The differential association between various anthropometric indices of obesity and subclinical atherosclerosis

Atherosclerosis 207 (2009) 232–238

Contents lists available at ScienceDirect

Atherosclerosis journal homepage: www.elsevier.com/locate/atherosclerosis

The differential association between various anthropometric indices of obesity and subclinical atherosclerosis Raymond T. Yan a,1 , Andrew T. Yan a,1 , Todd J. Anderson b,1 , Jean Buithieu c,1 , Francois Charbonneau b,1 , Lawrence Title d,1 , Subodh Verma e,1 , Eva M. Lonn f,∗,1 a

Division of Cardiology, St. Michael’s Hospital, University of Toronto, Toronto, Ontario, Canada Department of Cardiac Sciences and the Libin Cardiovascular Institute, University of Calgary, Calgary, Alberta, Canada Cardiology Division, McGill University, Montreal, Quebec, Canada d Department of Medicine, Division of Cardiology, University of Dalhousie, Halifax, Nova Scotia, Canada e Division of Cardiovascular Surgery, University of Toronto, Toronto, Ontario, Canada f Department of Medicine, Division of Cardiology, McMaster University, 237 Barton Street East, Hamilton, Ontario, Canada L8L 2X2 b c

a r t i c l e

i n f o

Article history: Received 22 September 2008 Received in revised form 10 February 2009 Accepted 31 March 2009 Available online 16 April 2009 Keywords: Carotid intimal-medial thickness Subclinical atherosclerosis Obesity Body mass index Waist-to-hip ratio Anthropometry

a b s t r a c t Background: Recent observational studies have reported differential quantitative relationships between the different anthropometric indices of obesity and risk for cardiovascular (CV) events. Specifically, waist circumference and waist-to-hip ratio (WHR) as crude measures of abdominal obesity were shown to be more predictive of CV events than body mass index (BMI). However, it remains undetermined whether indices of abdominal obesity are also more strongly associated with early subclinical atherosclerosis in asymptomatic individuals. Methods: The associations between carotid intimal-medial thickness (cIMT) as a validated marker of subclinical atherosclerosis and each of BMI, waist circumference and WHR were compared among 1578 middle-aged men free of clinical CV disease enrolled in the Fire Fighter and Their Endothelium (FATE) study. Results: In univariate analyses, the correlation with cIMT as well as the ability to predict substantially increased atherosclerotic burden (cIMT > 75% percentile of the cohort) was strongest for WHR, intermediate for waist circumference, and weakest for BMI (Pearson’s coefficient of 0.21, 0.18 and 0.12, respectively; area under the receiver operating characteristics curve [AUC] of 0.65, 0.62 and 0.58, respectively, P < 0.01 for differences). Within each traditional BMI category, WHR uniformly outperformed waist circumference in further refining discrimination for increased atherosclerotic burden. In multivariable analyses, WHR consistently demonstrated the strongest graded independent relationship with cIMT, beyond most of the established risk factors of atherosclerosis, and superseded both waist circumference and BMI. Conclusion: Our findings support the use of WHR for estimating adiposity-related atherosclerotic burden in clinical practice and in obesity research. Moreover, our study suggests that the increased CV risk associated with abdominal obesity may be mediated in part by the increased anatomic extent of atherosclerotic vascular disease. © 2009 Elsevier Ireland Ltd. All rights reserved.

1. Introduction Obesity is a public health problem of increasing prevalence worldwide [1–3]. It is often accompanied by a cluster of cardiovascular (CV) risk factors that define the metabolic syndrome and is associated with adverse CV outcomes [4]. Recent epidemiological studies have further suggested that the clinical risk attributable to obesity depends not simply on the extent, but importantly, the

∗ Corresponding author. Tel.: +1 905 526 0970; fax: +1 905 527 5380. E-mail address: [email protected] (E.M. Lonn). 1 On behalf of the Firefighter and Their Endothelium (FATE) Investigators. 0021-9150/$ – see front matter © 2009 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.atherosclerosis.2009.03.053

distribution of the excess adiposity. This is supported by the observations that compared to body mass index (BMI) as a predominant measure of general obesity, indices more reflective of abdominal obesity including waist circumference and waist-to-hip ratio (WHR) are stronger and more robust independent predictors of cardiovascular morbidity and mortality [5–7]. However, it remains uncertain whether there exist a corresponding differential association between these anthropometric indices of different adiposity distributions and the anatomic extent of subclinical atherosclerotic disease. Carotid intimal-medial thickness (cIMT) is a validated noninvasive biomarker of the anatomic extent of atherosclerotic vascular disease and was shown to independently predict CV events. Therefore, we compared the associations between these

R.T. Yan et al. / Atherosclerosis 207 (2009) 232–238

anthropometric indices reflective to different degrees of general vs. abdominal obesity with cIMT among middle-aged men without overt atherosclerotic disease enrolled in the prospective Fire Fighter and Their Endothelium (FATE) study. 2. Methods 2.1. Study design and subjects The design and rationale of the FATE study have been previously published [8]. In summary, it is a Canadian multicentre prospective cohort study of emerging risk factors and biomarkers of subclinical atherosclerosis as predictors of incident CV events. Between March 1999 and October 2003, 1578 active and retired Canadian fire fighters without overt CV disease were enrolled from four Canadian centres and extended follow-up for CV events is planned. Traditional and emerging risk factors were evaluated, and subclinical atherosclerosis was measured at baseline during the study enrollment visit by standardized B-mode ultrasound of the extracranial carotid arteries. At this same baseline enrollment visit, all participants underwent detailed physical examination according to a standardized protocol for measurements of three anthropometric indices widely used in clinical practice and epidemiological studies: (1) BMI—a well-accepted anthropometric index of general obesity; (2) waist circumference—a practical crude marker of abdominal adiposity which when interpreted in isolation also reflects total adiposity [3,9]; (3) WHR—an index of the relative amount of abdominal adiposity. BMI was calculated as weight (in kilogram) divided by height squared (in m2 ). Waist and hip circumferences were measured while standing with reference to external landmarks at the narrowest section of the waist, and at the level of the largest circumference between the iliac crest to the crotch, respectively. All measurements were performed by the same specifically trained study nurse at each site following this predefined measurement procedure. It has been established that measurement protocols based either on bony or external landmarks for waist circumference have no substantial influence on the association between waist circumference thus determined and CV events [10]. WHR was calculated as the ratio of waist divided by hip circumferences. 2.2. Imaging methods High resolution B-mode ultrasound of the common carotid artery was performed by trained sonographers on all participants at enrollment, according to a standardized and validated imaging protocol [8,11]. A transverse scan was first conducted to locate the flow divider between the internal and external carotid arteries and to identify the segment of interest, defined as the far (posterior) wall of the right common carotid artery starting 1 cm below the flow divider and extending 1 cm distally. A circumferential longitudinal scan focusing on this arterial segment was then performed, and at least three frames were recorded and measured. The region with the highest cIMT was then selected from which the mean cIMT of this segment as defined by the averaged distance between the leading edges of the lumen-intima and media-adventitia interfaces was measured. All ultrasound examinations were analyzed offline at the carotid ultrasound core laboratory at Hamilton Health Sciences, McMaster University, by one of two certified readers blinded to clinical data. This core laboratory has extensive experience in this technique as employed in prior epidemiologic studies and clinical trials with intra- and inter-reader intraclass correlation coefficients between 0.92 and 0.96 for repeat measurements among randomly chosen subjects [12,13]. All study participants provided written informed consent and the study was approved by the ethics boards of all participating institutions.

233

2.3. Statistical analysis Continuous variables are expressed as mean ± standard deviation and discrete variables as counts and percentages. Anthropometric indices and traditional risk factors were examined for trends across quartiles of cIMT using the Kendall tau-b correlation test. Logarithmic transformation of cIMT was used due to its skewed distribution. To examine and compare the strength of the unadjusted continuous relationship between each obesity index and cIMT, Pearson’s correlation coefficients were first evaluated and the results were verified by the non-parametric Spearman’s rank correlation test. We defined subclinical vascular disease as cIMT in the upper fourth of the distribution (highest quartile) for the entire cohort. We then computed the area under the receiver operating characteristics curve (AUC) for each obesity index as an unadjusted continuous variable to predict subclinical vascular disease thus defined. Significant differences in AUC were tested using the method described by DeLong et al. [14]. As waist circumference has been shown to further refine CV risk prediction within traditional clinically defined BMI categories (normal: <25 kg/m2 ; overweight: 25–29.9 kg/m2 ; obese: ≥30 kg/m2 ) [7], we additionally examined for trends of cIMT across tertiles of waist circumference and WHR within each of these BMI strata. Given that age is a well-recognized and dominating correlate of both body adiposity and subclinical atherosclerosis, we performed age-stratified analyses (by median of population) to characterize the effects of age on the relationships between indices of adiposity and cIMT. The independent association between each of these indices and cIMT among the entire population were further evaluated and compared in two different ways. First, multivariable linear regression models predictive of cIMT (as a continuous variable) were constructed separately for each one of the three adiposity indices entered as independent variables, with stepwise forward selection (P < 0.05) of traditional risk factors and biomarkers of known association with cIMT based on this and prior studies as covariates. To contrast the strength of association of the three obesity indices with cIMT, the extent of covariates adjusted beyond which they, respectively retained independent association with cIMT was compared. In the second method, a regression model was constructed with simultaneous entry of all three indices of adiposity to directly compare their relative independent association with cIMT beyond parameters of known association with cIMT. Statistical analyses were performed using Statistical Package for Social Science version 16.0 (SPSS Inc., Chicago, IL) and STATA 10 (STATA Inc., College Station, TX). Statistical significance was defined as 2-sided P-value < 0.05. 3. Results 3.1. Clinical characteristics of the study population The baseline characteristics of the FATE cohort overall and by quartiles of cIMT are shown in Table 1. On average, participants were overweight, but their blood pressure, glucose and cholesterol levels were within target ranges for primary disease prevention, and most participants were of low to intermediate risk. All three anthropometric indices of obesity, as well as age and major traditional risk factors exhibited significant graded increase across cIMT quartiles (Table 1). 3.2. Relationships between anthropometric indices and cIMT The cIMT measurements by quartiles of BMI, waist circumference and WHR among the entire cohort are depicted in Fig. 1a–c, respectively. cIMT increased across ascending quartiles of each obe-

234

R.T. Yan et al. / Atherosclerosis 207 (2009) 232–238

Table 1 Baseline characteristics of the entire FATE cohort and by subgroups stratified by quartiles of carotid intimal-medial thickness.

Age (year) History of smoking Systolic blood pressure (mmHg) Diastolic blood pressure (mmHg) Body mass index (BMI) (kg/m2 ) Waist circumference (cm) Waist-to-hip ratio (WHR) Total cholesterol (mmol/L) LDL cholesterol (mmol/L) HDL cholesterol (mmol/L) Triglyceride (mmol/L) Fasting glucose (mmol/L) Fasting insulin (pmol/L) HOMA-IR hs-CRP (mg/dL)

All subjects (N = 1574)

1st cIMT quartile (median = 0.51 mm; IQR 0.40–0.60 mm)

2nd cIMT quartile (median = 0.65 mm; IQR 0.61–0.70 mm)

3rd cIMT quartile (median = 0.76 mm; IQR 0.71–0.77 mm)

4th cIMT quartile (median = 0.90 mm; IQR 0.78–1.79 mm)

P-value for trenda

49 ± 10 859 (55%) 128 ± 17 82 ± 10 28.5 ± 3.6 97.0 ± 9.7 0.91 ± 0.06 5.27 ± 0.98 3.28 ± 0.84 1.24 ± 0.28 1.67 ± 1.19 5.34 ± 0.97 61.2 ± 40.2 2.09 ± 1.63 2.25 ± 4.58

45 ± 9 235 (46%) 127 ± 15 82 ± 10 28.1 ± 3.4 95.0± 9.0 0.90 ± 0.05 5.17 ± 0.97 3.19 ± 0.84 1.26 ± 0.29 1.58 ± 1.05 5.2 ± 0.9 54.4 ± 33.6 1.81 ± 1.38 2.11 ± 4.03

48 ± 8 143 (51%) 126 ± 15 82 ± 9 28.5 ± 3.4 97.0 ± 9.0 0.91 ± 0.05 5.35 ± 0.94 3.40 ± 0.81 1.23 ± 0.26 1.69 ± 1.30 5.3 ± 0.9 59.1 ± 36.8 1.99 ± 1.52 1.95 ± 3.35

49 ± 10 229 (56%) 126 ± 16 81 ± 10 28.2 ± 3.5 97.0 ± 10.0 0.91 ± 0.06 5.22 ± 0.99 3.22 ± 0.81 1.24 ± 0.27 1.75 ± 1.42 5.4 ± 1.1 62.1 ± 42.1 2.12 ± 1.67 2.43 ± 6.19

57 ± 9 240 (70%) 135 ± 20 82 ± 11 29.3 ± 3.9 100.0 ± 10.0 0.94 ± 0.05 5.44 ± 0.99 3.41 ± 0.87 1.25 ± 0.29 1.73 ± 0.98 5.5 ± 1.0 71.6 ± 48.7 2.50 ± 1.88 2.54 ± 4.06

<0.001 <0.001 <0.001 0.54 <0.001 <0.001 <0.001 0.001 0.002 0.48 0.003 <0.001 <0.001 <0.001 <0.001

cIMT = carotid intimal-medial thickness; HOMA-IR = homeostasis model of insulin resistance; hs-CRP = high sensitivity C-reactive protein; IQR = interquartile range. a Trend across subgroups stratified by carotid intimal-medial thickness quartiles.

Fig. 1. (a) Carotid intimal-medial thickness across quartiles of body mass index. (b) Carotid intimal-medial thickness across quartiles of waist circumference. (c) Carotid intima-media thickness across quartiles of waist-to-hip ratio.

R.T. Yan et al. / Atherosclerosis 207 (2009) 232–238

235

sity index (P < 0.001), albeit in the most consistent and stepwise manner with WHR in comparison to both waist circumference and BMI. In univariate analysis, the indices reflective more of abdominal obesity had stronger associations with cIMT, with WHR showing the strongest correlation: BMI (r = 0.12, P < 0.001), waist circumference (r = 0.18, P < 0.001) and WHR (r = 0.21, P < 0.001). Similar results were obtained using non-parametric analyses (data not shown). With reference to the respective ROC curves (Online Supplementary Fig. S1), the two indices predominantly reflective of abdominal adiposity demonstrated better discrimination (greater area under the curve—AUC) than BMI for subclinical vascular disease as defined by cIMT above the 75th percentile of this cohort (AUC BMI: 0.58, 95% CI = 0.55–0.62; waist circumference: 0.62, 95% CI = 0.59–0.65; WHR: 0.65, 95% CI = 0.62–0.68; P = 0.007 for differences between BMI and waist circumference, and P < 0.001 for differences between BMI and WHR). Between these two indices, WHR superseded waist circumference in discriminatory performance (P = 0.008). 3.3. Relationships between waist circumference, WHR and cIMT by traditional BMI categories Stratified according to traditional BMI categories, the distributions of cIMT across tertiles of waist circumference and WHR are illustrated in Fig. 2a and b, respectively. A significant trend of increasing cIMT across tertiles of waist circumference was observed within the overweight BMI category whereas such trend was, respectively weaker and less consistent within the subgroups of normal weight and obese BMI (Fig. 2a). In comparison, cIMT demonstrated more significant trend with stepwise increase across ascending tertiles of WHR consistently within all three BMI strata (Fig. 2b). Furthermore, within each BMI category, WHR uniformly outperformed waist circumference in further refining the discrimination for subclinical vascular disease (AUC waist circumference vs. AUC WHR by BMI strata: 0.58 95% CI = 0.48–0.69 vs. 0.64 95% CI = 0.54–0.74 for BMI < 25 kg/m2 ; 0.61 95% CI = 0.57–0.66 vs. 0.66 95% CI = 0.62–0.70 for BMI = 25–29.9 kg/m2 ; 0.57 95% CI = 0.50–0.63 vs. 0.60 95% CI = 0.53–0.66 for BMI ≥ 30 kg/m2 ). 3.4. Effects of age on the differential relationships between anthropometric indices and cIMT The mean values of cIMT across quartiles of anthropometric indices among the infra-median (<49 years old) and supra-median (≥49 years old) age-groups of the FATE cohort are shown in Fig. 3. Below median age (Fig. 3a), waist circumference and WHR, respectively exhibited borderline (P = 0.050) and more significant (P = 0.002) graded relationships with cIMT. Above median age (Fig. 3b), only WHR retained a significant trend with cIMT (P = 0.043). Compared to waist circumference and BMI, WHR was consistently a more robust correlate with cIMT least confounded by age. 3.5. Differential independent association between various anthropometric indices and IMT Multivariable regression models constructed to contrast the strength of the independent association between the three obesity indices and cIMT (entered as a continuous variable) beyond other CV risk factors are summarized in Table 2. BMI was weakly and independently associated with cIMT only beyond fasting plasma glucose and low-density lipoprotein cholesterol, but not after adjustment for blood pressure, smoking history and serum insulin (an indirect measure of insulin resistance). By comparison, waist circumference and WHR exhibited stronger and more robust independent rela-

Fig. 2. (a) Carotid intimal-medial thickness across tertiles of waist circumference by traditional body mass index categories. For BMI < 25 kg/m2 : P = 0.049 for trend across waist circumference tertiles. For BMI = 25–29.9 kg/m2 : P = 0.001 for trend across waist circumference tertiles. For BMI ≥ 30 kg/m2 : P = 0.01 for trend across waist circumference tertiles. (b) Carotid intimal-medial thickness across tertiles of waist-to-hip ratio by traditional body mass index categories. For BMI < 25 kg/m2 : P = 0.002 for trend across WHR tertiles. For BMI = 25–29.9 kg/m2 : P < 0.001 for trend across WHR tertiles. For BMI ≥ 30 kg/m2 : P = 0.007 for trend across WHR tertiles.

tionships with cIMT above and beyond all these variables (Table 2). Upon further inclusion of age into the models, the independent association between each of the anthropometric indices and cIMT were uniformly abrogated due to the dominant confounding effects of age, a surrogate of cumulative atherosclerotic risk factors exposure, which by far is the strongest correlate with both obesity and cIMT. Furthermore, the independent associations between the three anthropometric obesity indices and cIMT were directly compared in regression models as summarized in Table 3. After simultaneous adjustment for each other, WHR emerged to be the only anthropometric index of adiposity exhibiting an independent relationship with cIMT. This cross-sectional relationship was attenuated but remained significant after adjustment for the known measurable determinants of cIMT (Table 3), with the exception of age.

236

Body mass index (per 1 SD) Fasting glucose (per mmol/l) Low-density cholesterol (per mmol/L)

Model predictive of cIMT with body mass indexa

Model predictive of cIMT with waist circumferencea ␤-Coefficient (95% CI)

␤-Coefficient (95% CI)

P-value

0.010 (0.004–0.015)

<0.001

0.008 (0.002–0.014) 0.010 (0.004–0.016)

Model predictive of cIMT with waist-to-hip ratioa ␤-Coefficient (95% CI)

P-value

Waist circumference (per 1 SD)

0.008 (0.002–0.015)

0.013

Waist-to-hip ratio (per 1 SD)

0.012 (0.006–0.018)

<0.001

0.007

Fasting glucose (per mmol/L)

0.002 (−0.004–0.008)

0.45

Fasting glucose (per mmol/L)

0.002 (−0.004–0.008)

0.52

0.001

Low-density cholesterol (per mmol/L)

0.008 (0.002–0.014)

0.014

Low-density cholesterol (per mmol/L)

0.007 (0.001–0.013)

0.021

Fasting insulin (per pmol/L) Systolic blood pressure (per mmHg) History of smoking

0.00 (0.001–0.001) 0.001 (0.001–0.001) 0.032 (0.022–0.042)

0.15 <0.001 <0.001

Fasting insulin (per pmol/L) Systolic blood pressure (per mmHg) History of smoking

0.00 (0.00–0.00) 0.001(0.000–0.001) 0.030 (0.020–0.041)

0.13 0.001 <0.001

P-value

cIMT = carotid intimal-medial thickness; 95% CI = 95% confidence interval; SD = standard deviation. a Regression models constructed using forced entry of the corresponding obesity index with stepwise forward selection (P < 0.05) of covariates of known univariate association with cIMT—fasting glucose, low-density lipoprotein cholesterol, fasting insulin, systolic blood pressure, and history of smoking. The regression models as shown illustrate the extent of covariates adjusted beyond which the respective obesity index retained independent correlation with cIMT.

R.T. Yan et al. / Atherosclerosis 207 (2009) 232–238

Fig. 3. (a) Carotid intimal-medial thickness across quartiles of anthropometric indices among the infra-median age (<49 years old) population. P-values for trends: BMI: 0.21; waist circumference: 0.050; WHR: 0.002. (b) Carotid intimal-medial thickness across quartiles of anthropometric indices among the supra-median age (≥49 years old) population. P-values for trends: BMI: 0.29; waist circumference: 0.15; WHR: 0.043.

4. Discussion

In this middle-aged cohort of men free of clinical CV disease, we observed a differential strength of association between the various anthropometric indices of body adiposity with cIMT, an extensively validated surrogate marker of subclinical atherosclerosis [15–17]. The correlation with cIMT as well as the discriminatory performance for subclinical vascular disease was strongest for WHR, intermediate for waist circumference, and weakest for BMI. Importantly, both waist circumference and WHR add to BMI in further refining the discrimination for subclinical vascular disease, and WHR superseded waist circumference in this added value. Of these three simple bedside measures of obesity commonly used in contemporary clinical practice and epidemiological studies, WHR consistently demonstrated the strongest and most robust continuous independent relationship with cIMT, beyond most known determinants of subclinical atherosclerosis and least confounded

Table 2 Multivariable linear regression models of the associations between body mass index, waist circumference and waist-to-hip ratio to carotid intimal-medial thickness.

R.T. Yan et al. / Atherosclerosis 207 (2009) 232–238

237

Table 3 Models with carotid intimal-medial thickness (entered as a continuous variable) as the dependent variable and body mass index, waist circumference and waist-to-hip ratio as predictors with and without adjustment for parameters of known correlation with carotid intimal-medial thickness.

Fasting glucose (per mmol/L) Low-density cholesterol (per mmol/L) Fasting insulin (per pmol/L) Systolic blood pressure (per mmHg) History of smoking Body mass index (per 1 SD) Waist circumference (per 1 SD) Waist-to-hip ratio (per 1 SD)

Model predictive of cIMT with anthropometric indicesa

Model predictive of cIMT with anthropometric indices after adjustment for clinical parametersb

␤-Coefficient (95% CI)

P-value

␤-Coefficient (95% CI)

P-value

P = 0.47 P = 0.20 P < 0.001

0.002 (−0.004–0.008) 0.007 (0.001–0.013) 0.0002 (−0.0001–0.0003) 0.001 (0.0005–0.001) 0.030 (0.020–0.040) −0.007 (−0.017–0.002) 0.004 (−0.009–0.016) 0.012 (0.004–0.021)

0.48 0.019 0.06 <0.001 <0.001 0.13 0.54 0.005

– – – – – −0.003 (−0.012–0.006) 0.008 (−0.004–0.020) 0.017 (0.008–0.025)

cIMT = carotid intimal-medial thickness; 95% CI = 95% confidence interval; SD = standard deviation. a Regression models constructed with forced entry of all three obesity indices for comparative evaluation of their relative independent association with cIMT. b Regression models constructed with forced entry of BMI, waist circumference, WHR, and forward selection of parameters of known association with cIMT—fasting glucose, low-density lipoprotein cholesterol, fasting insulin, systolic blood pressure, and history of smoking. Additional parameters including homeostatic model assessment of insulin resistance (HOMA) and high-sensitivity C-reactive protein (hs-CRP) were not significant when entered into the model and did not alter the significance of the other covariates as shown.

by age, and surpassed waist circumference and in turn BMI as the anthropometric parameter most independently correlated with cIMT. These findings are concordant with recent observations from large epidemiological studies, which showed stronger associations between indices more specific of abdominal adiposity and particularly WHR to the risk of prevalent [5,7] and incident [6] CV disease. However, our study is novel and extends such observations by demonstrating that these indices and particularly WHR are also more closely linked to subclinical atherosclerotic vascular disease. This observation is important, as it suggests that the pathophysiological link between obesity, particularly excessive abdominal adiposity and CV events may be mediated at least in part by increased anatomic extent of atherosclerosis. Accumulating evidence from large population-based studies suggests that indices of primarily abdominal adiposity are more robustly associated with CV risks than indices of general adiposity, such as weight or BMI. The clinical implications and contribution of abdominal adiposity to global cardiometabolic risk have been extensively reviewed [3,9]. In the International Day for the Evaluation of Abdominal Obesity (IDEA) study of 168,000 primary care patients, a significant gradient was noted in the frequency of CV disease with both BMI and waist circumference, albeit a stronger relationship for waist circumference than for BMI, homogeneously across geographic regions and gender [7]. Findings from IDEA also importantly highlighted that waist circumference measurement, in addition to BMI, refines discrimination for prevalent CV disease [7,9]. Of note, WHR was not measured in this cross-sectional investigation. In the multi-ethnic case–control INTERHEART study involving more than 27,000 participants, as compared to both BMI and waist circumference, WHR showed a stronger independent association with the prevalence of myocardial infarction. This finding was consistent in both women and men, across all age and ethnic groups studied [5]. More recently, a prospective evaluation of the relationship between adiposity indices and incident coronary artery disease, the European Prospective Investigation Into Cancer and Nutrition in Norfolk (EPIC-Norfolk) study, also confirmed the superior prognostic value of WHR above that of waist circumference and BMI [6]. In this cohort, an independent continuous and graded relationship between WHR and incident coronary events was observed, which persisted after adjustment for BMI and conventional risk factors. These studies support the superior clinical utility and prognostic value of WHR in the context of obesity-related cardiovascular disease. Few studies have reported the relationship between different bedside anthropometric obesity indices and subclinical vascular disease, an early manifestation of atherosclerosis. Several such

studies have evaluated associations between indices of adiposity and cIMT [19–22]. However, these prior investigations [20–22] were much smaller than our study and some have not compared all three commonly used bedside adiposity indices [19,21], or have not adjusted for other risk factors for atherosclerosis and determinants of cIMT, such as cholesterol, insulin resistance and C-reactive protein [19,20,22]. Findings from these earlier reports have been inconclusive, as the independent association between atherosclerotic burden and the traditionally favored general obesity index BMI was shown to be similar to that of WHR and waist circumference in some studies [20,21], and non-significant in others [22], However, our findings of the strongest association between WHR and cIMT are concordant with the recent Dallas Heart Study, which demonstrated a stronger relationship between WHR (compared with waist circumference and BMI) and subclinical atherosclerosis determined qualitatively by the presence of coronary artery calcification on computed tomography and abdominal aortic plaque on magnetic resonance imaging [23]. These findings and our study suggest better discrimination of atherosclerotic burden afforded by WHR, which mirrors its superior prognostic value in the context of CV disease attributed to obesity. Visceral obesity, more specifically surrogated by WHR and waist circumference than BMI, has been related to adverse lipid and adipokine profiles, atherogenic cholesterol particles, insulin resistance, hemostatic disorders, hyperandrogenicity and glucocorticoid excess, hypertension as well as endothelial dysfunction [3,9], which are all deleterious metabolic and biological derangements implicated in the pathogenesis of atherosclerosis [24,25]. This may in part explain the stronger association between cIMT and indices of abdominal obesity as compared to BMI. Furthermore, some of these metabolic disturbances had also been shown to exhibit an inverse relationship with hip circumference, which is factored into WHR, but not the waist circumference measurement [26]. In our study population consisting largely of overweight (N = 902, 57%) asymptomatic men, the overall correlations between anthropometric indices and cIMT were only modest. We observed that the highly significant association between each anthropometric index and cIMT that is independent of major traditional and novel risk factors was no longer evident after additionally controlling for age—a dominant surrogate of cumulative lifelong exposure to all yet known and unknown atherosclerotic risk factors. It is known that age-associated increases in carotid wall thickness are also related to a multitude of age-associated morphologic, cellular, enzymatic, and biochemical changes in the arterial wall, which are not entirely explained by body adiposity and CV risk factors

238

R.T. Yan et al. / Atherosclerosis 207 (2009) 232–238

and may not necessarily reflect atherosclerosis [18]. It is thus conceivable that the modest yet important cross-sectional relationship between obesity indices and cIMT can be masked by the dominant, collectively known and unknown confounding effects accumulated over time with aging. Consistent with this notion is our finding, on age-stratified analysis, of less consistent relationships suggestive of greater confounding between all anthropometric indices and cIMT at more advanced age. However, importantly, our age-stratified analysis (below and above the median age for the cohort) explicitly illustrated that WHR (compared to waist circumference and BMI) was more closely correlated with cIMT among individuals of different ages. All our findings are consistently supportive of the superior value of WHR among the three anthropometric indices for estimating adiposity-related atherosclerotic burden in clinical practice and obesity research. Future investigations based on larger populations of more diversified demographics, risk factor and anthropometric profiles are encouraged to confirm and extend the generalizability of our findings. Our study is applicable only to the population studied, namely, apparently healthy middle-aged Caucasian males without clinical atherosclerotic vascular disease. Generalization of our findings to other populations should await confirmation from future studies. Furthermore, this cross-sectional analysis does not allow us to determine the temporal relationship between indices of obesity and the progression of subclinical atherosclerosis. In conclusion, our findings support the use of WHR over waist circumference and BMI both in clinical practice in the evaluation of CV risk and in obesity-related atherosclerosis research. The association between WHR and the extent of subclinical atherosclerosis may indicate the pathogenic link between excessive abdominal adiposity and the development of CV events. Role of the funding sources The Firefighter and Their Endothelium (FATE) study is supported by grants from The Canadian Institute of Health Research, The Heart and Stroke Foundation of Canada, and Pfizer Canada. The sponsors had no involvement in the study conception or design; collection, analysis, and interpretation of data; in the writing, review, or approval of the manuscript; and in the decision to submit the manuscript for publication. Acknowledgements We thank all the study nurses, research coordinators, sonographers and participants in the FATE study. We are most grateful to their participation and contribution. Dr. R. Yan is supported by Fellowship Award from the Canadian Institutes of Health Research and the Detweiler Travelling Fellowship Award from the Royal College of Physicians and Surgeons of Canada. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.atherosclerosis.2009.03.053. References [1] Ogden CL, Carroll MD, Curtin LR, et al. Prevalence of overweight and obesity in the United States 1999–2004. JAMA 2006;295(13):1549–55.

[2] Mclellan F. Obesity rising to alarming levels around the world. Lancet 2002;359: 1412. [3] Deprés J-P, Cartier A, Côté M, et al. The concept of cardiometabolic risk: bridging the fields of diabetology and cardiology. Ann Med 2008;40(7):514–23. [4] Grundy SM, Cleeman JI, Daniels SR, et al. American Heart Association; National Heart, Lung, and Blood Institute. Diagnosis and management of the metabolic syndrome: an American Heart Association/National Heart, Lung, and Blood Institute Scientific Statement. Circulation 2005;112(17):2735–52. [5] Yusuf S, Hawken S, Ounpuu S, et al. INTERHEART Study Investigators. Obesity and the risk of myocardial infarction in 27 000 participants from 52 countries: a case–control study. Lancet 2005;366(9497):1640–9. [6] Canoy D, Boekholdt SM, Wareham N, et al. Body fat distribution and risk of coronary heart disease in men and women in the European Prospective Investigation Into Cancer and Nutrition in Norfolk cohort: a population-based prospective study. Circulation 2007;116(25):2933–43. [7] Balkau B, Deanfield JE, Deprés J-P, et al. International day for the evaluation of abdominal obesity (IDEA). A study of waist circumference, cardiovascular disease, and diabetes mellitus in 168,000 primary care patients in 63 countries. Circulation 2007;116(17):1942–51. [8] Anderson TJ, Roberts AC, Hildebrand K, et al. The fate of endothelial function testing: rationale and design of the Firefighters And Their Endothelium (FATE) study. Can J Cardiol 2003;19(1):61–6. [9] Deprés J-P, Lemieux I, Bergeron J, et al. Abdominal obesity and the metabolic syndrome: contribution to global cardiometabolic risk. Arterioscler Thromb Vasc Biol 2008;28:1039–49. [10] Ross R, Berentzen T, Bradshaw AJ, et al. Does the relationship between waist circumference, morbidity and mortality depend on measurement protocol for waist circumference? Obes Rev 2008;9(4):312–25. [11] Yan RT, Anderson TJ, Charbonneau F, et al. Relationship between carotid artery intima-media thickness and brachial artery flow-mediated dilation in middleaged healthy men. J Am Coll Cardiol 2005;45:1980–6. [12] Anand SS, 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(9226): 279–84. [13] Lonn E, Yusuf S, Dzavik V, et al. SECURE Investigators. Effects of ramipril and vitamin E on atherosclerosis: the study to evaluate carotid ultrasound changes in patients treated with ramipril and vitamin E (SECURE). Circulation 2001; 103(7):919–25. [14] DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics 1988;44:837–45. [15] Taylor AJ, Merz CNB, Udelson JE. 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. [16] Redberg RF, Vogel RA, Criqui MH, et al. 34th Bethesda conference task force #3. What is the spectrum of current and emerging techniques for the noninvasive measurement of atherosclerosis? J Am Coll Cardiol 2003;41:1886–98. [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 Atherosclerosis Risk in Communities (ARIC) Study 1987–1993. Am J Epidemiol 1997;146(6):483–94. [18] Stein J, Korcarz CE, Hurst RT, et al. Use of carotid ultrasound to identify subclinical vascular disease and evaluate cardiovascular disease risk: a consensus statement from the American society of echocardiography carotid intimamedia thickness task force. J Am Soc of Echocardiogr 2008;21:93–111. [19] Folsom AR, Eckfeldt JH, Weitzman S, et al. Relation of carotid artery wall thickness to diabetes mellitus, fasting glucose and insulin, body size, and physical activity Atherosclerosis Risk in Communities (ARIC) Study Investigators. Stroke 1994;25(1):66–73. [20] Takami R, Takeda N, Havashi M, et al. Body fatness and fat distribution as predictors of metabolic abnormalities and early carotid atherosclerosis. Diabetes Care 2001;24(7):1248–52. [21] De Michele M, Panico S, Iannuzzi A, et al. Association of obesity and central fat distribution with carotid artery wall thickening in middle-aged women. Stroke 2002;33:2923–8. [22] Czernichow S, Bertrais S, Oppert JM, et al. Body composition and fat repartition in relation to structure and function of large arteries in middle-aged adults. Int J Obes (Lond) 2005;29(7):826–32. [23] See R, Abdullah SM, McGuire DK, et al. The association of differing measures of overweight and obesity with prevalent atherosclerosis. J Am Coll Cardiol 2007;50:752–9. [24] Carey DGP. Abdominal obesity. Curr Opin Lipidol 1998;9:35–40. [25] Steinberg HO, Chaker H, Leaming R, et al. Obesity/insulin resistance is associated with endothelial dysfunction: implication for the syndrome of insulin resistance. J Clin Invest 1996;97:2601–10. [26] Okura T, Nakata Y, Yamabuki K, et al. Regional body composition changes exhibit opposing effects on coronary heart disease risk factors. Arterioscler Thromb Vasc Biol 2004;24:923–9.