blood pressure) and may be a biomarker of cardiovascular risk in asymptomatic individuals. (Hypertension. 2006;47:174-179.) Key Words: blood pressure calcium epidemiology risk factors imaging Arterial stiffening occurs with aging and may be accelerated by other cardiovascular risk factors. 1,2 The resulting increase in velocity of the aortic pulse wave leads to an earlier reflection of the wave back toward the heart, which, in turn, augments systolic blood pressure (SBP), decreases diastolic blood pressure (DBP), and widens pulse pressure (PP). 3 PP, therefore, can be considered a measure of arterial stiffness and has been found to be predictive of several cardiovascular end points independent of conventional risk factors. 4 7 We have previously found PP to be related to the quantity of coronary artery calcium (CAC), 8 a noninvasive surrogate of coronary atherosclerotic burden 9 that is strongly correlated with angiographic coronary artery disease and is predictive of future coronary heart disease (CHD) events. 10 However, PP is a crude, indirect measure of arterial stiffness that is influenced by additional factors, such as left ventricular stroke volume, heart rate, and myocardial contractility. 11,12 A more direct measure of arterial stiffness, aortic pulse wave velocity (apwv), has been shown to predict adverse cardiovascular events in hypertensive 13 and elderly 14 subjects. Recently, Sutton-Tyrrell et al 15 found apwv to be associated with higher cardiovascular mortality and CHD events in older subjects (mean age, 73.7 years) from the community. The association of apwv with subclinical coronary atherosclerosis in the general population is unknown. apwv is a relatively simple, noninvasive test that could be a useful adjunct to conventional CHD risk factors in identifying patients and subgroups in the population that are at increased risk for cardiovascular events. The aim of the present study was to determine whether apwv was associated with the presence and quantity of CAC after consideration of CHD risk factors including indices of blood pressure (BP) in community-based asymptomatic adults without a history of myocardial infarction or stroke. Methods Study Population The study group included participants in the Functional Arterial Changes in Atherogenesis Study that is examining the relationship of measures of arterial function to subclinical coronary atherosclerosis and cardiovascular risk factors in non-hispanic white participants in the community-based Epidemiology of CAC (ECAC) study. The Received September 27, 2005; first decision October 13, 2005; revision accepted November 23, 2005. From the Division of Cardiovascular Diseases (I.J.K.), Department of Radiology (P.F.S.), and the Division of Nephrology and Hypertension, Department of Internal Medicine (S.T.T.), Mayo Clinic and Foundation, Rochester, Minn; and Department of Epidemiology (L.F.B. and P.A.P.), University of Michigan, Ann Arbor, Mich. Correspondence to Iftikhar J. Kullo, Division of Cardiovascular Diseases, Mayo Clinic, 200 First St Southwest, Rochester, MN 55905. E-mail kullo.iftikhar@mayo.edu 2006 American Heart Association, Inc. Hypertension is available at http://www.hypertensionaha.org DOI: 10.116110.1161/01.HYP.0000199605.35173.14 174

Kullo et al Arterial Stiffness and Coronary Calcium 175 ECAC Study is an ongoing study of the pathogenesis of CAC in Rochester, Minn. 16 Participants are not physician- or self-referred and were recruited from the community-based Rochester Family Heart Study 17 independent of risk factor and disease status, except that they do not have a history of previous coronary artery bypass surgery or angioplasty. Between September 2002 and December 2004, 438 participants from the ECAC study had completed the Functional Arterial Changes in Atherogenesis Study protocol. Study protocols were approved by the Mayo Clinic Institutional Review Board, and participants gave written informed consent. We excluded from the present analysis 10 participants who reported a history of previous myocardial infarction or stroke and 27 participants with missing or technically inadequate apwv measurements. The study group included 213 men and 188 women. Weight was measured by an electronic balance, height was measured by a stadiometer, and body mass index was calculated in units of kg/m 2. Resting BP levels were measured in the right arm with a random-zero sphygmomanometer (Hawksley and Sons). Three separate readings were taken 2 minutes apart, and the average of the second and third reading was taken. Blood samples were obtained by venipuncture after an overnight fast. Standard enzymatic methods were used to measure total cholesterol, highdensity lipoprotein (HDL) cholesterol, and plasma glucose. Information about use of hypertension medications, statins, oral hypoglycemic agents, and insulin was obtained at the time of the study. Diabetes was considered present if a subject was being treated with insulin or oral agents or had a fasting glucose level 125 mg/dl. History of smoking was defined as having smoked 100 cigarettes in the past. The diagnosis of hypertension was determined based on BP levels measured at the study visit (SBP 140 mm Hg or DBP 90 mm Hg) or report of a prior diagnosis of hypertension and current treatment with medications for hypertension. Electron Beam Computed Tomography of the Heart The quantity of CAC was measured with an Imatron C-150 electron beam computed tomography (EBCT) scanner (Imatron Inc) as described previously. 18 Calcification was defined as a hyperattenuating focus of 4 adjacent pixels in size (1.38 mm 2 under a field of view of 30 cm) with a computed tomography number 130 Hounsfield Units within 5 mm of the arterial midline. An experienced radiologist interpreted the findings of each tomogram. A score for each focus of CAC was determined as per the method of Agatston et al, 19 and the total calcium score was obtained by summing individual foci scores from each of the 4 epicardial arteries (left main, left anterior descending, circumflex, and right coronary arteries). apwv Measurement of apwv was performed a median of 8 months (range, 2 to 24 months) after EBCT for CAC. The measurement was performed after an overnight fast and 12 to 24 hours off of hypertension medications. Participants were asked to omit caffeinated beverages, smoking, and alcohol for 12 hours before the assessment. Carotid-femoral apwv was determined using the Sphygmocor system (AtCor Medical) by sequentially recording ECG-gated carotid and femoral artery waveforms by applanation tonometry. 20 Distances from the carotid sampling site to the manubrium sternum and from the manubrium sternum to the femoral artery were measured as straight lines between the points on the body surface using a tape measure. The time (t) between the onset of carotid and femoral waveforms was determined as the mean of 10 consecutive cardiac cycles. apwv was calculated from the distance between measurement points (D) and the measured time delay (t) as follows: apwv D/t (m/s), where D is distance in meters and t is the time interval in seconds. In 10 volunteers, the within-subject SD of apwv measured on successive days using this method was 0.66 m/s, which was 7% of the mean apwv in the present study. Statistical Methods Descriptive statistics are given as mean and SD (as well as median and ranges) or number and percentage. Because the distribution of CAC scores was positively skewed, and not all of the participants had detectable CAC, the scores were log-transformed after adding 1. Correlations among age, BP indices, apwv, and log(cac 1) were assessed by Spearman correlation coefficients. We compared apwv in participants with detectable CAC to those without detectable CAC using a t test and used linear regression to assess the relationship of apwv with log(cac 1). Multivariable logistic and linear regression was used to assess whether apwv was an independent predictor of the presence and quantity of CAC, respectively. Within each multivariable regression analysis, 4 models were constructed. In model 1, the variables included CHD risk factors (age, male sex, diabetes, total cholesterol, HDL cholesterol, history of smoking, SBP, and body mass index) and use of hypertension and statin medications. To assess whether PP was a significant predictor, we added DBP to the variables in model 1 rather than PP itself (model 2). Such an approach avoids the problem of potential collinearity between PP and SBP. In model 3, to assess whether apwv was an independent predictor of CAC measures, we added it to the variables in model 1. Finally, in model 4, to assess whether PWV remained a significant predictor after adjustment for PP, we included it in a model with both SBP and DBP. We also investigated interactions between apwv and age, male sex, diabetes, smoking history, total cholesterol, and HDL cholesterol in the statistical models. Statistical significance was determined at P 0.05. Because of sibships in the sample, population-averaged generalized estimating equations 21 were used for regression analyses. Statistical analyses were performed with SAS v 8.2 (SAS Institute). Results The mean age of participants was 60 years, 53% were men, 30.7% were hypertensive, and 6.2% had diabetes (Table 1). Approximately one fourth of participants reported statin use. The prevalence of detectable CAC was 64.3%. apwv was TABLE 1. Participant Characteristics (n 401) Variable Mean SD or No. (%) Median (Range) Age, y 59.8 9.6 60.1 (32.0 to 83.9) Men, n (%) 213 (53.1) Body mass index (kg/m 2 ) 28.5 4.7 27.9 (18.1 to 48.9) Heart rate, bpm 66.4 9.0 64.0 (46.0 to 116.0) Hypertension, n (%) 123 (30.7) Hypertension medication 122 (30.4) use, n (%) Total cholesterol (mg/dl) 198.1 36.1 196.5 (114.0 to 425.5) HDL cholesterol (mg/dl) 53.1 15.1 50.4 (21.2 to 106.7) Statin use, n (%) 95 (23.7) Diabetes, n (%) 25 (6.2) History of smoking, n (%) 192 (47.9) CAC present, n (%) 258 (64.3) Log(CAC 1) 2.8 2.6 2.5 (0.0 to 8.3) CAC score 189.3 467.3 10.7 (0.0 to 4206.7) SBP (mm Hg) 124.6 16.1 122.0 (82.0 to 176.0) DBP (mm Hg) 76.0 9.0 76.0 (54.0 to 108.0) PP (mm Hg) 48.6 13.0 48.0 (16.0 to 94.0) apwv (m/s) 9.5 2.6 8.9 (5.7 to 19.9) Continuous variables are presented as mean SD and median (range), whereas categorical variables are presented as counts and percentages.

176 Hypertension February 2006 TABLE 2. Spearman Correlation Coefficients Among Age, Indices of BP, apwv, and CAC Quantity Variable Age SBP DBP PP apwv Log(CAC 1) Age 0.32 0.07 0.43 0.58 0.47 SBP 0.58 0.81 0.44 0.24 DBP 0.05 0.11* 0.04 PP 0.47 0.31 apwv 0.41 *P 0.05. P 0.001. statistically significantly correlated with age, log(cac 1), and BP indices, including SBP, DBP, and PP (Table 2). Mean SD apwv in participants with detectable CAC (10.1 2.8 m/s) was significantly higher than in participants without detectable CAC (8.5 1.7 m/s; P 0.0001; Figure a). apwv was significantly related to log(cac 1) (r 0.41; P 0.0001; Figure b). In a multivariable logistic regression model, significant predictors of the presence of CAC were age, male sex, and statin use (model 1, Table 3). After the addition of DBP to the regression model, (higher) SBP and (lower) DBP both became significant predictors of CAC presence (P 0.026 and P 0.038, respectively; model 2, Table 3), suggesting that PP was related to the presence of CAC. apwv was significantly associated with the presence of CAC (P 0.011) after adjustment for CHD risk factors (including SBP), hypertension medication use, and statin use (model 3, Table 3). apvw remained significantly associated with the presence of CAC after additional adjustment for DBP (P 0.017; model 4, Table 3), suggesting that the association of apwv with the presence of CAC was independent of PP. In this model, neither SBP nor DBP were significantly associated with CAC presence, indicating that apwv was a better predictor than PP. In a multivariable linear regression model, variables significantly associated with greater log (CAC 1) included age, male sex, history of smoking, hypertension medication use, and statin use (model 1, Table 4). PP was associated with CAC quantity, because both SBP and DBP were statistically significant predictors of log(cac 1), SBP being positively associated and DBP being inversely associated (model 2, Table 4). apwv was significantly associated with CAC quantity (P 0.0001) after adjustment for CHD risk factors (including SBP), hypertension medication use, and statin use (model 3, Table 4). apvw remained significantly independently associated with CAC quantity after additional adjustment for DBP (P 0.0001; model 4, Table 4). In this model, neither SBP nor DBP were significantly associated with CAC quantity, suggesting that apwv was a better predictor of CAC quantity than PP. A significant interaction was noted between apwv and history of smoking in the prediction of presence and quantity of CAC. The regression model suggested that, in smokers, the probability of having detectable CAC increased at a higher rate as apwv increased compared with nonsmokers. Similarly, the association between CAC quantity and apwv was stronger in smokers than in nonsmokers (analyses not shown). Discussion We found apwv to be significantly and positively associated with the presence and quantity of CAC in participants drawn from the community and without a previous history of myocardial infarction or stroke. Importantly, the results provide evidence that apwv improves the ability to predict the presence and quantity of subclinical coronary atherosclerosis beyond what is possible with brachial measures of BP. Several studies have previously investigated the relationship of arterial stiffness and atherosclerotic burden. Herrington et al 22 demonstrated that a measure of lower extremity arterial stiffness was associated with aortic atherosclerotic burden quantified by MRI. In a study of 3000 elderly subjects aged 60 to 101 years, apwv was related to the extent of atherosclerotic plaque in the aorta and the carotid artery. 23 Two small studies that previously assessed the relationship of apwv and CAC yielded conflicting results. Megnien et al 24 found that apwv was not related to CAC quantity in 190 asymptomatic men with 1 CHD risk factor. However, Relation of apwv to the presence and quantity of CAC. (a) Box plots showing distribution of apwv in participants with and without detectable CAC. (b) Scatter plot showing the correlation between apwv and log(cac 1).

Kullo et al Arterial Stiffness and Coronary Calcium 177 TABLE 3. Association of BP Indices and apwv With Presence of CAC: Multivariable Logistic Regression Models Model 1 Model 2 Model 3 Model 4 Variable SE P SE P SE P SE P Age, y 0.100 0.018 0.0001 0.092 0.018 0.0001 0.084 0.019 0.0001 0.078 0.019 0.0001 Male sex 1.298 0.298 0.0001 1.436 0.304 0.0001 1.324 0.295 0.0001 1.450 0.303 0.0001 History of smoking 0.246 0.254 0.332 0.153 0.260 0.556 0.320 0.257 0.213 0.224 0.265 0.297 BMI 0.036 0.028 0.200 0.050 0.029 0.089 0.039 0.029 0.177 0.051 0.029 0.084 Total cholesterol (mg/dl) 0.002 0.003 0.511 0.003 0.003 0.377 0.002 0.003 0.516 0.003 0.003 0.402 HDL cholesterol (mg/dl) 0.011 0.009 0.238 0.008 0.009 0.376 0.011 0.010 0.249 0.008 0.010 0.390 Diabetes 0.477 0.788 0.545 0.516 0.833 0.536 0.394 0.860 0.647 0.422 0.894 0.637 Hypertension medication 0.215 0.322 0.504 0.229 0.328 0.486 0.089 0.327 0.786 0.108 0.333 0.745 use Statin use 1.184 0.367 0.001 1.150 0.367 0.002 1.208 0.367 0.001 1.187 0.369 0.001 SBP, mm Hg 0.011 0.009 0.194 0.026 0.012 0.026 0.004 0.009 0.641 0.019 0.013 0.136 DBP, mm Hg 0.043 0.021 0.038 0.040 0.021 0.056 apwv, m/s 0.182 0.071 0.011 0.172 0.072 0.017 BMI indicates body mass index. See text for model descriptions. Haydar et al 25 noted a significant association between apwv and CAC quantity in 55 patients with chronic kidney disease. We have previously found PP to be related to the presence and quantity of CAC in individuals 50 years of age. 8 In the present study, however, we did not find a significant interaction between apwv and age or sex in the prediction of CAC presence and quantity. apwv was a predictor of the presence and quantity of CAC in both men and women (analyses not shown). We did note a significant interaction between apwv and smoking history; the association between apwv and the presence and quantity of CAC was stronger in smokers than in nonsmokers (analyses not shown). apwv remained significantly associated with the presence and quantity of CAC regardless of which BP indices were used in the regression models. The BP indices included PP, mean arterial pressure, DBP, PP mean arterial pressure, PP SBP, and PP DBP (analyses not shown). Miura et al 26 have pointed out the difficulty of interpreting results of models that include PP along with SBP or DBP and suggest that the appropriate regression model is one that includes both SBP and DBP rather than their difference. The correct approach in assessing PP is to first estimate an individual s risk based on the level of SBP and then to adjust the risk upwards if there is a discordantly low DBP. In our analyses, when apwv was added to SBP and DBP in the multivariable regression models (model 4 in Tables 3 and 4), it remained a significant predictor of the presence and quantity of CAC, whereas SBP and DBP were not significant or only marginally (P 0.10) significant predictors. These results suggest that apwv is a better predictor of the presence and extent of subclinical coronary atherosclerosis than PP. The present study was based on cross-sectional data, and, therefore, we can only speculate on the pathophysiological TABLE 4. Association of BP Indices and apwv With Log(CAC 1): Multivariable Linear Regression Models Model 1 Model 2 Model 3 Model 4 Variable SE P SE P SE P SE P Age, y 0.115 0.011 0.0001 0.106 0.012 0.0001 0.095 0.012 0.0001 0.089 0.013 0.0001 Male sex 1.518 0.215 0.0001 1.623 0.220 0.0001 1.499 0.210 0.0001 1.586 0.215 0.0001 History of smoking 0.548 0.205 0.007 0.462 0.211 0.028 0.622 0.203 0.002 0.547 0.210 0.009 BMI 0.005 0.020 0.808 0.005 0.021 0.810 0.001 0.020 0.977 0.007 0.021 0.727 Total cholesterol (mg/dl) 0.001 0.003 0.743 0.002 0.003 0.603 0.001 0.003 0.820 0.001 0.003 0.688 HDL cholesterol (mg/dl) 0.011 0.007 0.134 0.009 0.007 0.222 0.009 0.007 0.202 0.008 0.007 0.289 Diabetes 0.702 0.473 0.138 0.699 0.473 0.140 0.567 0.478 0.236 0.575 0.478 0.229 Hypertension medication 0.526 0.259 0.043 0.541 0.258 0.036 0.386 0.261 0.140 0.407 0.260 0.118 use Statin use 1.287 0.269 0.0001 1.270 0.265 0.0001 1.278 0.257 0.0001 1.265 0.255 0.0001 SBP, mm Hg 0.013 0.007 0.064 0.025 0.009 0.006 0.004 0.007 0.547 0.015 0.009 0.110 DBP, mm Hg 0.034 0.015 0.026 0.028 0.015 0.059 apwv, m/s 0.182 0.045 0.0001 0.170 0.044 0.0001 The coefficients represent the amount of change in the predicted response (logcac 1) for a unit change in a continuous variable and the presence vs absence difference in the case of a categorical variable. BMI indicates body mass index. See text for model descriptions.

178 Hypertension February 2006 mechanisms that underlie the association between measures of arterial stiffness and CAC. One possibility is that atherosclerosis results in increased arterial stiffness. In cynomolgus monkeys, apwv increases when the animals are fed an atherogenic diet and subsequently decreases when these animals are fed an atherosclerosis regression diet. 27 Another possibility is that increased arterial stiffness may be an early change in response to risk factors and may promote atherosclerotic changes. 28 A third, more likely possibility is that because they share risk factors, such as age and hypertension, the 2 processes may develop concomitantly, arterial stiffness resulting predominantly from changes in the media and the atherosclerosis resulting predominantly from changes in the intima. Mediators such as endothelial NO, angiotensin II, and inflammatory markers may play an important role in both atherosclerosis and arterial stiffening. 29 Longitudinal studies starting in childhood or adolescence may help in elucidating the temporal relationship among risk factors, arterial stiffness, and atherosclerotic plaque formation. Recently, Juonala et al 30 demonstrated that CHD risk factors identified in childhood and adolescence were predictive of carotid artery stiffness in adulthood. We attempted to limit the influence of vasoactive medications by measuring apwv when participants were off antihypertensive medications for 12 to 24 hours. To adjust for the possibility of non-bp related effects of these medications on apwv, we incorporated a variable denoting hypertension medication use in the multiple regression models. We did not find any particular hypertension medication class to be independently related to apwv (analyses not shown). In addition, we repeated our analyses in the subset of participants who were not on hypertension medications. apwv was a significant and independent predictor of CAC presence and quantity in these normotensive participants (analyses not shown). Statin use was associated with the presence and quantity of CAC, likely because the use of these medications indexes the CHD risk factor burden of the participants. In our study, those taking statins were older, had a higher prevalence of hypertension and diabetes, and more often had a history of smoking compared with those not using statins. Indeed, in a retrospective study, statin use was associated with lesser progression in CAC quantity compared with placebo. 31 A strength of the present study is that it was conducted in a community-based group of non-hispanic white research participants without history of myocardial infarction or stroke. A limitation is that the results of this study may not be applicable to other racial or ethnic subgroups. apwv was measured a median of 8 months after EBCT and collection of risk factor data, and changes could have occurred in some of the CHD risk variables during this interval. We performed a separate set of multivariable regression analyses with apwv as the dependent variable and CAC (presence versus absence) or log(cac 1) as one of the independent variables along with other CHD risk factors and the time interval between EBCT and apwv measurement. In these analyses (data not shown), both the presence of CAC and the quantity of CAC were significantly related to apwv; however, the time interval between EBCT and apwv measurement was not significantly associated with apwv. Perspectives There is a need for simple, noninvasive tests that allow more accurate estimates of CHD risk in individual patients. Measurement of apwv is a relatively simple and noninvasive procedure that can be easily performed in an office setting. The device used is relatively small and portable; testing can be accomplished in minutes and is relatively inexpensive. Several studies have shown apwv to be an independent predictor of cardiovascular events. 13 15 Increased central aortic stiffness may also have an important role in the pathogenesis of atrial fibrillation, left ventricular hypertrophy, diastolic dysfunction, and congestive heart failure. 32,33 However, the incremental information about cardiovascular risk that is provided by measurement of apwv needs to be quantified. In addition, age- and sex-based cutoffs for normality need to be established to facilitate interpretation of the results of testing. In conclusion, apwv is related to subclinical coronary atherosclerotic burden in asymptomatic adults. This finding suggests that the association between arterial stiffness and CHD events may be, in part, through an association between arterial stiffness and coronary atherosclerosis. Additional studies are needed to understand the mechanisms underlying this association and the implications for assessment and management of CHD risk. The potential role for apwv in identifying individuals at risk of cardiovascular disease or individuals with early disease merits additional exploration. Acknowledgments This work was supported by a Mentored Patient-Oriented Research Career Development Award (K-23, RR17720) from the National Center for Research Resources (to I.J.K.) and by grant R01 HL46292 and the General Clinical Research Center Grant M01 RR00585 from National Institutes of Health. We acknowledge Libing Wang for performing the data analyses and Wendi Kitsteiner for help with article preparation. References 1. Oliver JJ, Webb DJ. Noninvasive assessment of arterial stiffness and risk of atherosclerotic events. Arterioscler Thromb Vasc Biol. 2003;23: 554 566. 2. Zieman SJ, Melenovsky V, Kass DA. Mechanisms, pathophysiology, and therapy of arterial stiffness. Arterioscler Thromb Vasc Biol. 2005;25: 932 943. 3. Nichols WW, O Rourke M. McDonald s Blood Flow in Arteries: Theoretical, Experimental and Clinical Principles. London: Edward Arnold; 1998. 4. Mitchell GF, Moye LA, Braunwald E, Rouleau JL, Bernstein V, Geltman EM, Flaker GC, Pfeffer MA. Sphygmomanometrically determined pulse pressure is a powerful independent predictor of recurrent events after myocardial infarction in patients with impaired left ventricular function. SAVE investigators. Survival and Ventricular Enlargement. Circulation. 1997;96:4254 4260. 5. Franklin SS, Khan SA, Wong ND, Larson MG, Levy D. Is pulse pressure useful in predicting risk for coronary heart disease? The Framingham heart study. Circulation. 1999;100:354 360. 6. O Rourke M, Frohlich ED. Pulse pressure: Is this a clinically useful risk factor? Hypertension. 1999;34:372 374. 7. Safar ME, St Laurent S, Safavian AL, Pannier BM, London GM. Pulse pressure in sustained essential hypertension: a haemodynamic study. J Hypertens. 1987;5:213 218. 8. Bielak LF, Turner ST, Franklin SS, Sheedy PF II, Peyser PA. Agedependent associations between blood pressure and coronary artery calcification in asymptomatic adults. J Hypertens. 2004;22:719 725. 9. Rumberger JA, Schwartz RS, Simons DB, Sheedy PF II, Edwards WD, Fitzpatrick LA. Relation of coronary calcium determined by electron

Kullo et al Arterial Stiffness and Coronary Calcium 179 beam computed tomography and lumen narrowing determined by autopsy. Am J Cardiol. 1994;73:1169 1173. 10. Arad Y, Spadaro LA, Goodman K, Newstein D, Guerci AD. Prediction of coronary events with electron beam computed tomography. J Am Coll Cardiol. 2000;36:1253 1260. 11. Safar ME, Levy BI, Struijker-Boudier H. Current perspectives on arterial stiffness and pulse pressure in hypertension and cardiovascular diseases. Circulation. 2003;107:2864 2869. 12. Safar ME, Boudier HS. Vascular development, pulse pressure, and the mechanisms of hypertension. Hypertension. 2005;46:205 209. 13. 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. 14. Meaume S, Benetos A, Henry OF, Rudnichi A, Safar ME. Aortic pulse wave velocity predicts cardiovascular mortality in subjects 70 years of age. Arterioscler Thromb Vasc Biol. 2001;21:2046 2050. 15. Sutton-Tyrrell K, Najjar SS, Boudreau RM, Venkitachalam L, Kupelian V, Simonsick EM, Havlik R, Lakatta EG, Spurgeon H, Kritchevsky S, Pahor M, Bauer D, Newman A. Elevated aortic pulse wave velocity, a marker of arterial stiffness, predicts cardiovascular events in wellfunctioning older adults. Circulation. 2005;111:3384 3390. 16. Peyser PA, Bielak LF, Chu JS, Turner ST, Ellsworth DL, Boerwinkle E, Sheedy PF II. Heritability of coronary artery calcium quantity measured by electron beam computed tomography in asymptomatic adults. Circulation. 2002;106:304 308. 17. Turner ST, Weidman WH, Michels VV, Reed TJ, Ormson CL, Fuller T, Sing CF. Distribution of sodium-lithium countertransport and blood pressure in Caucasians five to eighty-nine years of age. Hypertension. 1989;13:378 391. 18. Kullo IJ, Bailey KR, McConnell JP, Peyser PA, Bielak LF, Kardia SL, Sheedy PF II, Boerwinkle E, Turner ST. Low-density lipoprotein particle size and coronary atherosclerosis in subjects belonging to hypertensive sibships. Am J Hypertens. 2004;17:845 851. 19. Agatston AS, Janowitz WR, Hildner FJ, Zusmer NR, Viamonte M Jr, Detrano R. Quantification of coronary artery calcium using ultrafast computed tomography. J Am Coll Cardiol. 1990;15:827 832. 20. De Angelis L, Millasseau SC, Smith A, Viberti G, Jones RH, Ritter JM, Chowienczyk PJ. Sex differences in age-related stiffening of the aorta in subjects with type 2 diabetes. Hypertension. 2004;44:67 71. 21. Zeger SL, Liang KY. Longitudinal data analysis for discrete and continuous outcomes. Biometrics. 1986;42:121 130. 22. Herrington DM, Brown WV, Mosca L, Davis W, Eggleston B, Hundley WG, Raines J. Relationship between arterial stiffness and subclinical aortic atherosclerosis. Circulation. 2004;110:432 437. 23. van Popele NM, Grobbee DE, Bots ML, Asmar R, Topouchian J, Reneman RS, Hoeks APG, van der Kuip DAM, Hofman A, Witteman JCM. Association between arterial stiffness and atherosclerosis: The Rotterdam Study. Stroke. 2001;32:454 460. 24. Megnien JL, Simon A, Denarie N, Del-Pino M, Gariepy J, Segond P, Levenson J. Aortic stiffening does not predict coronary and extracoronary atherosclerosis in asymptomatic men at risk for cardiovascular disease. Am J Hypertens. 1998;11:293 301. 25. Haydar AA, Covic A, Colhoun H, Rubens M, Goldsmith DJ. Coronary artery calcification and aortic pulse wave velocity in chronic kidney disease patients. Kidney Int. 2004;65:1790 1794. 26. Miura K, Dyer AR, Greenland P, Daviglus ML, Hill M, Liu K, Garside DB, Stamler J. Pulse pressure compared with other blood pressure indexes in the prediction of 25-year cardiovascular and all-cause mortality rates: The Chicago Heart Association Detection Project in Industry Study. Hypertension. 2001;38:232 237. 27. Farrar D, Bond M, Riley W, Sawyer J. Anatomic correlates of aortic pulse wave velocity and carotid artery elasticity during atherosclerosis progression and regression in monkeys. Circulation. 1991;83:1754 1763. 28. Demer LL. Effect of calcification on in vivo mechanical response of rabbit arteries to balloon dilation. Circulation. 1991;83:2083 2093. 29. Wang YX, Fitch RM. Vascular stiffness: measurements, mechanisms and implications. Curr Vasc Pharmacol. 2004;2:379 384. 30. Juonala M, Jarvisalo MJ, Maki-Torkko N, Kahonen M, Viikari JS, Raitakari OT. Risk factors identified in childhood and decreased carotid artery elasticity in adulthood: the Cardiovascular Risk in Young Finns Study. Circulation. 2005;112:1486 1493. 31. Callister TQ, Raggi P, Cooil B, Lippolis NJ, Russo DJ. Effect of HMG-CoA reductase inhibitors on coronary artery disease as assessed by electron-beam computed tomography. N Eng J Med. 1998;339: 1972 1978. 32. Reiffel JA. Is arterial stiffness a contributing factor to atrial fibrillation in patients with hypertension? A preliminary investigation. Am J Hypertens. 2004;17:213 216. 33. Levy D, Garrison RJ, Savage DD, Kannel WB, Castelli WP. Prognostic implications of echocardiographically determined left ventricular mass in the Framingham Heart Study. N Engl J Med. 1990;322:1561 1566.