Brachial artery hyperemic blood flow velocities are related to carotid atherosclerosis Susann J. Järhult, Johan Sundström and Lars Lind

Clin Physiol Funct Imaging (29) doi: 1.1111/j.1475-97X.29.879.x Brachial artery hyperemic blood flow velocities are related to carotid atherosclerosis Susann J. Järhult, Johan Sundström and Lars Lind Department of Medicine, Faculty of Medical sciences, Uppsala University Hospital, Uppsala, Sweden Summary Correspondence Dr Susann J Järhult, Department of Medicine, Uppsala University Hospital 75185 Uppsala, Sweden E-mail: susann.jarhult@medsci.uu.se Accepted for publication Received 2 January 29; accepted 15 May 29 Key words atherosclerosis; bloodflow; brachial; carotid; hyperemia Objective: Cardiovascular (CV) risk relates to the blood flow velocity pattern in the brachial artery during hyperemia, especially to the hyperaemic systolic to diastolic mean blood flow velocity (SDFV) ratio. Here, we investigated the relations between SDFV in the brachial artery and different characteristics of carotid atherosclerosis. Material and methods: Data were collected from 116 7-year-olds participating in the Prospective Investigation of Uppsala Seniors study. Doppler recordings of blood flow velocity during hyperemia were analysed in the brachial artery. In the carotid artery, intima-media thickness (IMT) was recorded together with an assessment of echogenicity by the Grey scale median (GSM) method in both overt plaques and in the intima-media complex (IM-GSM). Results: The SDFV ratio was related to the number of carotid arteries affected by plaque (P = Æ18) and inversely to plaque echogenicity (P = Æ3). The SDFV ratio was also related to IMT (P =Æ22) and inversely to IM-GSM (P =Æ1). These relations were statistically significant also after adjusting for major CV risk factors, individually as well as summarised as the Framingham risk score. Conclusion: Our results indicate that the hyperemic systolic to diastolic blood flow velocity ratio in the brachial artery is related to atherosclerosis in the carotid artery. Introduction Flow-mediated dilation (FMD), normally within the range of 5 1% during reactive hyperemia. is related to the major cardiovascular (CV) risk factors as well as to coronary atherosclerosis (Celermajer et al., 1992; Anderson et al., 1995; Corretti et al., 22) CV event-prediction in selected patient populations by monitoring the compensatory mechanisms of a temporary ischemia of the brachial artery is known (Yeboah et al., 27). However, in a study of Framingham Offspring Study participants (Benjamin et al., 24), the previously reported associations between FMD and CV risk factors were mainly due to the relations of the risk factors to the hyperemic blood flow per se rather than a direct effect on FMD. In a recent study using data from 116 7-year-olds in the Prospective Investigation of the Vasculature in Uppsala Seniors (PIVUS) study, the hyperemic blood flow velocities in systole and diastole were treated as separate variables in the evaluation of their relationship to FMD and coronary risk in accordance with the Framingham findings. In that study, we found systolic and diastolic hyperemic blood flow velocities to be related to coronary risk in different directions, and the hyperemic systolic to diastolic blood velocity (SDFV) ratio to be associated with coronary risk independently of FMD (Jarhult et al., 28). SDFV, mirroring both vascular stiffness and peripheral resistance, might therefore provide useful information further enhancing the risk evaluation. The intima-media thickness (IMT) of the carotid artery measured by ultrasound is frequently used to evaluate atherosclerosis. Similar to brachial FMD, carotid IMT is related to the major CV risk factors and predicts CV events (OÕLeary et al., 1999). Also the echogenicity of the intima-media complex, evaluated as the Grey scale median (IM-GSM), might be of importance since IM-GSM is related to the echogenicity in overt plaques (Lind et al., 27), as well as to markers of lipid oxidation and is reduced in patients with stroke (Lind et al., 28). In the present study, aiming to assess the relationship between the characteristics of hyperemic blood flow velocities in the brachial artery and the ultrasonographic findings of the carotid arteries, we used data from the PIVUS study, hypothesising that the above mentioned SDFV ratio is related to the presence of carotid plaque and echogenicity, as well as to IMT and IM-GSM. 1

2 Brachial blood flow velocity and atherosclerosis, S. J. Järhult et al. Material and methods Participants This study used data from 116 seniors included in the PIVUS. All subjects were recruited by the register of community living and were invited within 2 months of their 7th birthday. The participation rate was 5Æ1% (116 225). The study was approved by the Ethics Committee of the University of Uppsala. Participants completed a questionnaire regarding their medical history, regular medication and smoking habits. Approximately 1% of the cohort reported a history of coronary heart disease, 4% reported stroke and 9% diabetes mellitus. Almost half the cohort reported any CV medication (45%), with antihypertensive medication being the most prevalent (32%). Fifteen percent reported use of statins, while use of insulin and oral antiglycemic drugs were reported by 2 and 6%, respectively (Lind et al., 28). For means see Table 1 and regarding distributions see Fig. 1. Methods The assessment of FMD in the brachial artery has been increasingly used in research to evaluate functional properties of the vasculature (Corretti et al., 22). When examined in the (a) 35 Count 3 25 2 15 1 5 (b) 4 Count 2 4 6 8 1 1 2 1 4 1 6 1 8 2 2 2 2 4 Systolic mean velocity during hyperemia (m s 1 ) 35 3 25 2 15 1 Table 1 Basic characteristics and major cardiovascular risk factors in the prospective investigation of the vasculature in Uppsala seniors sample. n 116 Females (%) 5Æ2 Height (cm) 169 (9Æ1) Weight (kg) 77 (14) Waist circumference (cm) 91 (12) BMI (kg m )2 ) 27Æ (4Æ3) Waist hip ratio Æ9 (Æ75) SBP (mmhg) 15 (23) DBP (mmhg) 79 (1) Heart rate (beats per min) 62 (8Æ7) Serum cholesterol (mmol l )1 ) 5Æ4 (1Æ) LDL cholesterol (mmol l )1 ) 3Æ3 (Æ88) HDL cholesterol (mmol l )1 ) 1Æ5 (Æ42) Serum triglycerides (mmol l )1 ) 1Æ3 (Æ6) Fasting blood glucose (mmol l )1 ) 5Æ3 (1Æ6) Current smoking (%) 11 IMT (mm) Æ89 (Æ16) IM-GSM 74Æ (31Æ6) Systolic mean velocity 1Æ17 (Æ25) Diastolic mean velocity Æ56 (Æ15) SDFV ratio 2Æ16 (Æ45) Blood flow increase (%) 6 (1Æ85) SBP, systolic blood pressure; DBP, diastolic blood pressure; BMI, body mass index; LDL, low density lipoproteins; HDL, high density lipoproteins; IMT, intima-media thickness of carotid artery as measured in far wall from probe; IM-GSM, intima-media grey scale median (given on a scale from 256, see legend); SDFV, systolic to diastolic blood flow velocity during induced reactive hyperemia of the brachial artery. Mean are given with SD in parentheses. 5 (c) 6 Count 2 4 6 8 1 1 2 1 4 Diastolic mean velocity during hyperemia (m s 1 ) 5 4 3 2 1 5 1 1 5 2 2 5 3 3 5 4 4 5 5 SDFV ratio Figure 1 Frequency distributions of systolic (a) and diastolic (b) mean velocity during hyperemia and the hyperemic systolic to diastolic mean blood flow velocity (SDFV) (c). morning, the participants were asked not to smoke or take any medication and to be fasting from midnight. Blood pressure was measured by a calibrated mercury sphygmomanometer to nearest mmhg after at least 3 min of rest in the supine position. The average of three recordings was used. Lipid variables and fasting blood glucose were measured by standard laboratory techniques. The Framingham risk score was calcu-

Brachial blood flow velocity and atherosclerosis, S. J. Järhult et al. 3 lated and used as a comprehensive measure of the major CV risk factors. Brachial artery monitoring The brachial artery was assessed by external B-mode ultrasound (Acuson XP128, 1 MHz linear transducer; Acuson, Mountain View, CA, USA). Imaging was performed 2 3 cm above the elbow for measures of the brachial artery diameter according to the recommendations of the International Brachial Artery Task Force (Corretti et al., 22). To document baseline blood flow velocity, a pulsed Doppler measurement of the brachial artery blood flow velocity was performed. Projection of the sampling volume was in the centre of the vessel. By inflating a blood pressure cuff placed distal to the elbow (to a pressure of at least 5 mm Hg above systolic blood pressure), the brachial artery was occluded, causing ischemia of the forearm. After 5 min, blood flow was re-established by rapidly deflating the cuff, causing a reactive hyperemia. Following cuff-release, blood flow velocity was measured for 15 s. The brachial diameter was measured at rest and at 3, 6 and 9 s following cuff-release. The ultrasonographic measurements were recorded on videotapes for later analysis. Recordings and analyses were performed on the same ultrasound device. Mean blood flow velocities in systole and diastole were measured at rest and after a few seconds following cuff-release. FMD was defined as the maximal diameter seen at either 3, 6 or 9 s following cuffrelease in relation to resting diameter. One individual analysed FMD and the indices of blood flow velocity were analysed by another examiner, by visually measuring an average of at least three different cardiac cycles. In the present study, we used the hyperemic blood flow velocities only, since they were more closely related to CV risk factors than the resting values (Benjamin et al., 24). The mean blood flow velocity was calculated as the area under the curve for the systolic and diastolic phases separately during the hyperemic phase (see Fig. 1). Blood volume flow at rest and during hyperemia was calculated from the mean blood flow velocity and the resting diameter. In accordance with previous investigators, the increment in blood volume flow induced by hyperemia was given as a percentage of resting blood flow, and was denoted the hyperemic blood flow increase. (Celermajer et al., 1992; Corretti et al., 1995). The s of variation for the hyperemic mean blood flow velocities in systole and diastole, the ratio thereof and the hyperemic blood volume flow increase were between 5Æ7 1Æ3% (Jarhult et al., 28). Eightyfour participants were excluded from SDFV calculations. Carotid artery imaging Ultrasonographic imaging of the carotid artery was obtained with the same equipment. The common carotid artery (CCA), the bulb and the internal carotid artery were visualised and the occurrence of plaque was recorded at both sides. The IMT was evaluated in the far wall of the CCA 1 2 cm proximal to the bulb. The given value for carotid artery IMT is the mean value from both sides. For IMT and echogenicity assessment of the carotid arteries, the image was then digitised and imported into the (artery measurement software) (Liang et al., 2; Lind et al., 27). A maximal 1 mm segment with good image quality was chosen for analysis. The borders of the IMT of the far wall and the inner diameter of the vessel were automatically identified by the programme. IMT was calculated through measuring the diameter from around 1 discrete points through the 1 mm long segment. If found not appropriate at visual inspection, this automated analysis could be manually corrected. An ultrasonographic region of interest (ROI) was depicted manually around the intima-media segment evaluated for IMT. IM-GSM was calculated by analysis of the individual pixels within the ROI on a scale from (black) to 256 (white). The blood was used as a reference for black and the adventitia of the carotid vessel for white. A ROI was also placed manually around plaques for measurement of plaque area and GSM. Plaque size was graded into four groups according to a previously used classification (Schmidt et al., 25). If the IMT was locally thickened more than 5% compared to the surrounding IMT, a plaque was considered small. A plaque was denoted moderate if the plaque area was more than 1 mm 2. Plaques were regarded as flow-limiting if the velocity was increased distally of the plaque. Occluded carotid arteries were also included in this latter group. The mean length of the evaluated segment was 6Æ9 (SD 1Æ9) mm when subjects with a segment recording <3 mm were excluded, leaving 945 subjects with valid recordings. The s of variation were 7Æ2% for carotid artery IMT, 7Æ5% for IM-GSM and 8Æ3% for GSM in plaque (Lind et al., 27). Statistical analysis The relations between brachial artery blood flow variables and carotid artery arteriosclerosis variables were evaluated by multiple regression analyses or ANOVA, adjusting for gender and Framingham risk score. Two-tailed significance values were given with P<Æ5 regarded as significant. StatView (SAS Inc., Cary, NC, USA) was used for calculations. Results Relations of flow velocity parameters to number of carotid arteries with plaque The SDFV was directly related (P =Æ18) to the number (, 1, 2) of carotid arteries affected by plaque, adjusting for gender and Framingham risk score. None of the other parameters analysed (mean velocity in systole and diastole respectively, or the blood flow increase from rest to hyperemia) was significantly related to the number of arteries affected with carotid plaque (Table 2).

4 Brachial blood flow velocity and atherosclerosis, S. J. Järhult et al. Number of carotid arteries with plaque n = 296 1 n = 299 2 n = 256 ANOVA Table 2 Relations of blood flow velocity parameters to number of carotid arteries with plaque. Systolic mean velocity 1Æ16 (Æ24) 1Æ18 (Æ24) 1Æ17 (Æ27) a: Æ48 b: Æ68 Diastolic mean velocity Æ57 (Æ14) Æ56 (Æ15) Æ55 (Æ18) a: Æ51 b: Æ7 SDFV ratio 2Æ1 (Æ4) 2Æ18 (Æ45) 2Æ23 (Æ48) a: Æ5 b: Æ2 Blood flow increase (%) 6Æ11 (1Æ77) 6Æ7 (1Æ84) 5Æ81 (1Æ91) a: Æ3 b: Æ9 SDFV, the hyperaemic systolic to diastolic mean blood flow velocity. ANOVA a: adjusted for gender only, b: adjusted for gender and Framingham risk score. Mean are given with SD in parentheses. Relations of flow velocity parameters to carotid plaque size A positive relation of SDFV to plaque size was observed (P = Æ35), adjusting for gender and Framingham risk score. None of the other parameters analysed (mean velocities in systole and diastole or the blood flow increase from rest to hyperemia) were significantly related to plaque size (Table 3). Relations of flow velocity parameters to plaque GSM, IM-GSM and IMT The mean blood velocity during the systolic phase of hyperemia was positively related to IMT (P = Æ9) but not significantly to plaque GSM or IM-GSM. The diastolic mean velocity during hyperemia was positively related to both plaque GSM (P =Æ4) and IM-GSM (P =Æ5), but not to IMT (see Table 4). The SDFV was significantly positively related to plaque GSM (P =Æ2), IM-GSM (P =Æ1) and to IMT (P =Æ22), adjusting for gender and Framingham risk score. The hyperemic blood flow increase was positively related to IM-GSM (P =Æ32), but not to GSM in plaque or IMT. IMT and the IM-GSM were both significantly related to the SDFV even after adjustment for the traditional CV risk factors used for calculating the Framingham Risk Score in two separate multiple regression models (see Table 5 for details). Discussion Summarising our results, one single parameter, the SDFV ratio, was consistently related to several indices of arteriosclerosis of the carotid artery. The SDFV ratio was related to the number of carotid arteries with plaques, as well as to plaque size and echogenicity. Furthermore, this ratio was also related to IMT and IM-GSM, independently of traditional CV risk factors. Cardiovascular disease (CVD) is the cause of around 3 per cent of the deaths world-wide and as high as 5% in the US and Europe (http://www.americanheart.org, http://www.who.int). Morbidity after CVD is a further challenge, personally and for the society, at an unestimatable supportive cost. Lean methods for Table 3 Blood flow velocity parameters versus carotid artery plaque size (no plaque, small, medium or large plaque). No plaque n = 35 Small plaques (<1 mm 2 ) n = 133 Worst plaque size Medium sized plaques (>1 mm 2 ) n = 398 Flow-limiting plaque n =26 ANOVA Systolic mean velocity 1Æ16 (Æ24) 1Æ18 (Æ24) 1Æ17 (Æ26) 1Æ19 (Æ27) a: Æ58 b: Æ78 Diastolic mean velocity Æ57 (Æ14) Æ57 (Æ16) Æ55 (Æ16) Æ56 (Æ17) a: Æ52 b: Æ74 SDFV ratio 2Æ1 (Æ33) 2Æ13 (Æ41) 2Æ21 (Æ46) 2Æ28 (Æ67) a: Æ3 b: Æ2 Blood flow increase (%) 6Æ9 (1Æ77) 6Æ16 (2Æ7) 5Æ93 (1Æ81) 5Æ36 (1Æ85) a: Æ7 b: Æ29 SDFV, the hyperemic systolic to diastolic mean blood flow velocity. ANOVA a: adjusted for gender only, b: adjusted for gender and Framingham risk score. Mean are given with SD in parentheses.

Brachial blood flow velocity and atherosclerosis, S. J. Järhult et al. 5 Table 4 Relationship between blood flow velocity variables and plaque echogenicity, echogenicity in the intima-media complex IM-GSM and IMT. Plaque echogenicity IM-GSM IMT Systolic mean velocity )Æ4 a: Æ23 b: Æ3 Diastolic mean velocity )Æ12 a: Æ5 b: Æ4 SDFV ratio Æ15 a: Æ1 b: Æ2 Blood flow increase (%) Æ8 a: Æ5 b: Æ7 )Æ3 a: Æ26 b: Æ25 )Æ12 a: Æ5 b: Æ5 Æ15 a:<æ1 b: Æ1 Æ1 a: Æ4 b: Æ3 Æ12 a:<æ1 b: Æ9 Æ4 a: Æ62 b: Æ9 Æ1 a:<æ1 b: Æ2 )Æ6 a: Æ5 b: Æ1 SDFV, the hyperemic systolic to diastolic mean blood flow velocity; IMT, intima-media thickness; IM-GSM, intima-media grey scale median. s are given following adjustment for a: gender and b: gender and major risk factors (Framingham risk score). detecting individuals at risk for CVD is therefore wanted. Screening of traditional risk factors (hyperlipidemia, hypertension, diabetes, smoking etc.) accompanied by several methods, such as ultrasonographic imaging, aims at finding subjects at high CV risk. The fact that FMD of the brachial artery is impaired before anatomical evidence or clinical symptoms of atherosclerosis are detected is well known since the early nineties (Celermajer et al., 1992). Non-invasive imaging by ultrasonography provides a possibility to, with a high accuracy, in detail assess the arterial wall properties and blood flow dynamic changes in response to various stimuli. In a previous publication we showed that the hyperemic blood flow in systole in the brachial artery was directly related to coronary risk, whereas higher diastolic blood flow velocity during hyperemia indicated a lower risk for CVD Table 5 Two different multiple regression models with IMT and IM- GSM as dependent variables in the two separate models with traditional cardiovascular risk factors and the SDFV as independent variables in both models. IMT IM-GSM SDFV Æ12 Æ6 Æ18 <Æ1 Gender )Æ8 Æ28 )Æ4 Æ26 HDL )Æ11 Æ3 Æ13 Æ2 LDL Æ9 Æ4 Æ5 Æ15 Diabetes Æ7 Æ46 )Æ4 Æ26 Hypertension Æ14 <Æ1 )Æ9 Æ8 Current smoking Æ9 Æ4 )Æ1 Æ76 (Jarhult et al., 28). We also found that the SDFV ratio was more closely related to coronary risk than the systolic and diastolic components separately (see Fig 2). In the present study, neither the systolic, nor the diastolic velocity was consistently related to arteriosclerosis while the ratio was, suggesting this ratio as a possible marker for atherosclerosis. SDFV ratio and atherosclerosis The systolic hyperemic blood flow velocity possibly reflects, at least in part, the elastic properties of the conduit arteries since we have found the systolic velocity to be related to systolic blood pressure, a marker of arterial stiffness in the elderly. The diastolic hyperemic velocity, on the other hand, possibly reflects the functional properties of the smaller arteries, since we found it to be mainly related to diabetes, known to induce small-vessel disease. Thus, the SDFV ratio possibly combines information on properties of both large and small-vessel status and will thereby serve as an integrated index of vascular function. The finding that this ratio was consistently related to different indices of atherosclerosis independently of CV risk factors further supports the usefulness of this ratio. Presence of carotid plaque, plaque size and echolucency (echolucent plaque indicating higher risk) has been shown to be indicative of future CVD. SDFV, the hyperemic systolic to diastolic mean blood flow velocity; IMT, intima-media thickness in the far wall of the carotid artery; IM-GSM, intima-media grey scale median in the far wall of the carotid artery; LDL, low density lipoproteins; HDL, high density lipoproteins. Diabetes was defined as previous diagnosis or diabetic fasting blood glucose at examination. Hypertension was defined as blood pressure above 14 9 mmhg or antihypertensive treatment. Figure 2 Schematic illustration of the measurement of blood flow velocity during hyperemia. The mean blood flow velocity was calculated as the area under the curve for the systolic and diastolic phases separately. The arrows indicate the maximal velocities, but the maximal velocity is not used in the present study.

6 Brachial blood flow velocity and atherosclerosis, S. J. Järhult et al. The SDFV ratio was related to all of these three plaque characteristics. The echogenicity of the CCA intima-media complex has been described to be in close correlation with the echogenicity of overt plaques (Lind et al., 27). The SDFV ratio was found in the present study to be related to IM-GSM, as well as to IMT, further emphasising the link between this ratio and atherosclerosis and plaque composition. It is however not known if a high SDFV ratio will cause atherosclerosis by mechanical influences on the vascular wall, or if atherosclerosis would lead to changes in the blood flow velocity pattern. Only a longitudinal approach with repeated measurements of the SDFV ratio and atherosclerosis would give the answers to that question. Limitations The present sample is limited to Caucasians aged 7. Caution should therefore be made to draw conclusions to other ethnicand age- groups. The PIVUS study had a moderate participation rate. However, an analysis of non-participants showed the present sample to be fairly representative of the total population regarding most CVD s and drug intake (Lind et al., 26). In the present study several statistical tests have been performed. Since the study was explorative in its character, we did not apply any strict correction for multiple testing and thereby positive chance findings may have occurred. However, the SDFV ratio was related to all five evaluated indices of atherosclerosis in a consistent way, possibly indicating that this is not a chance finding. In conclusion, our results indicate that SDFV ratio in the brachial artery is related to atherosclerosis in the carotid artery. Acknowledgments The outstanding work at the endothelium laboratory performed by Nilla Fors, Jan Hall, Kerstin Marttala and Anna Stenborg is highly acknowledged. Uppsala University hospital, Thureus foundation and AstraZeneca R&D Mölndal have kindly supported the study. References Anderson TJ, Uehata A, Gerhard MD, Meredith IT, Knab S, Delagrange D, Lieberman EH, Ganz P, Creager MA, Yeung AC. Close relation of endothelial function in the human coronary and peripheral circulations. J Am Coll Cardiol (1995); 26: 1235 1241. Benjamin EJ, Larson MG, Keyes MJ, Mitchell GF, Vasan RS, Keaney JF Jr, Lehman BT, Fan S, Osypiuk E, Vita JA. Clinical correlates and heritability of flow-mediated dilation in the community: the Framingham Heart Study. Circulation (24); 19: 613 619. Celermajer DS, Sorensen KE, Gooch VM, Spiegelhalter DJ, Miller OI, Sullivan ID, Lloyd JK, Deanfield JE. Non-invasive detection of endothelial dysfunction in children and adults at risk of atherosclerosis. Lancet (1992); 34: 1111 1115. Corretti MC, Plotnick GD, Vogel RA. Technical aspects of evaluating brachial artery vasodilatation using high-frequency ultrasound. Am J Physiol (1995); 268(4 Pt 2): H1397 H144. Corretti MC, Andersson TJ, Benjamin EJ, Celermajer D, Charbonneau F, Creager MA, Deanfield J, Drexler H, Gerhard-Herman M, Herrington D, Vallance P, Vita J, Vogel R. Guidelines for the ultrasound assessment of endothelial-dependent flow-mediated vasodilation of the brachial artery: a report of the International Brachial Artery Reactivity Task Force. J Am Coll Cardiol (22); 39: 257 265. Jarhult SJ, Hall J, Lind L. Hyperaemic blood-flow velocities in systole and diastole relate to coronary risk in divergent ways. Clin Physiol Funct Imaging (28); 28: 189 195. Liang Q, Wendelhag I, Wikstrand J, Gustavsson T. A multiscale dynamic programming procedure for boundary detection in ultrasonic artery images. IEEE Trans Med Imaging (2); 19: 127 142. Lind L, Fors N, Hall J, Marttala K, Stenborg A. A comparison of three different methods to determine arterial compliance in the elderly: the Prospective Investigation of the Vasculature in Uppsala Seniors (PI- VUS) study. J Hypertens (26); 24: 175 182. Lind L, Andersson J, Ronn M, Gustavsson T. The echogenecity of the intima-media complex in the common carotid artery is closely related to the echogenecity in plaques. Atherosclerosis (27); 195: 411 414. Lind L, Andersson J, Ronn M, Gustavsson T, Holdfeldt P, Hulthe J, Elmgren A, Zilmer K, Zilmer M. Brachial artery intima-media thickness and echogenicity in relation to lipids and markers of oxidative stress in elderly subjects: the prospective investigation of the vasculature in Uppsala Seniors (PIVUS) Study. Lipids (28); 43: 133 141. OÕLeary DH, Polak JF, Kronmal RA, Manolio TA, Burke GL, Wolfson SK Jr. Carotid-artery intima and media thickness as a risk factor for myocardial infarction and stroke in older adults. Cardiovascular Health Study Collaborative Research Group. N Engl J Med (1999); 34: 14 22. Schmidt C, Fagerberg B, Hulthe J. Non-stenotic echolucent ultrasoundassessed femoral artery plaques are predictive for future cardiovascular events in middle-aged men. Atherosclerosis (25); 181: 125 13. Yeboah J, Crouse JR, Hsu FC, Burke GL, Herrington DM. Brachial flowmediated dilation predicts incident cardiovascular events in older adults: the Cardiovascular Health Study. Circulation (27); 115: 239 2397.