- Email: [email protected]
Prevalence and predictors of subclinical atherosclerosis among asymptomatic “low risk” individuals in a multiethnic population
Atherosclerosis 197 (2008) 435–442
Prevalence and predictors of subclinical atherosclerosis among asymptomatic “low risk” individuals in a multiethnic population Jasmine Grewal a,b , Sonia Anand a,b , Shofiqul Islam a , Eva Lonn a,b,∗ , On behalf of the SHARE and SHARE-AP Investigators b
a The Population Health Research Institute, McMaster University, Hamilton, Ontario, Canada The Department of Medicine, Division of Cardiology, McMaster University, Hamilton, Ontario, Canada
Received 16 February 2007; received in revised form 8 May 2007; accepted 21 June 2007 Available online 6 August 2007
Abstract Background: Current approaches to cardiovascular (CV) risk assessment have limitations. Subclinical atherosclerosis (SCA) as determined by carotid ultrasound is an independent predictor of myocardial infarction and stroke and can refine CV risk assessment. Objectives: We aimed to determine the prevalence and predictors of SCA in a multiethnic population classified as low risk for coronary heart disease (CHD) events by the Framingham Risk Assessment Model. Methods: We conducted a cross-sectional population study in 1015 Canadian adults of Caucasian European, South Asian, Chinese and Aboriginal ancestry. CHD risk was calculated by the 10-year Framingham Risk Score (FRS). Novel and conventional CHD risk factors were measured and high-resolution carotid ultrasound was performed. SCA was defined as carotid intima media thickness (IMT) ≥75th percentile adjusted for age, sex and ethnicity. Results: Seven hundred and fifty two (74%) participants were classified as low risk by FRS. Of these, 175 (23%) had evidence of SCA. Independent predictors of SCA among low-risk subjects included female sex, systolic blood pressure, and apolipoprotein B. Conclusions: Many individuals classified at low CHD risk by the FRS have SCA and are at increased long-term risk for vascular events. Carotid IMT can identify subjects with SCA, who may benefit from early intervention. © 2007 Elsevier Ireland Ltd. All rights reserved. Keywords: Cardiovascular risk; Carotid IMT; Risk factors; Ethnicity
1. Introduction Cardiovascular (CV) disease is a leading cause of death and disability [1]. For many individuals the first symptom of heart disease is sudden cardiac death or myocardial infarction (MI) [2]. Therefore, there is great interest in identifying asymptomatic individuals at risk, who would be candidates for more intensive, evidence-based medical interventions ∗ Corresponding author at: Hamilton Health Sciences, General Site, 237 Barton Street East, Hamilton, Ontario, Canada L8L 2X2. Tel.: +1 905 526 0970; fax: +1 905 527 5380. E-mail address: [email protected] (E. Lonn).
0021-9150/$ – see front matter © 2007 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.atherosclerosis.2007.06.020
aimed at preventing death and disability from coronary heart disease (CHD) and stroke [3]. The traditional approach to CHD risk assessment is based on identifying and to a certain extent quantifying established CV risk factors. Several algorithms based on this approach are used [4,5]. Among these, the Framingham Risk Assessment Model is the most widely accepted. Although used extensively and generally accepted, this model (as well as other algorithms based on similar approaches) has limitations. It is derived from a white Caucasian population in the US and may be less applicable to other ethnic groups. Family history, abdominal adiposity, inflammation and other factors shown to predict CV risk [6,7] are not incorporated in the Framingham
436
J. Grewal et al. / Atherosclerosis 197 (2008) 435–442
risk score (FRS). Diabetes and smoking are identified only as present or absent, although current evidence supports a continuous relationship between glycemia and tobacco exposure to CHD risk [8,9]. Age is the overriding FRS determinant, ignoring greater inter-individual variation in atherosclerotic burden at older ages and often providing false reassurance at younger ages. Importantly, the FRS only predicts shortterm (10-year) risk, although from a clinical perspective the life-long risk of developing CHD events is equally relevant [10]. Imaging of arteries to identify and quantify the presence of subclinical atherosclerosis (SCA) has been suggested to further refine CV risk assessment [3,4,11]. Among available imaging techniques, measurement of carotid intima-media thickness (IMT) with B-mode ultrasound is a non-invasive, sensitive, and highly reproducible technique for identifying and quantifying atherosclerotic burden. It is a well-validated research tool, increasingly used also in clinical practice [12–14]. Indeed, carotid IMT was shown to accurately represent anatomic structural abnormalities [12,13], that correlate with various classical and emerging CV risk factors and with prevalent CV disease [12–15] and importantly, to be an independent predictor of MI and stroke [16–18]. The American Heart Association (AHA) and the Third Adult Treatment Panel of the National Cholesterol Education Program (NCEP ATP III) have endorsed the use of carotid IMT in CV risk assessment [3,4]. To date, the prevalence and predictors of SCA in a “low risk” multiethnic western population have not been examined.
2. Methods The Study of Health Assessment and Risk in Ethnic groups (SHARE) and the Study of Health Assessment and Risk in Aboriginal Peoples (SHARE-AP) were cross-sectional investigations of atherosclerosis, CV risk factors and CV disease in four ethnic groups living in Canada—Caucasian Europeans, South Asians, Chinese and Aboriginal people. The design and major findings of these studies have been previously reported [19–21]. A brief description follows. 2.1. Study population Between 1996 and 2000 participants were randomly sampled from 4 communities. The Ethics Boards of all participating institutions approved the studies and all participants provided written informed consent. A careful sampling methodology was used, aimed at assembling a true population sample, as previously described [19–21]. For the purposes of this analysis we excluded all participants with known CV disease and/or diabetes. CV disease was defined as CHD (angina by Rose questionnaire, history of MI, silent MI [major Q waves by Minnesota criteria], prior percutaneous coronary intervention or coronary artery bypass graft surgery), or his-
tory of stroke. Diabetes was defined as self-reported history of diabetes, use of oral hypoglycemic drugs or insulin, fasting plasma glucose concentration ≥7.0 mmol/L or plasma glucose ≥11.1 mmol/L on a 2-h oral glucose tolerance test. 2.2. Assessment of risk factors Demographic, lifestyle, personal and family history were obtained using validated questionnaires (translated into Tamil, Punjabi, Hindi and Chinese and including validated ethnic specific food questionnaires) [19]. Regular alcohol use was defined as consumption ≥3 times/week. Fruit and vegetable intake was recorded as servings per day. Individuals were judged to be physically active if they engaged in moderate (walking, cycling) or strenuous exercise (jogging, swimming) for ≥3 h per week. Blood pressure (BP), weight, waist to hip ratio (WHR), fasting plasma lipid and glucose levels, hemoglobin A1C (HbA1C ), lipoprotein (a), apolipoprotein B, homocysteine, C-reactive protein (hsCRP), fibrinogen, plasminogen activator inhibitor 1 (PAI-1), white blood cell count and norepinephrine were measured using validated and reproducible methods described previously [19–21]. 2.3. Carotid ultrasonography SCA was measured using high-resolution ultrasound and standardized and extensively validated scanning and measurement protocols [19–21]. A transverse scan was followed by a longitudinal circumferential scan, aimed at imaging the thickest arterial wall of 12 well defined segments, the near and the far walls of the right and left common carotid (1 cm below the bulb), bulb (1 cm below the flow divider) and internal carotid (1 cm above the flow divider) arteries. Offline measurements were done at the Core Laboratory in Hamilton using custom-built software. A minimum of 3 frames from each pre-specified arterial segment with the thickest IMT were measured. The maximum IMT for each of the 12 segments was selected and the mean maximum IMT was computed for each subject by averaging the segment maximum IMT measurements of all segments visualized. There were 4 sonographers and 2 readers, specifically trained and certified. Between and within sonographer and reader reproducibility was high (intraclass correlation coefficients ≥0.90 and coefficients of variation < 5%). We computed the distribution of the mean maximum carotid IMT for groups of individuals defined by age (in decades), sex and ethnicity. Participants with a mean maximum carotid IMT ≥75th percentile for their age, sex and ethnicity were defined as having SCA (Table 1). 2.4. Cardiovascular risk assessment Framingham sex-specific equations were used to predict the 10-year risk for fatal and non-fatal CHD [4,22]. Participants were classified as low, intermediate and high risk if
J. Grewal et al. / Atherosclerosis 197 (2008) 435–442
437
Table 1 Normative values for carotid IMT in the SHARE cohort in subjects without cardiovascular disease and diabetesa (N = 1015) Age
Caucasian
South Asian
Chinese
Aboriginal
Women
Men
Women
Men
Women
Men
Women
Men
Mean (S.D.) 35–45 46–55 56–65 66–75
0.63 (0.11) 0.70 (0.12) 0.80 (0.18) 0.90 (0.22)
0.66 (0.08) 0.74 (0.12) 0.92 (0.23) 1.12 (0.37)
0.60 (0.11) 0.69 (0.15) 0.84 (0.30) 0.90 (0.38)
0.63 (0.12) 0.73 (0.15) 0.82 (0.18) 0.92 (0.15)
0.59 (0.07) 0.65 (0.08) 0.70 (0.07) 0.86 (0.25)
0.64 (0.11) 0.71 (0.10) 0.80 (0.14) 0.90 (0.24)
0.61 (0.09) 0.71 (0.13) 0.89 (0.26) 1.08 (0.26)
0.67 (0.14) 0.78 (0.19) 0.98 (0.32) 0.95 (0.18)
75% Percentile 35–45 46–55 56–65 66–75
0.68 0.74 0.88 0.97
0.70 0.81 0.97 1.33
0.63 0.75 0.92 1.10
0.68 0.80 0.95 0.99
0.63 0.68 0.75 1.00
0.68 0.76 0.86 1.16
0.67 0.79 0.98 1.23
0.73 0.89 1.11 1.18
a The presence of carotid IMT ≥75% percentile for subject’s age, sex and race/ethnicity are indicative of increased CV risk and may signify the need for aggressive risk-reduction interventions. The values should only be applied if utilizing the carotid ultrasound imaging and reading protocols outlined in Section 2.
their CHD risk was calculated at <10%, 10% and 20% and >20%, respectively. 2.5. Statistical analysis Characteristics of subjects in the low versus the medium/high FRS groups and between subjects with and without SCA were compared using ANOVA or logistic regression, as appropriate. Multivariate linear and logistic regression with backward elimination was used to determine independent predictors of carotid IMT (entered as a continuous variable in the model) and of SCA (defined as carotid IMT ≥75th percentile for their age, sex and ethnicity), respectively. Independent variables were selected on the basis of their clinical importance as CV risk factors and their significance in bivariate analyses (p ≤ 0.10). Because carotid IMT, lipoprotein (a) and hs-CRP measurements were not normally distributed, logarithmic transformation was used for all computations involving these variables. White European ethnicity was used as the referent population in determining the influence of ethnicity on carotid IMT (as a continuous variable) and of SCA. Statistical significance was defined as two-sided p < 0.05. All statistical analyses were performed using SPSS version 12.0 (SPSS Inc, Chicago, IL).
3. Results 3.1. Study participants The SHARE and SHARE-AP studies enrolled 1285 participants. CV disease was present in 172, diabetes in 88 (CV disease and diabetes were more common in men and Aboriginal people [20,21]), and carotid IMT recordings were suboptimal in 10, leaving a total of 1015 subjects for this analysis. These were 484 men and 531 women, 285 of Caucasian European ethnicity, 258 South Asians, 284 Chi-
nese and 188 Aboriginals. Of these 60 were classified as high-risk, 203 as intermediate risk and 752 as low-risk by FRS. In this cohort of subjects without CV disease and diabetes, 60% of men (291 of 484) and 87% of women (461 of 531) were classified as low risk; also, 75% of participants of Caucasian European ethnicity (213 of 285), 72% of South Asians (187 of 258), 78% of Chinese (223 of 285), and 69% of Aboriginals (129 of 188), were classified as low-risk (p = 0.10). As expected, those in the medium/highrisk groups were older, predominantly male, more likely to smoke and had higher levels of conventional and novel CV risk factors, such as blood pressure, cholesterol, fasting plasma glucose, waist to hip ratio (WHR), apolipoprotein B, PAI-1, fibrinogen, homocysteine, hs-CRP and had higher carotid IMT. There were no significant differences between the low and high-risk groups in lifestyle factors (alcohol consumption, physical activity, fruit and vegetable intake) (Table 2). 3.2. Prevalence of subclinical atherosclerosis among low-risk subjects Among the 752 low-risk individuals, SCA (carotid IMT ≥75th percentile for age, sex and ethnicity) was present in 175 (23%), 20% of men (58 of 291) and 25% of women (117 of 461), 25% participants of European ethnicity (53 of 213), 22% of South Asians (41 of 187), 24% of Chinese (55 of 223) and 20% of Aboriginals (26 of 129) (p = 0.66). Characteristics of low-risk subjects with and without SCA are shown in Table 2. Those with SCA tended to be older, were more likely to be women and to have a family history of premature CHD (although these differences were not statistically significant). Systolic BP, total cholesterol, apolipoprotein B and fasting plasma glucose were higher among subjects with SCA, but generally within “normal ranges”. Lifestyle fac-
438
J. Grewal et al. / Atherosclerosis 197 (2008) 435–442
Table 2 Characteristics of subjects classified by the Framingham Risk Model as medium/high-risk (10-year risk of fatal and non-fatal CHD ≥10%) and low-risk (10 year risk of fatal and non-fatal CHD <10%), with and without subclinical atherosclerosis (SCA) Characteristic
FRS ≥ 10% (N = 263)
Demographic Age Women
57.1 ± 8.8 70 (27%)
Clinical Family history CHD Systolic blood pressure (mmHg) Diastolic blood pressure (mmHg) Waist to hip ratio ≥0.90 Men Women Exercise ≥3 h/week Alcohol consumption ≥3×/week Fruit intake (servings/day) Vegetable intake (servings/day) Hormone replacementa No menstrual period for 1 yeara Former/current smoking Men Women Biochemical Total cholesterol (mmol/L) LDL-cholesterol (mmol/L) Triglycerides (mmol/L) HDL-cholesterol (mmol/L) Total cholesterol/HDL ratio Fasting glucose (mmol/L) PAI (units/mL) Fibrinogen (g/L) Lipoprotein (a) (mg/dL) Apolipoprotein B (mg/dL) Homocysteine (mol/L) hs-CRP (g/L) Mean maximum IMT (mm)
p value
FRS < 10% With SCA (N = 175)
FRS < 10% Without SCA (N = 577)
p value
46.1 ± 8.4 461 (61%)
0.0001 0.0001
46.9 ± 8.0 117 (67%)
45.8 ± 8.4 344 (60%)
0.15 0.08
110 (42%) 132 ± 19 78 ± 12
230 (30%) 112 ± 14 70 ± 10
0.001 0.0001 0.0001
61 (35%) 115 ± 16 72 ± 10
169 (29%) 111 ± 14 70 ± 10
0.14 0.0001 0.17
149 (77%) 27 (39%)
179 (62%) 55 (12%)
0.0001 0.0001
39 (67%) 18 (15%)
140 (60%) 37 (11%)
0.21 0.25
179 (24%) 73 (9.6%) 2.2 ± 2.3 2.7 ± 2.0 85 (18%) 153 (33%)
0.60 0.78 0.12 0.39 0.08 0.0001
39 (22%) 20 (11%) 1.9 ± 1.4 2.8 ± 1.7 25 (21%) 39 (33%)
140 (24%) 53 (9%) 1.9 ± 2.1 2.8 ± 2.2 60 (17%) 114 (33%)
0.50 0.21 0.50 0.92 0.22 0.60
112 (58%) 25 (36%)
115 (40%) 139 (30%)
0.0001 0.31
23 (40%) 33 (28%)
92 (39%) 106 (31%)
0.34 0.60
5.5 ± 0.9 3.7 ± 0.8 1.9 ± 0.8 1.0 ± 0.2 5.8 ± 1.5 5.4 ± 0.6 17.3 ± 9.8 3.4 ± 1.3 165 (58–371) 1.2 ± 0.2 11.3 ± 3.8 1.6 (0.9–3.1)
4.9 ± 0.9 3.0 ± 0.7 1.5 ± 0.8 1.2 ± 0.4 4.4 ± 1.4 5.1 ± 0.5 15.3 ± 9.3 3.0 ± 0.6 156 (67–363) 0.9 ± 0.2 9.5 ± 4.2 1.1 (0.5–2.6)
0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.005 0.0001 0.93 0.001 0.001 0.008
4.9 ± 0.9 3.0 ± 0.8 1.5 ± 0.7 1.2 ± 0.4 4.4 ± 1.4 5.1 ± 0.5 15.1 ± 9.7 2.9 ± 0.6 158 (69–353) 0.9 ± 0.2 9.4 ± 4.3 1.0 (0.4–2.4)
0.01 0.11 0.15 0.23 0.50 0.01 0.39 0.11 0.54 0.04 0.37 0.88
70 (27%) 31 (12%) 2.0 ± 2.3 2.6 ± 2.0 18 (26%) 60 (86%)
0.78 (0.69–0.94)
FRS < 10% (N = 752)
0.64 (0.58–0.73)
5.1 ± 0.8 3.1 ± 0.7 1.6 ± 0.9 1.2 ± 0.3 4.7 ± 1.5 5.4 ± 0.6 16.1 ± 8.9 3.0 ± 0.6 148 (59–446) 1.0 ± 0.2 9.9 ± 3.5 1.1 (0.5–3.4)
0.0001
0.75 (0.69–0.87)
0.61 (0.56–0.67)
0.0001
Data are shown as mean ± S.D. or median (interquartile range) or number (%) FRS = Framingham Risk Score. a For women.
tors did not differ significantly among participants with and without SCA. As expected, participants in the medium and high-risk FRS groups had a higher prevalence of SCA than those in the low-risk group, 29% and 33%, respectively. 3.3. Predictors of carotid IMT and of subclinical atherosclerosis among low-risk subjects Among subjects in the low-risk group, the conventional risk factors and some novel risk factors (WHR, homocysteine, apolipoprotein B and hs-CRP) were predictors of carotid IMT in univariate analysis. Among lifestyle factors only exercise showed a weak univariate association with carotid IMT (Table 3). In the multivariate model age, male sex, systolic BP and apolipoprotein B were independent predictors of carotid IMT (Table 4). South Asian and Chinese ethnicity were predictive of lower carotid IMT as compared to the ref-
erent European Caucasian population, both in the univariate and multivariate analyses. Conventional risk factors and ethnicity accounted for 36% of the variability in carotid IMT. When considering the entire study cohort, including participants classified as low, medium and high risk, the independent predictors of carotid IMT were similar (data not shown). Univariate and multivariate predictors of SCA (defined as carotid IMT ≥75th percentile for age, sex and ethnic group) in the low-risk group are shown in Table 5.
4. Discussion Our main finding is the high prevalence of SCA in a “low-risk” multiethnic Canadian population. Indeed, 23% of individuals classified as low-risk for CHD by FRS had SCA. This is in contrast to the expectation, that most individuals classified as low risk, would be in the lower carotid
J. Grewal et al. / Atherosclerosis 197 (2008) 435–442
439
Table 3 Univariate predictors of carotid IMT (entered as a continuous variable) in the low-risk group (N = 752) Risk factor
Univariate  coefficient (S.E.)a
Standardized  coefficient
p value
Age Sex (women vs. men) Systolic blood pressure Total cholesterol/HDL cholesterol ratio LDL cholesterol (mmol/L) HDL cholesterol (mmol/L) Triglycerides (mmol/L) Smoking (former/current vs. never) Fasting glucose (mmol/L) Family history of premature CHD Waist to hip ratio ≥0.9 Homocysteine (mol/L) Lipoprotein (a) (g/L) Apolipoprotein B (g/L) Fibrinogen (g/L) hs-CRP (g/L) PAI (units/mL)
0.012 (0.001) −0.02 (0.01) 0.004 (0.000) 0.005 (0.005) 0.035 (0.009) 0.05 (0.02) 0.037 (0.009) 0.06 (0.01) 0.03 (0.01) 0.07 (0.01) 0.04 (0.01) 0.005 (0.002) −0.006 (0.006) 0.15 (0.03) 0.01 (0.04) 0.019 (0.005) −0.006 (0.001)
0.55 −0.05 0.31 0.04 0.14 0.10 0.15 0.15 0.10 0.17 0.10 0.11 −0.04 0.19 0.01 0.13 −0.003
0.0001 0.14 0.0001 0.25 0.0001 0.004 0.0001 0.0001 0.009 0.0001 0.005 0.002 0.32 0.0001 0.22 0.001 0.94
Ethnicity Chinese vs. European Caucasian South Asian vs. European Caucasian Aboriginal vs. European Caucasian
−0.08 (0.02) −0.07 (0.02) −0.002 (0.02)
−0.21 −0.17 −0.005
0.001 0.001 0.90
Alcohol ≥3 drinks/week Exercise ≥3 h/week Fruit servings/week Vegetable servings/week a
0.04 (0.02) 0.03 (0.02) 0.003 (0.004) 0.007 (0.004)
0.06 0.07 0.03 0.007
0.12 0.04 0.48 0.85
Per unit change for continuous variables.
IMT distribution. Moreover, the prevalence of SCA was similar across ethnic groups, including European Caucasians, for whom the Framingham model should be most accurate. The parent SHARE study found a strong association between carotid IMT and prevalent CV disease in this population, which was significant even after adjusting for the FRS [20]. Previous studies indicate that elevated carotid IMT predicts increased risk for CHD and stroke among men and women of various ages, even after adjusting for traditional risk factors and for the FRS [16–18]. Therefore, conventional risk stratification using the Framingham model misses an important proportion of individuals at risk of future CV events. To our knowledge, our study is unique in exploring the prevalence of SCA in low risk asymptomatic individuals belonging to these specific ethnic groups and residing in a Western environment. Carotid IMT is a validated method for identifying SCA and this technique was applied with high
reproducibility in our study. We defined SCA as carotid IMT ≥75th percentile of a population without CVD and diabetes, a commonly used definition [4], which is justified by prior studies which found asymptomatic individuals in the fourth quartile of the carotid IMT distribution (age and sex adjusted) to be at high risk for future CV events [17,18]. Similar results were reported by Michos et al., who used coronary artery calcium scoring for the detection of SCA, although this study was restricted to women, used a cohort referred for CV risk assessment and did not explore the influence of ethnicity [23]. Systolic blood pressure and apolipoprotein B were predictive of SCA in our study, while none of the other novel and lifestyle factors improved the prediction of SCA. Our results are consistent with large epidemiological investigations, which show that classical risk factors explain most of the population attributable risk for clinical CHD [4,6], although some studies found certain lifestyle factors and
Table 4 Multivariate independent predictors of carotid IMT (entered as a continuous variable) in the low-risk group (N = 752; 291 men and 461 women) Risk factor
Multivariate  coefficient (S.E.)
Standardized  coefficient
P value
Age Sex (women vs. men) Systolic blood pressure Apolipoprotein B (g/L)
0.011 (0.001) −0.03 (0.01) 0.002 (0.000) 0.08 (0.02)
0.55 −0.05 0.31 0.04
0.0001 0.01 0.0001 0.001
Ethnicity Chinese vs. European Caucasian South Asian vs. European Caucasian
−0.05 (0.01) −0.05 (0.01)
−0.14 −0.12
0.0001 0.001
Adjusted R2 = 0.36.
440
J. Grewal et al. / Atherosclerosis 197 (2008) 435–442
Table 5 Predictors of subclinical atherosclerosis defined as mean maximum carotid IMT ≥75th percentile for age, gender and ethnicity in the low-risk group (N = 752) Factor
Univariate OR (95%CI)
p
Multivariate OR (95% CI)
p
Age Sex—F vs. M Systolic blood pressure Cholesterol/HDL ratio LDL (cholesterol (mmol/L) HDL cholesterol (mmol/L) Triglycerides (mmol/L) Smoking—former and current Fasting glucose (mmol/L) Family history Waist to hip ratio ≥0.9 Homocysteine (mol/L) Lipoprotein (a) (g/L) Apolipoprotein B (g/L) Fibrinogen (g/L) hs-CRP (mg/L) PAI (units/mL)
1.02 (0.97–1.04) 1.40 (0.99–2.00) 1.02 (1.01–1.03) 1.04 (0.92–1.12) 1.20 (0.96–1.50) 1.27 (0.82–1.96) 1.17 (0.94–1.45) 0.87 (0.61–1.24) 0.95 (0.68–1.33) 1.33 (0.94–1.89) 0.92 (0.64–1.32) 1.01 (0.98–1.05) 1.04 (0.90–1.20) 2.15 (1.05–4.39) 0.83 (0.62–1.09) 1.08 (0.97–1.04) 0.99 (0.98–1.01)
0.11 0.05 0.0001 0.49 0.10 0.29 0.15 0.40 0.75 0.14 0.64 0.46 0.58 0.04 0.18 0.26 0.40
0.99 (0.98–1.02) 1.67 (1.15–2.42) 1.02 (1.01–1.04)
0 78 0.007 0.0001
1.00 (0.65–1.55)
0.99
2.38 (1.13–5.03)
0.02
Ethnicity Chinese vs. European Caucasian South Asian vs. European Caucasian Aboriginal vs. European Caucasian
0.97 (0.63–1.50) 0.85 (0.54–1.34) 0.77 (0.48–1.30)
0.91 0.49 0.32
Alcohol ≥3 drinks/week Exercise ≥3 h/week Fruit servings/week Vegetable servings/week
1.30 (0.75–2.24) 0.89 (0.59–1.33) 0.97 (0.88–1.07) 1.02 (0.91–1.09)
0.35 0.57 0.58 0.99
novel biochemical markers to have additional prognostic value for clinical CHD events [6,7]. These differences may be related to the size of our study, which is too small to conclusively evaluate the contribution of all potential risk factors to SCA. This may be particularly important in the evaluation of the importance of lifestyle factors. Thus, there is generally considerable “noise” in the ascertainment and quantitation of lifestyle factors and much larger studies would be required to study their effect on SCA. Moreover, although SCA may lead to clinical events, not all subjects with SCA will develop clinical CHD and the converse is also true. Although systolic blood pressure and apolipoprotein B were independent predictors of SCA determined by carotid IMT, these risk factors identified only a minority of participants with SCA. For example, only 130 (17%) of all study participants at low risk by FRS and only 38 of the 175 (21%) individuals with SCA by carotid IMT in this group had apolipoprotein B levels ≥75% percentile for age, sex and ethnicity. We found that, although among subjects in the low-risk group those with SCA had in general higher levels of traditional risk factors (BP, lipids, fasting plasma glucose) than those without SCA, these risk factors were on average within “normal” ranges. Such subtle alterations in the levels of these risk factors may not be reflected in the FRS, but will frequently affect carotid IMT, an integrated measure of lifelong exposure to risk factors and a direct measure of “occult” vascular disease. Therefore, carotid IMT is likely to be more informative as a predictor of risk than any CV risk factor
in isolation and than commonly used global risk assessment algorithms. In our study the prevalence of SCA among low-risk subjects by FRS tended to be higher among women and female sex emerged as an independent predictor of SCA. We believe that this finding reflects limitations of the Framingham risk model, which classifies most young and middle-aged women as being at low risk for CV events, although many may have subclinical vascular disease and their life-time risk may be substantial. However, our findings should not be misinterpreted to suggest, that among low-risk subjects by FRS, women are at higher risk for CHD. Indeed, male sex was an independent predictor of carotid IMT in our study and previous prospective investigations suggest that among subjects with SCA by carotid IMT, as compared to men, women have lower rates of incident CV events [17]. Rather, our findings suggest that arterial imaging may be particularly important in risk stratification in women, who in the absence of advanced age or extremes of risk factors are frequently classified as low risk. The prevalence of SCA among asymptomatic low-risk subjects did not differ significantly among the four ethnic groups evaluated and ethnicity was not an independent predictor of SCA. This may be related to the fact that we used ethnic specific normative values as previously suggested [12,13,15]. Indeed, using uniform thresholds values to define SCA based on the referent European Caucasian population (i.e. defining SCA as carotid IMT ≥75% for age and sex in the European Caucasian group), the prevalence of SCA
J. Grewal et al. / Atherosclerosis 197 (2008) 435–442
was 19% overall, 16% in South Asians, 11% in Chinese and 26% in Aboriginal people, respectively, versus 23% overall, 22% in South Asians, 24% in Chinese and 20% in Aboriginal people, when using ethnic specific normative values. In addition, our study is likely too small to conclusively exclude ethnic differences in the prevalence of SCA. Also, Aboriginal people were more likely to have CHD and diabetes and to be excluded from this analysis, further decreasing the power of the analysis for this ethnic group. In the analyses which used carotid IMT as a continuous variable, participants of European Caucasian ethnicity had higher carotid IMT than Chinese and South Asian participants and similar to subjects of Aboriginal ancestry. This was true for both univariate analysis and after controlling for age, sex and other risk factors, suggesting that ethnicity is an important determinant of carotid IMT. This finding supports the use of ethnic adjusted normative values for carotid IMT and the need for larger studies to explore differences in SCA among ethnic groups. We used carotid IMT as a simple, non-invasive imaging technique, which can reliably visualize early vascular abnormalities. Among additional imaging modalities used to identify and quantitate early atherosclerosis, only coronary calcium measurements by computed tomography has been prospectively validated and is and endorsed by current AHA and NCEP ATP III guidelines for refining CV risk [3,4]. Carotid ultrasound has several advantages. It does not involve any exposure to ionizing radiation, an important consideration when imaging healthy individuals for preventive purposes, while calcium screening using fast computed tomography is associated with some radiation exposure [24]. In addition, the predictive value of coronary calcium in younger men, women and African-Americans is less clear because of a lower prevalence of coronary artery calcium in these groups. Carotid IMT has the advantage of being a continuous measure that could be used to stratify risk in individuals where coronary artery calcium scoring may have limited discriminatory power because of a high predicted prevalence of a zero calcium score [25].
441
5. Conclusions About a fourth of subjects classified as low risk using traditional risk stratification approaches had prognostically important SCA in our study. Although conventional risk factors are the major determinants of SCA, these risk factors are often within “normal” ranges. Therefore, the use of arterial wall imaging, as an adjunct to traditional approaches, may aid in risk stratification and early disease detection. Carotid IMT is a simple, non-invasive, reliable and validated method to evaluate SCA and may enhance current risk stratification approaches. Use of carotid IMT for CV risk stratification requires reference to age, sex and ethnic specific normative values, such as those established in our laboratory (Table 1) and other large epidemiological studies [13,26–28]. Alternately, individual expert laboratories can use their own normative values. Current guidelines support the use of carotid IMT in CV risk stratification in intermediate risk individuals without established CV disease or diabetes and with 10-year estimated risk for CHD of 6–20%. These individuals may benefit most from additional measures of subclinical vascular disease to further refine their risk estimate and therefore the intensity of preventive treatments, as decision making in this group may be uncertain [11]. In our study we found that a considerable proportion of individuals classified as low risk by the FRS have SCA. While carotid IMT can be considered in selected low-risk subjects, such as subjects with a family history of premature CV disease, those with severe abnormalities in a single risk factor (e.g. genetic dyslipidemia) and women <60 years old with additional CV risk factors, we do not advocate at present the widespread use of carotid IMT in low risk groups, as the therapeutic consequences of identifying SCA in low risk individuals are uncertain. The study of therapies in low risk populations will require innovative approaches, because traditional clinical endpoint trials are difficult to conduct in low risk populations. Ultimately however, proof that risk stratification algorithms using arterial wall imaging such as carotid IMT can reduce CV events and are cost effective, is needed.
4.1. Limitations References We attempted to assemble a cohort representative of a true population by using a wide random sampling frame. However, only subjects who agreed to participate in the study were evaluated (volunteer bias). Key socio-demographic data were collected from all subjects invited to participate and no substantial differences were identified between responders versus non-responders, with the exception of lower rates of employment and of post-secondary education among nonresponders. Data on medical history and personal information were gathered by self-report and may have decreased the precision of risk factor ascertainment, especially for lifestyle factors. Finally, a larger study is needed to evaluate more conclusively possible ethnic differences in the prevalence of SCA among low-risk subjects.
[1] World Health Statistics. http://www.who.int/healthinfo/en/index.html [accessed, January 23, 2007]. [2] Thom T, Haase N, Rosamond W, et al. Heart disease and stroke statistics-2006 update: a report from the American Heart Association Statistics Committee and Stroke Statistics Subcommittee. Circulation 2006;113:e85–151. [3] Greenland P, Smith SC, Grundy SM. Improving coronary heart disease risk assessment in asymptomatic people: role of traditional risk factors and noninvasive cardiovascular tests. Circulation 2001;104: 1863–7. [4] National Cholesterol Education Program (NCEP) expert panel on detection, evaluation and treatment of high blood pressure in adults (ATP III). Third report of the National Cholesterol Education Program (NCEP) expert panel on detection, evaluation and treatment of high cholesterol in adults (ATP III) final report. Circulation 2002;106:3121–43.
442
J. Grewal et al. / Atherosclerosis 197 (2008) 435–442
[5] Third Joint Task Force of European and other Societies on Cardiovascular Prevention in Clinical Practice. European guidelines on cardiovascular prevention in clinical practice. Eur Heart J 2003;24:1601–10. [6] Yusuf S, Hawken S, Ounpuu S, et al. Effect of potentially modifiable risk factors associated with myocardial infarction in 52 countries (the INTERHEART study): case control study. Lancet 2004;364: 937–52. [7] Ridker PM, Hennekens CH, Buring JE, Cook NR, Rifai N. C-reactive protein and other markers of inflammation in the prediction of cardiovascular disease in women. N Engl J Med 2000;342:836–43. [8] Teo KK, Ounpuu S, Hawken S, et al. Tobacco use and risk of myocardial infarction in 52 countries in the INTERHEART study: a case-control study. Lancet 2006;368:647–58. [9] Coutinho M, Gerstein HC, Wang Y, Yusuf S. The relationship between glucose and incident cardiovascular events. A metaregression analysis of published data from 20 studies of 95,783 individuals followed for 12.4 years. Diab Care 1999;22:233–40. [10] Lloyd-Jones DM, Leip EP, Larson MG, et al. Prediction of lifetime risk for cardiovascular disease by risk factor burden at 50 years of age. Circulation 2006;113:791–8. [11] Taylor AJ, Merz CN, Udelson JE. 34th Bethesda Conference: Executive summary—can atherosclerosis imaging techniques improve the detection of patients at risk for ischemic heart disease? J Am Coll Cardiol 2003;41:1860–2. [12] Gepner AD, Keevil JG, Wyman RA, et al. Use of carotid intima-media thickness and vascular age to modify cardiovascular risk prediction. J Am Soc Echocardiogr 2006;19:1170–4. [13] Simon A, Gariepy J, Chironi G, Megnien JL, Levenson J. Intima-media thickness: a new tool for diagnosis and treatment of cardiovascular risk. J Hypertens 2002;20:159–69. [14] Roman MJ, Naqvi TZ, Gardin JM, et al. Clinical application of noninvasive vascular ultrasound in cardiovascular risk stratification: a report from the American Society of Echocardiography and the Society of Vascular Medicine and Biology. J Am Soc Echocardiogr 2006;19: 943–54. [15] Heiss G, Sharrett AR, Barnes R, et al. Carotid atherosclerosis measured by B-mode ultrasound in populations: associations with cardiovascular risk factors in the ARIC study. Am J Epidemiol 1991;134: 250–6. [16] Salonen JT, Salonen R. Ultrasonographically assessed carotid morphology and the risk of coronary heart disease. Arterioscler Thromb 1991;11:1245–9.
[17] Chambless LE, Heiss G, Folsom AR, et al. Association of coronary heart disease incidence with carotid arterial wall thickness and major risk factors: the ARIC study. Am J Epidemiol 1997;146: 483–94. [18] O’Leary DH, Polak JF, Kronaml RA, et al. Carotid-artery intima and media thickness as a risk factor for myocardial infarction and stroke in older adults. N Engl J Med 1999;340:14–22. [19] Anand S, Yusuf S, Vuksan V, et al. The Study of Health Assessment and Risk in Ethnic groups (SHARE): Rationale and design. Can J Cardiol 1998;14:1349–57. [20] Anand S, Yusuf S, Vuksan V, et al. Differences in risk factors, atherosclerosis, and cardiovascular disease between ethnic groups in Canada: the Study of Health Assessment and Risk in Ethnic Groups (SHARE). Lancet 2000;356:278–84. [21] Anand S, Yusuf S, Jacobs R, et al. Risk factors, atherosclerosis, and cardiovascular disease among Aboriginal people in Canada: the Study of Health Assessment and Risk Evaluation in Aboriginal Peoples (SHARE-AP). Lancet 2001;358:1147–53. [22] Wilson PW, D’Agostino RB, Levy D, Belanger AM, Silbershatz H, Kannel WB. Prediction of coronary heart disease using risk factor categories. Circulation 1998;97:1837–47. [23] Michos ED, Nasir K, Braunstein JB, et al. Framingham risk underestimates subclinical atherosclerosis risk in asymptomatic women. Atherosclerosis 2006;184:201–6. [24] Coles DR, Smail MA, Negus IS, Wilde P, et al. Comparison of radiation doses from multislice computed tomography coronary angiography and conventional diagnostic angiography. J Am Coll Cardiol 2006;47:1840–5. [25] McClelland RL, Chung H, Detrano R, et al. Distribution of coronary artery calcium by race, gender, and age: results from the Multi-Ethnic Study of Atherosclerosis (MESA). Circulation 2006;113:30–7. [26] Howard G, Sharrett A, Heiss G, Evans G, Chambless L, Riley W, Burke G. For the ARIC Investigators. Carotid artery intimal-medial thickness distribution in general populations as evaluated by B-mode ultrasound. Stroke 1993;24:1297–304. [27] Stein JH, Tzou WS, Douglas PS, Srinivasan S, Li S, Berenson GS. Predictors of carotid artery intima-media thickness in a large cohort of young adults: The Bogalusa Heart Study. Circulation 2005;112:II-819 [abstract]. [28] Allan PL, Mowbray PI, Lee AJ, Fowkes FG. Relationship between carotid intima-media thickness and symptomatic and asymptomatic peripheral arterial disease. The Edinburgh Artery Study. Stroke 1997;28:348–53.
Prevalence and predictors of subclinical atherosclerosis among asymptomatic “low risk” individuals in a multiethnic population Jasmine Grewal a,b , Sonia Anand a,b , Shofiqul Islam a , Eva Lonn a,b,∗ , On behalf of the SHARE and SHARE-AP Investigators b
a The Population Health Research Institute, McMaster University, Hamilton, Ontario, Canada The Department of Medicine, Division of Cardiology, McMaster University, Hamilton, Ontario, Canada
Received 16 February 2007; received in revised form 8 May 2007; accepted 21 June 2007 Available online 6 August 2007
Abstract Background: Current approaches to cardiovascular (CV) risk assessment have limitations. Subclinical atherosclerosis (SCA) as determined by carotid ultrasound is an independent predictor of myocardial infarction and stroke and can refine CV risk assessment. Objectives: We aimed to determine the prevalence and predictors of SCA in a multiethnic population classified as low risk for coronary heart disease (CHD) events by the Framingham Risk Assessment Model. Methods: We conducted a cross-sectional population study in 1015 Canadian adults of Caucasian European, South Asian, Chinese and Aboriginal ancestry. CHD risk was calculated by the 10-year Framingham Risk Score (FRS). Novel and conventional CHD risk factors were measured and high-resolution carotid ultrasound was performed. SCA was defined as carotid intima media thickness (IMT) ≥75th percentile adjusted for age, sex and ethnicity. Results: Seven hundred and fifty two (74%) participants were classified as low risk by FRS. Of these, 175 (23%) had evidence of SCA. Independent predictors of SCA among low-risk subjects included female sex, systolic blood pressure, and apolipoprotein B. Conclusions: Many individuals classified at low CHD risk by the FRS have SCA and are at increased long-term risk for vascular events. Carotid IMT can identify subjects with SCA, who may benefit from early intervention. © 2007 Elsevier Ireland Ltd. All rights reserved. Keywords: Cardiovascular risk; Carotid IMT; Risk factors; Ethnicity
1. Introduction Cardiovascular (CV) disease is a leading cause of death and disability [1]. For many individuals the first symptom of heart disease is sudden cardiac death or myocardial infarction (MI) [2]. Therefore, there is great interest in identifying asymptomatic individuals at risk, who would be candidates for more intensive, evidence-based medical interventions ∗ Corresponding author at: Hamilton Health Sciences, General Site, 237 Barton Street East, Hamilton, Ontario, Canada L8L 2X2. Tel.: +1 905 526 0970; fax: +1 905 527 5380. E-mail address: [email protected] (E. Lonn).
0021-9150/$ – see front matter © 2007 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.atherosclerosis.2007.06.020
aimed at preventing death and disability from coronary heart disease (CHD) and stroke [3]. The traditional approach to CHD risk assessment is based on identifying and to a certain extent quantifying established CV risk factors. Several algorithms based on this approach are used [4,5]. Among these, the Framingham Risk Assessment Model is the most widely accepted. Although used extensively and generally accepted, this model (as well as other algorithms based on similar approaches) has limitations. It is derived from a white Caucasian population in the US and may be less applicable to other ethnic groups. Family history, abdominal adiposity, inflammation and other factors shown to predict CV risk [6,7] are not incorporated in the Framingham
436
J. Grewal et al. / Atherosclerosis 197 (2008) 435–442
risk score (FRS). Diabetes and smoking are identified only as present or absent, although current evidence supports a continuous relationship between glycemia and tobacco exposure to CHD risk [8,9]. Age is the overriding FRS determinant, ignoring greater inter-individual variation in atherosclerotic burden at older ages and often providing false reassurance at younger ages. Importantly, the FRS only predicts shortterm (10-year) risk, although from a clinical perspective the life-long risk of developing CHD events is equally relevant [10]. Imaging of arteries to identify and quantify the presence of subclinical atherosclerosis (SCA) has been suggested to further refine CV risk assessment [3,4,11]. Among available imaging techniques, measurement of carotid intima-media thickness (IMT) with B-mode ultrasound is a non-invasive, sensitive, and highly reproducible technique for identifying and quantifying atherosclerotic burden. It is a well-validated research tool, increasingly used also in clinical practice [12–14]. Indeed, carotid IMT was shown to accurately represent anatomic structural abnormalities [12,13], that correlate with various classical and emerging CV risk factors and with prevalent CV disease [12–15] and importantly, to be an independent predictor of MI and stroke [16–18]. The American Heart Association (AHA) and the Third Adult Treatment Panel of the National Cholesterol Education Program (NCEP ATP III) have endorsed the use of carotid IMT in CV risk assessment [3,4]. To date, the prevalence and predictors of SCA in a “low risk” multiethnic western population have not been examined.
2. Methods The Study of Health Assessment and Risk in Ethnic groups (SHARE) and the Study of Health Assessment and Risk in Aboriginal Peoples (SHARE-AP) were cross-sectional investigations of atherosclerosis, CV risk factors and CV disease in four ethnic groups living in Canada—Caucasian Europeans, South Asians, Chinese and Aboriginal people. The design and major findings of these studies have been previously reported [19–21]. A brief description follows. 2.1. Study population Between 1996 and 2000 participants were randomly sampled from 4 communities. The Ethics Boards of all participating institutions approved the studies and all participants provided written informed consent. A careful sampling methodology was used, aimed at assembling a true population sample, as previously described [19–21]. For the purposes of this analysis we excluded all participants with known CV disease and/or diabetes. CV disease was defined as CHD (angina by Rose questionnaire, history of MI, silent MI [major Q waves by Minnesota criteria], prior percutaneous coronary intervention or coronary artery bypass graft surgery), or his-
tory of stroke. Diabetes was defined as self-reported history of diabetes, use of oral hypoglycemic drugs or insulin, fasting plasma glucose concentration ≥7.0 mmol/L or plasma glucose ≥11.1 mmol/L on a 2-h oral glucose tolerance test. 2.2. Assessment of risk factors Demographic, lifestyle, personal and family history were obtained using validated questionnaires (translated into Tamil, Punjabi, Hindi and Chinese and including validated ethnic specific food questionnaires) [19]. Regular alcohol use was defined as consumption ≥3 times/week. Fruit and vegetable intake was recorded as servings per day. Individuals were judged to be physically active if they engaged in moderate (walking, cycling) or strenuous exercise (jogging, swimming) for ≥3 h per week. Blood pressure (BP), weight, waist to hip ratio (WHR), fasting plasma lipid and glucose levels, hemoglobin A1C (HbA1C ), lipoprotein (a), apolipoprotein B, homocysteine, C-reactive protein (hsCRP), fibrinogen, plasminogen activator inhibitor 1 (PAI-1), white blood cell count and norepinephrine were measured using validated and reproducible methods described previously [19–21]. 2.3. Carotid ultrasonography SCA was measured using high-resolution ultrasound and standardized and extensively validated scanning and measurement protocols [19–21]. A transverse scan was followed by a longitudinal circumferential scan, aimed at imaging the thickest arterial wall of 12 well defined segments, the near and the far walls of the right and left common carotid (1 cm below the bulb), bulb (1 cm below the flow divider) and internal carotid (1 cm above the flow divider) arteries. Offline measurements were done at the Core Laboratory in Hamilton using custom-built software. A minimum of 3 frames from each pre-specified arterial segment with the thickest IMT were measured. The maximum IMT for each of the 12 segments was selected and the mean maximum IMT was computed for each subject by averaging the segment maximum IMT measurements of all segments visualized. There were 4 sonographers and 2 readers, specifically trained and certified. Between and within sonographer and reader reproducibility was high (intraclass correlation coefficients ≥0.90 and coefficients of variation < 5%). We computed the distribution of the mean maximum carotid IMT for groups of individuals defined by age (in decades), sex and ethnicity. Participants with a mean maximum carotid IMT ≥75th percentile for their age, sex and ethnicity were defined as having SCA (Table 1). 2.4. Cardiovascular risk assessment Framingham sex-specific equations were used to predict the 10-year risk for fatal and non-fatal CHD [4,22]. Participants were classified as low, intermediate and high risk if
J. Grewal et al. / Atherosclerosis 197 (2008) 435–442
437
Table 1 Normative values for carotid IMT in the SHARE cohort in subjects without cardiovascular disease and diabetesa (N = 1015) Age
Caucasian
South Asian
Chinese
Aboriginal
Women
Men
Women
Men
Women
Men
Women
Men
Mean (S.D.) 35–45 46–55 56–65 66–75
0.63 (0.11) 0.70 (0.12) 0.80 (0.18) 0.90 (0.22)
0.66 (0.08) 0.74 (0.12) 0.92 (0.23) 1.12 (0.37)
0.60 (0.11) 0.69 (0.15) 0.84 (0.30) 0.90 (0.38)
0.63 (0.12) 0.73 (0.15) 0.82 (0.18) 0.92 (0.15)
0.59 (0.07) 0.65 (0.08) 0.70 (0.07) 0.86 (0.25)
0.64 (0.11) 0.71 (0.10) 0.80 (0.14) 0.90 (0.24)
0.61 (0.09) 0.71 (0.13) 0.89 (0.26) 1.08 (0.26)
0.67 (0.14) 0.78 (0.19) 0.98 (0.32) 0.95 (0.18)
75% Percentile 35–45 46–55 56–65 66–75
0.68 0.74 0.88 0.97
0.70 0.81 0.97 1.33
0.63 0.75 0.92 1.10
0.68 0.80 0.95 0.99
0.63 0.68 0.75 1.00
0.68 0.76 0.86 1.16
0.67 0.79 0.98 1.23
0.73 0.89 1.11 1.18
a The presence of carotid IMT ≥75% percentile for subject’s age, sex and race/ethnicity are indicative of increased CV risk and may signify the need for aggressive risk-reduction interventions. The values should only be applied if utilizing the carotid ultrasound imaging and reading protocols outlined in Section 2.
their CHD risk was calculated at <10%, 10% and 20% and >20%, respectively. 2.5. Statistical analysis Characteristics of subjects in the low versus the medium/high FRS groups and between subjects with and without SCA were compared using ANOVA or logistic regression, as appropriate. Multivariate linear and logistic regression with backward elimination was used to determine independent predictors of carotid IMT (entered as a continuous variable in the model) and of SCA (defined as carotid IMT ≥75th percentile for their age, sex and ethnicity), respectively. Independent variables were selected on the basis of their clinical importance as CV risk factors and their significance in bivariate analyses (p ≤ 0.10). Because carotid IMT, lipoprotein (a) and hs-CRP measurements were not normally distributed, logarithmic transformation was used for all computations involving these variables. White European ethnicity was used as the referent population in determining the influence of ethnicity on carotid IMT (as a continuous variable) and of SCA. Statistical significance was defined as two-sided p < 0.05. All statistical analyses were performed using SPSS version 12.0 (SPSS Inc, Chicago, IL).
3. Results 3.1. Study participants The SHARE and SHARE-AP studies enrolled 1285 participants. CV disease was present in 172, diabetes in 88 (CV disease and diabetes were more common in men and Aboriginal people [20,21]), and carotid IMT recordings were suboptimal in 10, leaving a total of 1015 subjects for this analysis. These were 484 men and 531 women, 285 of Caucasian European ethnicity, 258 South Asians, 284 Chi-
nese and 188 Aboriginals. Of these 60 were classified as high-risk, 203 as intermediate risk and 752 as low-risk by FRS. In this cohort of subjects without CV disease and diabetes, 60% of men (291 of 484) and 87% of women (461 of 531) were classified as low risk; also, 75% of participants of Caucasian European ethnicity (213 of 285), 72% of South Asians (187 of 258), 78% of Chinese (223 of 285), and 69% of Aboriginals (129 of 188), were classified as low-risk (p = 0.10). As expected, those in the medium/highrisk groups were older, predominantly male, more likely to smoke and had higher levels of conventional and novel CV risk factors, such as blood pressure, cholesterol, fasting plasma glucose, waist to hip ratio (WHR), apolipoprotein B, PAI-1, fibrinogen, homocysteine, hs-CRP and had higher carotid IMT. There were no significant differences between the low and high-risk groups in lifestyle factors (alcohol consumption, physical activity, fruit and vegetable intake) (Table 2). 3.2. Prevalence of subclinical atherosclerosis among low-risk subjects Among the 752 low-risk individuals, SCA (carotid IMT ≥75th percentile for age, sex and ethnicity) was present in 175 (23%), 20% of men (58 of 291) and 25% of women (117 of 461), 25% participants of European ethnicity (53 of 213), 22% of South Asians (41 of 187), 24% of Chinese (55 of 223) and 20% of Aboriginals (26 of 129) (p = 0.66). Characteristics of low-risk subjects with and without SCA are shown in Table 2. Those with SCA tended to be older, were more likely to be women and to have a family history of premature CHD (although these differences were not statistically significant). Systolic BP, total cholesterol, apolipoprotein B and fasting plasma glucose were higher among subjects with SCA, but generally within “normal ranges”. Lifestyle fac-
438
J. Grewal et al. / Atherosclerosis 197 (2008) 435–442
Table 2 Characteristics of subjects classified by the Framingham Risk Model as medium/high-risk (10-year risk of fatal and non-fatal CHD ≥10%) and low-risk (10 year risk of fatal and non-fatal CHD <10%), with and without subclinical atherosclerosis (SCA) Characteristic
FRS ≥ 10% (N = 263)
Demographic Age Women
57.1 ± 8.8 70 (27%)
Clinical Family history CHD Systolic blood pressure (mmHg) Diastolic blood pressure (mmHg) Waist to hip ratio ≥0.90 Men Women Exercise ≥3 h/week Alcohol consumption ≥3×/week Fruit intake (servings/day) Vegetable intake (servings/day) Hormone replacementa No menstrual period for 1 yeara Former/current smoking Men Women Biochemical Total cholesterol (mmol/L) LDL-cholesterol (mmol/L) Triglycerides (mmol/L) HDL-cholesterol (mmol/L) Total cholesterol/HDL ratio Fasting glucose (mmol/L) PAI (units/mL) Fibrinogen (g/L) Lipoprotein (a) (mg/dL) Apolipoprotein B (mg/dL) Homocysteine (mol/L) hs-CRP (g/L) Mean maximum IMT (mm)
p value
FRS < 10% With SCA (N = 175)
FRS < 10% Without SCA (N = 577)
p value
46.1 ± 8.4 461 (61%)
0.0001 0.0001
46.9 ± 8.0 117 (67%)
45.8 ± 8.4 344 (60%)
0.15 0.08
110 (42%) 132 ± 19 78 ± 12
230 (30%) 112 ± 14 70 ± 10
0.001 0.0001 0.0001
61 (35%) 115 ± 16 72 ± 10
169 (29%) 111 ± 14 70 ± 10
0.14 0.0001 0.17
149 (77%) 27 (39%)
179 (62%) 55 (12%)
0.0001 0.0001
39 (67%) 18 (15%)
140 (60%) 37 (11%)
0.21 0.25
179 (24%) 73 (9.6%) 2.2 ± 2.3 2.7 ± 2.0 85 (18%) 153 (33%)
0.60 0.78 0.12 0.39 0.08 0.0001
39 (22%) 20 (11%) 1.9 ± 1.4 2.8 ± 1.7 25 (21%) 39 (33%)
140 (24%) 53 (9%) 1.9 ± 2.1 2.8 ± 2.2 60 (17%) 114 (33%)
0.50 0.21 0.50 0.92 0.22 0.60
112 (58%) 25 (36%)
115 (40%) 139 (30%)
0.0001 0.31
23 (40%) 33 (28%)
92 (39%) 106 (31%)
0.34 0.60
5.5 ± 0.9 3.7 ± 0.8 1.9 ± 0.8 1.0 ± 0.2 5.8 ± 1.5 5.4 ± 0.6 17.3 ± 9.8 3.4 ± 1.3 165 (58–371) 1.2 ± 0.2 11.3 ± 3.8 1.6 (0.9–3.1)
4.9 ± 0.9 3.0 ± 0.7 1.5 ± 0.8 1.2 ± 0.4 4.4 ± 1.4 5.1 ± 0.5 15.3 ± 9.3 3.0 ± 0.6 156 (67–363) 0.9 ± 0.2 9.5 ± 4.2 1.1 (0.5–2.6)
0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.005 0.0001 0.93 0.001 0.001 0.008
4.9 ± 0.9 3.0 ± 0.8 1.5 ± 0.7 1.2 ± 0.4 4.4 ± 1.4 5.1 ± 0.5 15.1 ± 9.7 2.9 ± 0.6 158 (69–353) 0.9 ± 0.2 9.4 ± 4.3 1.0 (0.4–2.4)
0.01 0.11 0.15 0.23 0.50 0.01 0.39 0.11 0.54 0.04 0.37 0.88
70 (27%) 31 (12%) 2.0 ± 2.3 2.6 ± 2.0 18 (26%) 60 (86%)
0.78 (0.69–0.94)
FRS < 10% (N = 752)
0.64 (0.58–0.73)
5.1 ± 0.8 3.1 ± 0.7 1.6 ± 0.9 1.2 ± 0.3 4.7 ± 1.5 5.4 ± 0.6 16.1 ± 8.9 3.0 ± 0.6 148 (59–446) 1.0 ± 0.2 9.9 ± 3.5 1.1 (0.5–3.4)
0.0001
0.75 (0.69–0.87)
0.61 (0.56–0.67)
0.0001
Data are shown as mean ± S.D. or median (interquartile range) or number (%) FRS = Framingham Risk Score. a For women.
tors did not differ significantly among participants with and without SCA. As expected, participants in the medium and high-risk FRS groups had a higher prevalence of SCA than those in the low-risk group, 29% and 33%, respectively. 3.3. Predictors of carotid IMT and of subclinical atherosclerosis among low-risk subjects Among subjects in the low-risk group, the conventional risk factors and some novel risk factors (WHR, homocysteine, apolipoprotein B and hs-CRP) were predictors of carotid IMT in univariate analysis. Among lifestyle factors only exercise showed a weak univariate association with carotid IMT (Table 3). In the multivariate model age, male sex, systolic BP and apolipoprotein B were independent predictors of carotid IMT (Table 4). South Asian and Chinese ethnicity were predictive of lower carotid IMT as compared to the ref-
erent European Caucasian population, both in the univariate and multivariate analyses. Conventional risk factors and ethnicity accounted for 36% of the variability in carotid IMT. When considering the entire study cohort, including participants classified as low, medium and high risk, the independent predictors of carotid IMT were similar (data not shown). Univariate and multivariate predictors of SCA (defined as carotid IMT ≥75th percentile for age, sex and ethnic group) in the low-risk group are shown in Table 5.
4. Discussion Our main finding is the high prevalence of SCA in a “low-risk” multiethnic Canadian population. Indeed, 23% of individuals classified as low-risk for CHD by FRS had SCA. This is in contrast to the expectation, that most individuals classified as low risk, would be in the lower carotid
J. Grewal et al. / Atherosclerosis 197 (2008) 435–442
439
Table 3 Univariate predictors of carotid IMT (entered as a continuous variable) in the low-risk group (N = 752) Risk factor
Univariate  coefficient (S.E.)a
Standardized  coefficient
p value
Age Sex (women vs. men) Systolic blood pressure Total cholesterol/HDL cholesterol ratio LDL cholesterol (mmol/L) HDL cholesterol (mmol/L) Triglycerides (mmol/L) Smoking (former/current vs. never) Fasting glucose (mmol/L) Family history of premature CHD Waist to hip ratio ≥0.9 Homocysteine (mol/L) Lipoprotein (a) (g/L) Apolipoprotein B (g/L) Fibrinogen (g/L) hs-CRP (g/L) PAI (units/mL)
0.012 (0.001) −0.02 (0.01) 0.004 (0.000) 0.005 (0.005) 0.035 (0.009) 0.05 (0.02) 0.037 (0.009) 0.06 (0.01) 0.03 (0.01) 0.07 (0.01) 0.04 (0.01) 0.005 (0.002) −0.006 (0.006) 0.15 (0.03) 0.01 (0.04) 0.019 (0.005) −0.006 (0.001)
0.55 −0.05 0.31 0.04 0.14 0.10 0.15 0.15 0.10 0.17 0.10 0.11 −0.04 0.19 0.01 0.13 −0.003
0.0001 0.14 0.0001 0.25 0.0001 0.004 0.0001 0.0001 0.009 0.0001 0.005 0.002 0.32 0.0001 0.22 0.001 0.94
Ethnicity Chinese vs. European Caucasian South Asian vs. European Caucasian Aboriginal vs. European Caucasian
−0.08 (0.02) −0.07 (0.02) −0.002 (0.02)
−0.21 −0.17 −0.005
0.001 0.001 0.90
Alcohol ≥3 drinks/week Exercise ≥3 h/week Fruit servings/week Vegetable servings/week a
0.04 (0.02) 0.03 (0.02) 0.003 (0.004) 0.007 (0.004)
0.06 0.07 0.03 0.007
0.12 0.04 0.48 0.85
Per unit change for continuous variables.
IMT distribution. Moreover, the prevalence of SCA was similar across ethnic groups, including European Caucasians, for whom the Framingham model should be most accurate. The parent SHARE study found a strong association between carotid IMT and prevalent CV disease in this population, which was significant even after adjusting for the FRS [20]. Previous studies indicate that elevated carotid IMT predicts increased risk for CHD and stroke among men and women of various ages, even after adjusting for traditional risk factors and for the FRS [16–18]. Therefore, conventional risk stratification using the Framingham model misses an important proportion of individuals at risk of future CV events. To our knowledge, our study is unique in exploring the prevalence of SCA in low risk asymptomatic individuals belonging to these specific ethnic groups and residing in a Western environment. Carotid IMT is a validated method for identifying SCA and this technique was applied with high
reproducibility in our study. We defined SCA as carotid IMT ≥75th percentile of a population without CVD and diabetes, a commonly used definition [4], which is justified by prior studies which found asymptomatic individuals in the fourth quartile of the carotid IMT distribution (age and sex adjusted) to be at high risk for future CV events [17,18]. Similar results were reported by Michos et al., who used coronary artery calcium scoring for the detection of SCA, although this study was restricted to women, used a cohort referred for CV risk assessment and did not explore the influence of ethnicity [23]. Systolic blood pressure and apolipoprotein B were predictive of SCA in our study, while none of the other novel and lifestyle factors improved the prediction of SCA. Our results are consistent with large epidemiological investigations, which show that classical risk factors explain most of the population attributable risk for clinical CHD [4,6], although some studies found certain lifestyle factors and
Table 4 Multivariate independent predictors of carotid IMT (entered as a continuous variable) in the low-risk group (N = 752; 291 men and 461 women) Risk factor
Multivariate  coefficient (S.E.)
Standardized  coefficient
P value
Age Sex (women vs. men) Systolic blood pressure Apolipoprotein B (g/L)
0.011 (0.001) −0.03 (0.01) 0.002 (0.000) 0.08 (0.02)
0.55 −0.05 0.31 0.04
0.0001 0.01 0.0001 0.001
Ethnicity Chinese vs. European Caucasian South Asian vs. European Caucasian
−0.05 (0.01) −0.05 (0.01)
−0.14 −0.12
0.0001 0.001
Adjusted R2 = 0.36.
440
J. Grewal et al. / Atherosclerosis 197 (2008) 435–442
Table 5 Predictors of subclinical atherosclerosis defined as mean maximum carotid IMT ≥75th percentile for age, gender and ethnicity in the low-risk group (N = 752) Factor
Univariate OR (95%CI)
p
Multivariate OR (95% CI)
p
Age Sex—F vs. M Systolic blood pressure Cholesterol/HDL ratio LDL (cholesterol (mmol/L) HDL cholesterol (mmol/L) Triglycerides (mmol/L) Smoking—former and current Fasting glucose (mmol/L) Family history Waist to hip ratio ≥0.9 Homocysteine (mol/L) Lipoprotein (a) (g/L) Apolipoprotein B (g/L) Fibrinogen (g/L) hs-CRP (mg/L) PAI (units/mL)
1.02 (0.97–1.04) 1.40 (0.99–2.00) 1.02 (1.01–1.03) 1.04 (0.92–1.12) 1.20 (0.96–1.50) 1.27 (0.82–1.96) 1.17 (0.94–1.45) 0.87 (0.61–1.24) 0.95 (0.68–1.33) 1.33 (0.94–1.89) 0.92 (0.64–1.32) 1.01 (0.98–1.05) 1.04 (0.90–1.20) 2.15 (1.05–4.39) 0.83 (0.62–1.09) 1.08 (0.97–1.04) 0.99 (0.98–1.01)
0.11 0.05 0.0001 0.49 0.10 0.29 0.15 0.40 0.75 0.14 0.64 0.46 0.58 0.04 0.18 0.26 0.40
0.99 (0.98–1.02) 1.67 (1.15–2.42) 1.02 (1.01–1.04)
0 78 0.007 0.0001
1.00 (0.65–1.55)
0.99
2.38 (1.13–5.03)
0.02
Ethnicity Chinese vs. European Caucasian South Asian vs. European Caucasian Aboriginal vs. European Caucasian
0.97 (0.63–1.50) 0.85 (0.54–1.34) 0.77 (0.48–1.30)
0.91 0.49 0.32
Alcohol ≥3 drinks/week Exercise ≥3 h/week Fruit servings/week Vegetable servings/week
1.30 (0.75–2.24) 0.89 (0.59–1.33) 0.97 (0.88–1.07) 1.02 (0.91–1.09)
0.35 0.57 0.58 0.99
novel biochemical markers to have additional prognostic value for clinical CHD events [6,7]. These differences may be related to the size of our study, which is too small to conclusively evaluate the contribution of all potential risk factors to SCA. This may be particularly important in the evaluation of the importance of lifestyle factors. Thus, there is generally considerable “noise” in the ascertainment and quantitation of lifestyle factors and much larger studies would be required to study their effect on SCA. Moreover, although SCA may lead to clinical events, not all subjects with SCA will develop clinical CHD and the converse is also true. Although systolic blood pressure and apolipoprotein B were independent predictors of SCA determined by carotid IMT, these risk factors identified only a minority of participants with SCA. For example, only 130 (17%) of all study participants at low risk by FRS and only 38 of the 175 (21%) individuals with SCA by carotid IMT in this group had apolipoprotein B levels ≥75% percentile for age, sex and ethnicity. We found that, although among subjects in the low-risk group those with SCA had in general higher levels of traditional risk factors (BP, lipids, fasting plasma glucose) than those without SCA, these risk factors were on average within “normal” ranges. Such subtle alterations in the levels of these risk factors may not be reflected in the FRS, but will frequently affect carotid IMT, an integrated measure of lifelong exposure to risk factors and a direct measure of “occult” vascular disease. Therefore, carotid IMT is likely to be more informative as a predictor of risk than any CV risk factor
in isolation and than commonly used global risk assessment algorithms. In our study the prevalence of SCA among low-risk subjects by FRS tended to be higher among women and female sex emerged as an independent predictor of SCA. We believe that this finding reflects limitations of the Framingham risk model, which classifies most young and middle-aged women as being at low risk for CV events, although many may have subclinical vascular disease and their life-time risk may be substantial. However, our findings should not be misinterpreted to suggest, that among low-risk subjects by FRS, women are at higher risk for CHD. Indeed, male sex was an independent predictor of carotid IMT in our study and previous prospective investigations suggest that among subjects with SCA by carotid IMT, as compared to men, women have lower rates of incident CV events [17]. Rather, our findings suggest that arterial imaging may be particularly important in risk stratification in women, who in the absence of advanced age or extremes of risk factors are frequently classified as low risk. The prevalence of SCA among asymptomatic low-risk subjects did not differ significantly among the four ethnic groups evaluated and ethnicity was not an independent predictor of SCA. This may be related to the fact that we used ethnic specific normative values as previously suggested [12,13,15]. Indeed, using uniform thresholds values to define SCA based on the referent European Caucasian population (i.e. defining SCA as carotid IMT ≥75% for age and sex in the European Caucasian group), the prevalence of SCA
J. Grewal et al. / Atherosclerosis 197 (2008) 435–442
was 19% overall, 16% in South Asians, 11% in Chinese and 26% in Aboriginal people, respectively, versus 23% overall, 22% in South Asians, 24% in Chinese and 20% in Aboriginal people, when using ethnic specific normative values. In addition, our study is likely too small to conclusively exclude ethnic differences in the prevalence of SCA. Also, Aboriginal people were more likely to have CHD and diabetes and to be excluded from this analysis, further decreasing the power of the analysis for this ethnic group. In the analyses which used carotid IMT as a continuous variable, participants of European Caucasian ethnicity had higher carotid IMT than Chinese and South Asian participants and similar to subjects of Aboriginal ancestry. This was true for both univariate analysis and after controlling for age, sex and other risk factors, suggesting that ethnicity is an important determinant of carotid IMT. This finding supports the use of ethnic adjusted normative values for carotid IMT and the need for larger studies to explore differences in SCA among ethnic groups. We used carotid IMT as a simple, non-invasive imaging technique, which can reliably visualize early vascular abnormalities. Among additional imaging modalities used to identify and quantitate early atherosclerosis, only coronary calcium measurements by computed tomography has been prospectively validated and is and endorsed by current AHA and NCEP ATP III guidelines for refining CV risk [3,4]. Carotid ultrasound has several advantages. It does not involve any exposure to ionizing radiation, an important consideration when imaging healthy individuals for preventive purposes, while calcium screening using fast computed tomography is associated with some radiation exposure [24]. In addition, the predictive value of coronary calcium in younger men, women and African-Americans is less clear because of a lower prevalence of coronary artery calcium in these groups. Carotid IMT has the advantage of being a continuous measure that could be used to stratify risk in individuals where coronary artery calcium scoring may have limited discriminatory power because of a high predicted prevalence of a zero calcium score [25].
441
5. Conclusions About a fourth of subjects classified as low risk using traditional risk stratification approaches had prognostically important SCA in our study. Although conventional risk factors are the major determinants of SCA, these risk factors are often within “normal” ranges. Therefore, the use of arterial wall imaging, as an adjunct to traditional approaches, may aid in risk stratification and early disease detection. Carotid IMT is a simple, non-invasive, reliable and validated method to evaluate SCA and may enhance current risk stratification approaches. Use of carotid IMT for CV risk stratification requires reference to age, sex and ethnic specific normative values, such as those established in our laboratory (Table 1) and other large epidemiological studies [13,26–28]. Alternately, individual expert laboratories can use their own normative values. Current guidelines support the use of carotid IMT in CV risk stratification in intermediate risk individuals without established CV disease or diabetes and with 10-year estimated risk for CHD of 6–20%. These individuals may benefit most from additional measures of subclinical vascular disease to further refine their risk estimate and therefore the intensity of preventive treatments, as decision making in this group may be uncertain [11]. In our study we found that a considerable proportion of individuals classified as low risk by the FRS have SCA. While carotid IMT can be considered in selected low-risk subjects, such as subjects with a family history of premature CV disease, those with severe abnormalities in a single risk factor (e.g. genetic dyslipidemia) and women <60 years old with additional CV risk factors, we do not advocate at present the widespread use of carotid IMT in low risk groups, as the therapeutic consequences of identifying SCA in low risk individuals are uncertain. The study of therapies in low risk populations will require innovative approaches, because traditional clinical endpoint trials are difficult to conduct in low risk populations. Ultimately however, proof that risk stratification algorithms using arterial wall imaging such as carotid IMT can reduce CV events and are cost effective, is needed.
4.1. Limitations References We attempted to assemble a cohort representative of a true population by using a wide random sampling frame. However, only subjects who agreed to participate in the study were evaluated (volunteer bias). Key socio-demographic data were collected from all subjects invited to participate and no substantial differences were identified between responders versus non-responders, with the exception of lower rates of employment and of post-secondary education among nonresponders. Data on medical history and personal information were gathered by self-report and may have decreased the precision of risk factor ascertainment, especially for lifestyle factors. Finally, a larger study is needed to evaluate more conclusively possible ethnic differences in the prevalence of SCA among low-risk subjects.
[1] World Health Statistics. http://www.who.int/healthinfo/en/index.html [accessed, January 23, 2007]. [2] Thom T, Haase N, Rosamond W, et al. Heart disease and stroke statistics-2006 update: a report from the American Heart Association Statistics Committee and Stroke Statistics Subcommittee. Circulation 2006;113:e85–151. [3] Greenland P, Smith SC, Grundy SM. Improving coronary heart disease risk assessment in asymptomatic people: role of traditional risk factors and noninvasive cardiovascular tests. Circulation 2001;104: 1863–7. [4] National Cholesterol Education Program (NCEP) expert panel on detection, evaluation and treatment of high blood pressure in adults (ATP III). Third report of the National Cholesterol Education Program (NCEP) expert panel on detection, evaluation and treatment of high cholesterol in adults (ATP III) final report. Circulation 2002;106:3121–43.
442
J. Grewal et al. / Atherosclerosis 197 (2008) 435–442
[5] Third Joint Task Force of European and other Societies on Cardiovascular Prevention in Clinical Practice. European guidelines on cardiovascular prevention in clinical practice. Eur Heart J 2003;24:1601–10. [6] Yusuf S, Hawken S, Ounpuu S, et al. Effect of potentially modifiable risk factors associated with myocardial infarction in 52 countries (the INTERHEART study): case control study. Lancet 2004;364: 937–52. [7] Ridker PM, Hennekens CH, Buring JE, Cook NR, Rifai N. C-reactive protein and other markers of inflammation in the prediction of cardiovascular disease in women. N Engl J Med 2000;342:836–43. [8] Teo KK, Ounpuu S, Hawken S, et al. Tobacco use and risk of myocardial infarction in 52 countries in the INTERHEART study: a case-control study. Lancet 2006;368:647–58. [9] Coutinho M, Gerstein HC, Wang Y, Yusuf S. The relationship between glucose and incident cardiovascular events. A metaregression analysis of published data from 20 studies of 95,783 individuals followed for 12.4 years. Diab Care 1999;22:233–40. [10] Lloyd-Jones DM, Leip EP, Larson MG, et al. Prediction of lifetime risk for cardiovascular disease by risk factor burden at 50 years of age. Circulation 2006;113:791–8. [11] Taylor AJ, Merz CN, Udelson JE. 34th Bethesda Conference: Executive summary—can atherosclerosis imaging techniques improve the detection of patients at risk for ischemic heart disease? J Am Coll Cardiol 2003;41:1860–2. [12] Gepner AD, Keevil JG, Wyman RA, et al. Use of carotid intima-media thickness and vascular age to modify cardiovascular risk prediction. J Am Soc Echocardiogr 2006;19:1170–4. [13] Simon A, Gariepy J, Chironi G, Megnien JL, Levenson J. Intima-media thickness: a new tool for diagnosis and treatment of cardiovascular risk. J Hypertens 2002;20:159–69. [14] Roman MJ, Naqvi TZ, Gardin JM, et al. Clinical application of noninvasive vascular ultrasound in cardiovascular risk stratification: a report from the American Society of Echocardiography and the Society of Vascular Medicine and Biology. J Am Soc Echocardiogr 2006;19: 943–54. [15] Heiss G, Sharrett AR, Barnes R, et al. Carotid atherosclerosis measured by B-mode ultrasound in populations: associations with cardiovascular risk factors in the ARIC study. Am J Epidemiol 1991;134: 250–6. [16] Salonen JT, Salonen R. Ultrasonographically assessed carotid morphology and the risk of coronary heart disease. Arterioscler Thromb 1991;11:1245–9.
[17] Chambless LE, Heiss G, Folsom AR, et al. Association of coronary heart disease incidence with carotid arterial wall thickness and major risk factors: the ARIC study. Am J Epidemiol 1997;146: 483–94. [18] O’Leary DH, Polak JF, Kronaml RA, et al. Carotid-artery intima and media thickness as a risk factor for myocardial infarction and stroke in older adults. N Engl J Med 1999;340:14–22. [19] Anand S, Yusuf S, Vuksan V, et al. The Study of Health Assessment and Risk in Ethnic groups (SHARE): Rationale and design. Can J Cardiol 1998;14:1349–57. [20] Anand S, Yusuf S, Vuksan V, et al. Differences in risk factors, atherosclerosis, and cardiovascular disease between ethnic groups in Canada: the Study of Health Assessment and Risk in Ethnic Groups (SHARE). Lancet 2000;356:278–84. [21] Anand S, Yusuf S, Jacobs R, et al. Risk factors, atherosclerosis, and cardiovascular disease among Aboriginal people in Canada: the Study of Health Assessment and Risk Evaluation in Aboriginal Peoples (SHARE-AP). Lancet 2001;358:1147–53. [22] Wilson PW, D’Agostino RB, Levy D, Belanger AM, Silbershatz H, Kannel WB. Prediction of coronary heart disease using risk factor categories. Circulation 1998;97:1837–47. [23] Michos ED, Nasir K, Braunstein JB, et al. Framingham risk underestimates subclinical atherosclerosis risk in asymptomatic women. Atherosclerosis 2006;184:201–6. [24] Coles DR, Smail MA, Negus IS, Wilde P, et al. Comparison of radiation doses from multislice computed tomography coronary angiography and conventional diagnostic angiography. J Am Coll Cardiol 2006;47:1840–5. [25] McClelland RL, Chung H, Detrano R, et al. Distribution of coronary artery calcium by race, gender, and age: results from the Multi-Ethnic Study of Atherosclerosis (MESA). Circulation 2006;113:30–7. [26] Howard G, Sharrett A, Heiss G, Evans G, Chambless L, Riley W, Burke G. For the ARIC Investigators. Carotid artery intimal-medial thickness distribution in general populations as evaluated by B-mode ultrasound. Stroke 1993;24:1297–304. [27] Stein JH, Tzou WS, Douglas PS, Srinivasan S, Li S, Berenson GS. Predictors of carotid artery intima-media thickness in a large cohort of young adults: The Bogalusa Heart Study. Circulation 2005;112:II-819 [abstract]. [28] Allan PL, Mowbray PI, Lee AJ, Fowkes FG. Relationship between carotid intima-media thickness and symptomatic and asymptomatic peripheral arterial disease. The Edinburgh Artery Study. Stroke 1997;28:348–53.