The Relationship Between Flow‐Mediated Dilatation of the Brachial Artery and Intima‐Media Thickness of the Carotid Artery to Framingham Risk Scores in Older African Americans

Abstract

The objective of this study was to investigate the relationship of flow‐mediated dilatation and intima‐media thickness (IMT) with coronary risk in African Americans (AAs). Endothelial dysfunction and IMT of carotid arteries are considered early steps in atherosclerotic disease process and have been used as surrogate markers of subclinical atherosclerosis. Data were collected on 106 AAs with a mean age of 64.0±6.6 years. Carotid artery IMT was measured with B‐mode ultrasonography, as was brachial artery diameter at rest, during reactive hyperemia, and after nitroglycerin. Percent change in flow‐mediated dilatation (%FMD) was defined as 100×(diameter during reactive hyperemia – resting diameter)/resting diameter. Percent change in nitroglycerin‐mediated dilatation (%NMD) was defined as 100×(diameter with nitroglycerin‐resting diameter)/resting diameter. The Framingham 10‐year risk score (FRS) was calculated for each patient using the National Cholesterol Education Program (NCEP) risk score calculator and participants were categorized into 3 groups with FRS as <10%, 10% to 20%, and >20%. Thirty‐eight participants had risk scores <10%, 26 had 10% to 20%, and 42 >20%. There was a significant inverse relation between %FMD and FRS (P<.0001) and between %NMD and FRS (P<.001). IMT was not statistically different among the risk groups. Endothelial dysfunction assessed by FMD significantly correlates inversely with FRS in AAs. FMD, an index of arterial compliance, appears to be a sensitive and reliable index of cardiovascular disease.

Atherosclerotic cardiovascular disease (CVD) is a major cause of morbidity and mortality in America and the effect is disproportionally higher in African Americans (AAs). Minimizing the disproportionately higher morbidity and mortality from CVD in AAs requires identification and addressing the causes of excess mortality. Early identification of the disease may help to understand the magnitude and severity of the problem, the progression of the disease, and the effect of intervention, particularly in populations with disproportionately higher morbidity and mortality rates.

Endothelium‐dependant vasodilatation (EDV) can be induced in peripheral arteries by producing ischemia. The term endothelial dysfunction refers to impaired EDV. Measurement of vasodilation produced by ischemia (also termed flow‐mediated vasodilatation [FMD]) can be determined with high‐resolution B‐mode ultrasound. This noninvasive method has been used to assess brachial artery endothelial function. ¹ , ² Atherosclerotic risk factors cause endothelial dysfunction by increasing oxidative stress. ³ Higher oxidative stress and impaired forearm EDV has been noted in AAs compared with whites. ¹ , ² , ³ , ⁴ , ⁵ , ⁶ , ⁷ , ⁸ , ⁹ , ¹⁰ Endothelial dysfunction assessed by FMD of the brachial artery ¹¹ , ¹² , ¹³ , ¹⁴ and intima‐media thickness (IMT) of the carotid artery ¹⁵ , ¹⁶ , ¹⁷ are considered to be early changes of atherosclerotic vascular disease and have been used in epidemiologic studies as a surrogate marker of subclinical atherosclerosis.

The coronary artery disease (CAD) risk of an individual is the result of the global atherosclerotic risk burden. The Framingham 10‐year CAD risk score (FRS) is an important tool to assess the global risk burden of an individual ¹⁸ , ¹⁹ and its validity in different races has been reported. ²⁰ In this study, we used high‐resolution B‐mode ultrasonography to measure FMD in the brachial artery and IMT of carotid artery to assess their relation to the FRS.

Methods

The study population consisted of AAs 55 years or older who participated in a study on the effect of herbal treatment on cardiovascular disease.

The inclusion criteria were age 55 years or older, presence of ≥1 CAD risk factors, and absence of clinical atherosclerosis. The study participants were recruited from the outpatient clinic at Howard University Hospital. A total of 106 patients were included in the study. Each study patient was identified by a unique identification number.

The medical history and physical examination were reviewed and the entry risk factors identified. Age, sex, weight, and height were recorded and body mass index (BMI) calculated. Blood pressure (BP) was measured in the left arm 3 times at 30‐second intervals after at least 5 minutes of rest, with the patient in a sitting position. The averaged BP and a lipid profile determined in a fasting state were recorded. The FRS was calculated using baseline CVD risk factors and the risk factor calculator available at the National Cholesterol Education Program (NCEP) Web site. The FRS was used to characterize participants into 3 risk groups with a <10%, 10% to 20%, and >20% probability of developing CAD in 10 years. Even though not all diabetes patients fall in the FRS risk of >20%, ¹⁸ the participants in this study were older and most also had other concomitant CAD risk factors and therefore were included in the risk group with a >20% risk score.

Flow‐ and Nitroglycerine‐Mediated Dilatation of Brachial Artery

Brachial artery reactivity testing (BART) was performed using high‐resolution B‐mode images of the right brachial artery to measure the change in diameter. The right brachial artery was positioned horizontal and the transducer was directed perpendicular to the vessel walls to image the double lining (lumen/intima‐media‐adventitia) in the near and far walls. The images were acquired with the participant in a supine position with the right arm fully extended on the examination table facing the cubical area. The tourniquet cuff was placed on the arm above the location where scanning takes place. BP was measured on the left arm before beginning scanning. Two‐dimensional (2D) images were optimized before the procedure started. When the near and far walls of the vessel were clearly visualized over a segment of at least 10 mm, the images were recorded on videotape. The prehyperemia, ischemic condition, in the right arm was then induced by increasing the pressure of the BP/tourniquet cuff to at least 50 mm Hg above the systolic BP. After 5 minutes of prehyperemia ischemia, the tourniquet cuff pressure was released and 2D images of the brachial artery were recorded in a similar fashion. After the recording of the hyperemia phase was complete, the participants rested for 10 minutes. To study the relation of endothelium‐dependent (brachial artery reactivity, ischemic‐mediated) and endothelium‐independent (nitroglycerine‐mediated) brachial artery dilatation to the FRS, a 0.4 mg nitroglycerin tablet was administered sublingually and 2D images of the right brachial artery were recorded on videotape. All images were digitized on a video frame grabber and installed on a personal computer. The percent change in brachial artery diameter during reactive testing relative to the resting diameter was calculated as follows:

Similarly, the percent change in brachial artery diameter following nitroglycerine administration relative to the resting diameter was calculated as:.

Analysis of the images was performed with Prosound quantitative software (University of Southern California, Los Angeles, CA).

Carotid IMT

Carotid artery IMT was measured by B‐mode ultrasonography. The distal 1 cm of the near and far walls of the right common carotid artery was studied with participants in the supine position and the head rotated away from the side examined. IMT was defined as the distance between interfaces 2 and 3 (near wall) and 4 and 5 (far wall) on the longitudinal view. Frozen images, at the site of maximal wall thickness or stenosis in anterior and lateral positions and short segment of real‐time scanning with continuous electrocardiography, were recorded on super VHS videotape. For analysis of wall thickness and percent luminal stenosis, the images were transferred to a computer system where they were digitized and stored on disk. For image analysis, studies were identified by study code and examined in a blinded fashion using a fully automated, computerized, edge‐detection tracking method. Carotid IMT images were digitized on a video frame grabber, installed on a personal computer, and analyzed with a data translation (DT 2862) image processing board and Prosound system.

Statistical Analysis

Univariate summaries of clinical and demographic characteristics included means and standard deviations or frequencies and percents, as appropriate. Pearson correlation coefficient was used to examine the relationship of selected variables. One‐way analysis of variance (ANOVA) was used to separately determine differences in means of IMT, FMD, and nitroglycerin‐mediated dilatation (NMD) among the 3 FRS groups. Bonferroni’s post hoc analysis was applied for pairwise comparison whenever the main effect of ANOVA was significant. The independent effect of FRS on IMT, percent of FMD (%FMD), and percent of NMD (%NMD) was separately examined in regression models that adjust for age, sex, and relevant clinical risk factor. Clinical factors used for covariates adjustment included BMI, dyslipidemia, hypertension, diabetes mellitus, smoking status, pulse pressure, and use of antihypertensives and statins. To test for potential collinearity of variables that may affect IMT or %FMD, variance inflation factor (VIF), were invoked in these regression models. In all regression diagnostics, after examination of the VIF, any variable with the condition of collinearity with other risk factors was excluded. Total cholesterol was the only variable with such condition. All statistical analysis was performed using SAS software version 13 (SAS Institute, Inc, Cary, NC). A P value <.05 was considered statistically significant. Data are generally presented as mean ± standard deviation.

Results

The mean age of the 106 AAs participating in this study was 64.0±6.60 years and consisted of 51.9% women. The mean BMI was 31.43±6.18 kg/m2. The Table shows the clinical and demographic characteristics of patients. Hypertension was present in 74 (70.1%), dyslipidemia in 87 (82.1%), and type 2 diabetes mellitus in 38 (35.51%) patients. Eighteen percent of the study participants were smokers. Sixty‐nine (65%) were taking antihypertensive treatment, and 30 (28%) were taking statins. Thirty‐eight (36%) were taking neither antihypertensive medication nor statins. Thirty‐eight (35.85%) of the study participants had a risk score of <10%, 26 (24.53%) had risk scores of 10% to 20%, and 42 (39.62%) had scores >20%.

Table.

 Baseline Characteristics

Characteristics	Framingham 10‐Y CAD Risk Score
	<10% (n=38)	10%–20% (n=26)	>20% (n=46)	All (N=106)
Age, y	62.5±6.2	64.6±6.6	65.9±6.7	64.0±6.6
Females, No. (%)	32 (84.21)	3 (11.5)	20 (47.62)	55 (51.9)
Body mass index, kg/m2	30.1±5.15	31.7±6.7	32.41±6.6	31.43±6.2
Hypertension, %	57.9	69.2	81.4	70.09
Diabetes mellitus, %	0	0	90.5	35.51
Dyslipidemia, %	78.4	88.5	81.4	82.1
Smoking, %	18.9	23.1	14.0	17.9
SBP, mm Hg	122.7±10.9	131.71±15.5	132.12±14. 34	128.70±14.12
DBP, mm Hg	76.7±6.9	80.8±7.9	73.7±10.7	76. 50±9.19
Total cholesterol, mg/dL	216.0±43.2	200.7±33.1	187.1±48.7	200.7±44.8
HDL, mg/dL	60.74±12.5	49.58±19.63	45.67±11.2	51. 97±15.5
LDL, mg/dL	137.1±37.8	129.04±27.5	116.23±41.8	126.75±38.2
Triglycerides, mg/dL	90.7±32.3	109.5±51.7	126.26±95.8	109.6±69.8
IMT, mm	0.87±0.13	0.93±0.15	0.93±0.18	0.91±0.16
%ΔFMD	8.72±5.5	6.3±4.5	3.50±4.14	6.03±5.22
%ΔNMD	12.8±7.02	7.6±4.7	8.3±6.1	9.72±6.51
B1, mm	3.5±0.82	4.3±1.9	3.8±0.84	3.8±0.91

Abbreviations: B1, brachial artery resting diameter; DBP, diastolic blood pressure; HDL, high‐density lipoprotein; IMT, intima‐media thickness; LDL, low‐density lipoprotein; %ΔFMD, percent change flow mediated dilatation; %ΔNMD, percent change nitroglycerine mediated dilatation; SBP, systolic blood pressure. Values are expressed as mean ± standard deviation unless otherwise indicated.

A decrease in NMD was observed in 2 participants. Moreover, a decrease in FMD was observed in 10 participants and all of these patients were diabetics. However, the NMD in these 10 diabetic patients was normal.

In univariate analysis, FRS correlated negatively with %FMD (r=−0.41, P<.001) and %NMD (r=−0.31, P=.001), and there was no correlation with IMT (r=0.03, P=.71). Results also showed a correlation between %FMD and %NMD (r=0.22, P<.02). There was no correlation between IMT and FMD (r=0.06, P=.52) nor %NMD (r=0.09, P=.35).

The Table shows the mean percent changes of FMD, NMD, and IMT with respect to risk score category. Results of ANOVA showed a significant decrease in %FMD and %NMD with increasing risk score category and are depicted in Figure 1. As seen in the Figure 1, mean %FMD was smaller than mean %NMD in all risk groups. There was no significant difference in IMT among the 3 risk categories (Figure 2).

Figure 1.

 Percent change of brachial artery diameter following brachial artery reactivity testing (percent change in flow‐mediated dilatation [%FMD]) and following nitroglycerin administration (percent change in nitroglycerin‐mediated dilatation [%NDM]) per Framingham Risk Score category. CAD indicates coronary artery disease.

Figure 2.

 Carotid intima‐media thickness (IMT) per Framingham Risk Score category. CAD indicates coronary artery disease.

In multivariate regression analysis, %FMD correlated negatively with FRS (P<.01) when diabetes mellitus is not in the model. The independent association was, however, lost when diabetes mellitus was in the model. This is explained in part by the fact that all the diabetics were in the higher‐risk score category.

In addition, %NMD correlated significantly with age (P<.01), female sex (P<.01), and presence of hypertension (P<.01). There was no independent correlation of %NMD with FRS (P=.30). IMT correlated independently with age (P<.01) and dyslipidemia (P=.02). Results of the regression analysis, however, showed no relationship of FRS with IMT.

Discussion

The study participants were older AAs who varied in their global CAD risk. Endothelial dysfunction, indicated by impaired FMD, has been noted to be more prevalent in AAs compared with whites. ⁵ , ⁶ In the current study, the percentage increase in FMD is inversely related to the FRS.

This inverse relation between FMD and FRS indicates that the degree of endothelial dysfunction is associated with the total coronary risk burden of an individual. This is important because it confirms the need for global CAD risk assessment and the use of lifestyles and/or treatments known to preserve or improve endothelial dysfunction.

Nitroglycerin (donor of NO) was used to evaluate the NMD as a control to assess whether the vascular smooth muscle dysfunction is the cause of impaired FMD. The mean percentage of NMD for the entire study population was higher than the mean percent change of ischemic FMD. This indicates that vascular smooth muscle dysfunction alone cannot explain the reduced FMD and, hence, impaired endothelium‐dependant NO‐mediated dilatation is largely responsible for the reduced FMD.

However, the percentage of NMD correlated with the percent of FMD and inversely with the FRS. This indicates the possibility of concomitant vascular smooth muscle dysfunction. Adams and colleagues ²¹ studied the vasodilator response to 400 μgm of nitroglycerin in 800 asymptomatic patients with atherosclerotic risk and showed an independent relation between reduced endothelium‐dependant dilatation and reduced vasodilator response to nitroglycerin, indicating a concomitant smooth muscle dysfunction. The FRS consists of different clusters of CVD risk factors, thus it is possible that 2 risk factors, eg, diabetes and hypertension, may impact vascular function in the same direction by NMD and FMD, respectively, but may impact FRS differently. Since different risk factors may effect vascular function in the same direction but contribute unequally to the FRS, this could possibly explain the observed inverse relationship between NMD and FRS. However, further studies are needed to precisely clarify the mechanism(s) relating atherosclerotic risk factors and atherosclerotic vascular disease to vascular endothelial and smooth muscle function.

From a physiologic point of view, both BART and nitroglycerine cause vasodilatation and hence increase in the brachial artery diameter. A decrease in the brachial artery diameter was observed in 10 participants following BART and in 2 participants following nitroglycerine administration. However, all 10 of the participants had diabetes, thus our observation in this high‐risk population may be due to an adverse effect of diabetes on endothelium function, which was greater in the group with diabetes. Further studies are needed to investigate the smaller vasodilator response and the occasional lack of any increase in brachial artery diameter following BART in diabetic patients with a high FRS.

A significant progressive increase in IMT was not observed as the FRS increased. However, with multiple regression analysis, 2 determinants of FRS, age and dyslipidemia, were significantly related to IMT of the carotid artery (P<.0001 and P<.044, respectively). A strong relationship between age and IMT of the carotid artery has been reported previously. ²² , ²³ , ²⁴ , ²⁵ , ²⁶ In this study of elderly AAs with multiple risk factors and a mean age of 64.0±6.60 years, the mean carotid IMT was 0.91±0.16. Veller and colleagues ²⁷ described a plaque incidence in carotid or femoral arteries of 95% if mean carotid IMT exceeded 0.8 mm in a group of symptom‐free healthy volunteers. O’Leary and colleagues ²⁸ measured carotid IMT with high‐resolution ultrasonography in 5858 patients 65 years or older and showed that increases in the thickness of the IMT of the carotid artery were directly associated with an increased risk of myocardial infarction and stroke in older adults without a history of CVD. The increased IMT, significantly related to age and dyslipidemia in the current study, is consistent with the observations of previous studies, which have demonstrated increased IMT to be highly specific for predicting CVD outcomes. However, the evaluation of IMT and FMD in the same high‐risk patients at the same time provided an opportunity to compare an arterial structural (IMT) and functional (FMD) parameter, which may provide different prognostic information about vascular health and for therapeutic options.

This study and others ²⁹ , ³⁰ , ³¹ have documented an association among CVD risk factors, IMT, and FMD. All of these markers are used to predict, identify, and/or quantify the onset, stage, and/or severity of atherosclerosis and to monitor the progression or regression rates of the same. Although IMT appears to be the most reliable with the greatest specificity for identifying the presence of atherosclerosis, FMD appears to be a more sensitive detector of decreasing vascular function and the progression of the atherosclerotic process. The linear significant inverse relation of FMD and the number of risk factors present as the FRS increased is consistent with a synergistic atherosclerotic process among multiple risk factors. The lack of a significant relationship between FRS and individual risk factors, but not when clustered, suggest that the earlier effective risk factor reduction is started, the better. The independent significant relation between diabetes mellitus alone and percent change in FMD in multivariate analysis may indicate that diabetes is a very powerful modulator of the progressive loss of normal vascular function and an earlier marker of atherosclerotic CVD than IMT. These observations also suggest that FMD, an index of arterial compliance, is a sensitive and useful method for tracking changes in the arterial system prior to the development of atherosclerotic structural changes detected by IMT, which is less sensitivity.

The strengths of this work include the following: (1) this is the first study to demonstrate an association of a progressive decrease in endothelial function and an increase in IMT with an increase in the FRS; (2) it demonstrates a difference in sensitivity and specificity of FMD and IMT to characterize changes in vascular function and structural properties in a population at high risk for developing atherosclerotic disease; (3) the study suggests that FMD is sensitive enough to detect changes in vascular function before the development of overt disease and, therefore, that it is possible to monitor the progression or regression of vascular lesions and intervene, if desired, at an early stage; and (4) because endothelial dysfunction can be an early marker of CVD, a leading cause of morbidity and mortality, the observations in this report may lead to earlier and more frequent evaluation of FMD using a more accurate, new e‐tracking technique. ³²

Conclusions

The observations of this study suggest that in spite of the increased disparity of cardiovascular conditions and consequences in this high‐risk population, efforts to reduce cardiovascular risk components earlier in life with subsequent reduction in the FRS may preserve endothelial function and reduce cardiovascular morbidity and mortality.