Prediction of Cardiovascular Events with Consideration of General and Central Obesity Measures in Diabetic Adults: Results of the 8.4-Year Follow-Up

Abstract

Background:

Obesity is one of the most important cardiovascular disease (CVD) risk factors among diabetic populations. We evaluated the ability of different anthropometric measures for predicting CVD among type 2 diabetic patients.

Methods:

The study consisted of 411 men and 599 women, aged ≥30 years, free of CVD at baseline with a median follow-up of 8.4 years. The adjusted hazard ratios (HRs) for CVD were calculated for a 1 standard deviation change in body mass index (BMI), waist circumference (WC), waist-to-hip ratio (WHR), and waist-to-height ratio (WHtR) using Cox proportional regression analysis.

Results:

A total of 188 CVD events occurred (men, 90; women, 98). In women, in confounder-adjusted analysis [age, fasting plasma glucose (instead of glycosylated hemoglobin), and positive family history of CVD], WHR was associated with incident CVD [1.32 (1.06–1.65)], followed by WC and WHtR, which were marginally significant (P=0.06 and 0.08, respectively); after adjustment for hypertension and hypercholesterolemia, only WHR predicted CVD significantly. In men, the confounder-adjusted (age, fasting plasma glucose, and aspirin use) HR to predict CVD was significant only for WHR [HR 1.21(1.00–1.48)].

Conclusion:

This study showed WHR was the most powerful predictor of CVD among anthropometric measures, followed by WHtR, in diabetic population.

Introduction

T he prevalence of type 2 diabetes is increasing worldwide as well as in Middle Eastern countries. It is estimated that the number of diabetic patients will increase from 170 million in 2000 to 366 million in 2030, and it is expected that the prevalence of diabetes will reach to the highest by 2030 in Middle Eastern countries.^1
–3 Cardiovascular disease (CVD) is the major cause of death in diabetic individuals, causing more than 75% of total mortality in diabetic patients.^4
–6 Hence, much attention must be paid to reduce CVD risk factors in this population,^7,8 and one of the most important of these risk factors is obesity.⁹

Although there is much data on the link between various components of general and central obesity with CVD in the general population,^10,11 there are controversies over this relationship in diabetic populations.^1,6 Some studies have demonstrated that body mass index (BMI) is positively associated with CVD, whereas others have reported that there is no association or there is an inverse association between the two.^12,13 Also, the effect of central obesity on CVD among diabetic populations has been rarely studied.¹⁴

Considering the ethnic differences in CVD risk factors between different cultures or between diabetic and nondiabetic subjects,¹⁵ the heavy burden that CVD imposes in developing countries, and the limited available studies on obesity in this population,^16,17 the aim of this study was to evaluate the ability of different parameters of central and general obesity to predict CVD among Iranian men and women with type 2 diabetes.

Materials and Methods

Study population

Subjects of this study were selected from among participants of the Tehran Lipid and Glucose Study (TLGS), a prospective study aimed at determining the risk factors and outcomes for noncommunicable disease, performed on a representative sample of 15,005 people, aged 3 years and over, who were residents of district 13 of Tehran. Subjects were categorized into the cohort and intervention groups, the latter to be educated for implementation of lifestyle modifications. Among this overall group, 1,358 subjects, aged ≥30 years, had diabetes and were evaluated in a cross-sectional phase of the TLGS (February, 1999, to August, 2001). Subjects with a history of CVD at baseline (n=216) and those who had missing data on anthropometric variables (n=15) were excluded, leaving 1,127 subjects, of whom 1,010 subjects (men, 411; women, 599) were followed up until March, 2009, for a median of 8.4 years (response rate ≈89.6 %). The ethical committee of the Research Institute for Endocrine Sciences approved the proposal of this study, and informed written consent was obtained from all subjects.

Clinical and laboratory measurements

Using a pretested questionnaire, a trained interviewer collected information that included demographic data, past medical history, family history of CVD, medication use, and smoking behavior. Weight was measured using digital scales (Seca 707, Seca Corp., Hanover, MD; range 0.1–150 kg) while subjects were minimally clothed and without shoes and recorded to the nearest 100 grams. Height was measured in a standing position without shoes and with shoulders in normal alignment, using a tape meter. BMI was calculated as weight (kg) divided by square of height (m²). Anthropometric measurements are taken with shoes removed and the participants wearing light clothing in standing position; waist circumference (WC) was measured at umbilical level, using an unstretched tape meter, without any pressure to body surface and hip circumference (HC) at the at the widest girth of the hip. Waist-to-hip ratio (WHR) and waist-to-height ratio (WHtR) were calculated as WC (cm) divided by HC (cm) and height (cm), respectively. On the right arm, after a 15-min rest in a sitting position, two measurements of systolic blood pressure (SBP) and diastolic blood pressure (DBP) were taken using a standardized mercury sphygmomanometer (calibrated by the Iranian Institute of Standards and Industrial Researches), and the mean of the two measurements was considered as subject's blood pressure.

A blood sample was drawn between 7:00 and 9:00 a.m. from all study participants after 12–14 h of overnight fasting. All of the blood analyses were done at the TLGS research laboratory on the day of blood collection. Plasma glucose was measured using an enzymatic colorimetric method with glucose oxidase. The standard oral glucose tolerance test (OGTT) was performed for all participants who were not on glucose-lowering drugs. Total cholesterol (TC) was assayed using the enzymatic colorimetric method with cholesterol esterase and cholesterol oxidase. These analyses were performed using commercial kits (Pars Azmoon Inc., Tehran, Iran) and a Selectra 2 auto-analyzer (Vital Scientific, Spankeren, The Netherlands). The intra- and interassay coefficients of variation (CV) were both 2.2% for glucose. For TC, intra- and interassay CVs were 0.5% and 2%, respectively.

CVD outcome

Details of the collection of cardiovascular outcome data have been published earlier.¹⁸ To summarize, each participant was followed up annually by phone call for any medical event. A trained nurse questioned participants regarding any medical condition and then a trained physician collected complementary data about that event, during a home visit and by acquisition of data from medical files. The collected data were then evaluated by an outcome committee consisting of an internist, an endocrinologist, a cardiologist, an epidemiologist, and other experts, when needed, to assign a specific outcome for every event. In the current study, the events targeted were the first CVD events, including definite myocardial infarction [MI; with diagnostic electrocardiogram (ECG) and biomarkers] (22 patients, 11.7%), probable MI (positive ECG findings plus cardiac symptoms or signs plus missing biomarkers or positive ECG findings plus equivocal biomarkers) (9 patients, 4.7%), unstable angina (new cardiac symptoms or changing symptom patterns and positive ECG findings with normal biomarkers) (36 patients, 19.1%), angiographic-proven coronary heart disease (CHD) (65 patients,34.5%), stroke (as defined by a new neurological deficit that lasted more than 24 h) (20 patients, 10.6%), and death from CVD (36 patients, 19.1%).

Definition of terms

Diabetes mellitus was defined as fasting plasma glucose (FPG) ≥126 mg/dL, 2-h plasma glucose ≥200 mg/dL, current use of antidiabetic drugs, or a positive answer to the question “have you ever been told by a doctor that you have diabetes (high blood sugar disease)?” Hypertension (HTN) was defined as SBP ≥140 mmHg and/or DBP ≥90 mmHg, or current use of antihypertensive medication. Subjects with TC ≥240 mg/dL or those using antilipid drugs were defined as being hypercholesterolemic. A positive family history of premature CVD reflected any prior diagnosis of CVD by a physician, in any female first-degree relative under 65 years old and any male first-degree relative aged less than 55 years. Menopause was defined as absence of spontaneous menstrual bleeding for more than 12 months. Smoking status included current or past regular use of cigarettes.

Statistics

Mean [standard deviation (SD)] values for continuous and frequencies (%) for categorical variables of the baseline characteristics were expressed for diabetic men and women with and without CVD. Because FPG had a skewed distribution, it was shown as median (interquartile range). Comparison of baseline characteristics between diabetic men and women with and without CVD was done by the Student t-test for continuous variables, chi-squared test for categorical variables, and Mann–Whitney U-test for the skewed variable.

The Cox proportional hazard model was used to study the association between various anthropometric variables and CVD outcome. Follow-up duration was defined as the period between entrance to the study and the end points; the end point was considered as the first CVD event and censoring, defined as leaving the residence area, loss to follow-up or non-CVD death or follow-up until March 8, 2009. To select the covariates to be included in the multivariate Cox models, univariate analysis was used for each candidate covariate (age, positive family history of premature CVD, intervention group, FPG, smoking status, aspirin use in both genders, and menopause status in women). Then, each covariate with a P value less than 0.2, was selected to be included in a stepwise backward (P remove, 0.1) multivariate Cox regression analyses.

Final selected confounders were age, FPG, and aspirin use in diabetic men, and age, FPG, and positive family history of premature CVD in diabetic women. Adjusted hazard ratios (HRs), with 95% confidence intervals (CI), were calculated for every 1 SD increase in the value of each anthropometric variables in three models (model 1, adjusted for age; model 2, adjusted for selected confounders in each gender; model 3, adjusted for confounders and HTN and hypercholesterolemia as mediators). Interactions between each anthropometric measure and the other variables were tested using log-likelihood ratio test of models containing first order interactions.

Likelihood ratio chi-squared and the Akaike Information Criterion (AIC), as a statistical estimate of the trade-off between the likelihood of a model against its complexity, were used to assess the goodness of fit of the predictive models. A lower value of AIC and higher value of likelihood ratio chi-squared indicates a better model fitness. The discriminatory power of the models was calculated by the C index, which was estimated with the “somersd” STATA command that yields the confidence interval for the Harrell C index using jackknife variance estimation. A value of 1 denotes perfect discrimination and a value of 0.5 is no better than chance. The Schoenfild residual test was used for assessment of the proportional hazards assumption in the Cox models, and all proportionality assumptions were appropriate.

STATA software (version 10) was used for data analysis and P values ≤0.05 were considered statistically significant.

Results

The study population consisted of 1,010 diabetic subjects (men, 411; women, 599) with a mean age of 54.8 years. Among them, 241 subjects just had positive OGTT values, 159 patients were identified only by the positive response to the relevant question at baseline, 24 subjects only had high FPG, 3 patients were on antidiabetic drugs, and the other patients had a combination of these criteria. There was no significant difference in age, CVD risk factors, and anthropometric variables between subjects followed and those not followed. During a median follow-up of 8.4 years, 188 first CVD events (men, 90; women, 98) occurred (the rate/year event=2.6%).

Baseline characteristics of the diabetic men and women, with and without CVD events, are presented in Table 1. In men, in comparison to subjects without CVD, those with the condition were older and had significantly higher WC, WHR, WHtR, FPG, hypercholesterolemia, HTN, and rate of aspirin consumption. Diabetic women with CVD were older and had higher WC, WHR, WHtR, FPG, hypercholesterolemia, and HTN than the diabetic women without CVD. No significant differences in BMI levels and in prevalence of smoking were found between participants with and without CVD in either gender.

Table 1.

Baseline Characteristics of Diabetic Men and Women, with and without Incident Cardiovascular Events

	Men			Women
	Without CVD (n=321)	With CVD (n=90)	P value	Without CVD (n=501)	With CVD (n=98)	P value
Age	55.7±11.8	59.6±10.4	0.005	52.7±10.6	57.9±8.9	<0.001
FH premature CVD (%)	11.5	15.6	0.30	23.0	31.6	0.06
Cigarette smoking (%)	43.6	46.7	0.6	5.6	4.1	0.53
Aspirin use (%)	16.5	28.7	0.01	16.8	21.3	0.29
Hypertension (%)	35.8	58.9	<0.001	46.4	74.5	<0.001
Hypercholesterolemia (%)	29.6	53.9	<0.001	49.3	71.1	<0.001
Intervention (%)	44.2	38.9	0.36	40.3	39.8	0.92
Body mass index (kg/m²)	27.3±3.9	28.0±3.5	0.13	29.8±4.9	29.7±4.6	0.9
Waist circumference (cm)	94.3±10.7	97.3±9.4	0.02	96.3±11.5	98.8±11.3	0.051
Waist/hip	0.96±0.07	0.99±0.06	0.007	0.90±0.07	0.93±0.08	<0.001
Waist/height	0.56±0.06	0.58±0.05	0.008	0.62±0.07	0.64±0.07	0.02
Fasting plasma glucose (mg/dL)	129 (101–172)	146 (122–204)	0.001	130 (107–177)	162 (125–217)	<0.001

Mean±standard deviation (SD) are shown for continuous variables and the P value is calculated with a t-test; % is shown for categorical variables with the P value according to a chi-squared test; fasting plasma glucose is shown as median (interquartile range) and the P value according to the Mann–Whitney test.

Hypertension was defined as systolic blood pressure ≥140 mmHg or diastolic blood pressure ≥90 mmHg, or current use of antihypertensive medication. Subjects with total cholesterol (TC) ≥240 mg/dL or those using antilipid drugs were defined as hypercholesterolemic. Smoking status included current or past regular use of cigarettes.

CVD, cardiovascular disease; FH, family history.

Table 2 shows the HRs of a 1 SD increase in each anthropometric variable for the first CVD event in three models in diabetic women. In this group, in an age-adjusted model, a 1 SD increase in WHR caused a 32% increase in risk of CVD (P=0.01). After further adjustment according to model 2 (adjusted for confounders), the HR of WHR did not attenuate and remained still significant [HR=1.32 (1.06–1.65), P=0.01], although in this model, WC and WHtR were marginally associated with incident CVD with HR of 1.21 (P=0.06) and 1.18 (P=0.08), respectively. The marginal association of WC and WHtR in women with incident CVD might be partly attributed to the limited statistical power. With our sample size in diabetic women and an α of 0.05, we had the statistical power to detect an HR of 1.32 and 1.20 per 1 SD increment of 60% and 30%, respectively.

Table 2.

Hazard Ratios and Discrimination of Anthropometric Measures for Predicting Cardiovascular Events in Diabetic Women

	Hazard ratio (95% confidence interval)	P value	C Index (95% confidence interval)	Likelihood ratio chi-squared	AIC
Model 1
BMI	1.03 (0.83–1.26)	0.78	0.64 (0.58–0.69)	22.11	1195
WC	1.16 (0.95–1.43)	0.13	0.64 (0.59–0.69)	24.28	1193
WHR	1.32 (1.06–1.64)	0.01	0.66 (0.61–0.71)	28.52	1189
WHtR	1.14 (0.95–1.38)	0.15	0.64 (0.59–0.69)	24.08	1193
Model 2
BMI	1.08 (0.87–1.33)	0.47	0.69 (0.64–0.74)	39.92	1168
WC	1.21 (0.98–1.50)	0.06	0.69 (0.64–0.74)	42.86	1165
WHR	1.32 (1.06–1.65)	0.01	0.70 (0.65–0.75)	45.69	1162
WHtR	1.18 (0.98–1.43)	0.08	0.69 (0.64–0.74)	42.47	1166
Model 3
BMI	0.97 (0.78–1.20)	0.80	0.73 (0.68–0.77)	60.53	1151
WC	1.11 (0.89–1.38)	0.32	0.73 (0.68–0.77)	61.46	1150
WHR	1.29 (1.02–1.61)	0.03	0.73 (0.069–0.78)	65.32	1146
WHtR	1.08 (0.89–1.31)	0.41	0.73 (0.68–0.77)	61.15	1151

Model 1 was adjusted for age. Model 2 was adjusted for age, fasting plasma glucose, and positive family history of premature cardiovascular. Model 3 was adjusted for mentioned variables plus hypertension and hypercholesterolemia. Hazard ratio indicates the increased risk for 1 standard deviation increase in each anthropometric variable. A higher value of likelihood ratio chi-squared indicates a better fit. AIC is a statistical estimate of the trade-off between the likelihood of a model against its complexity; a lower value of AIC indicates a better model fit. The discrimination ability of the models was calculated using the C index. A value of 1 denotes perfect discrimination and a value of 0.5 is no better than chance.

AIC, Akaike information criterion; BMI, body mass index; WC, waist circumference; HC, hip circumference; WHR, waist-to-hip ratio; WHtR, waist-to-height ratio.

In model 2, WC, WHR and WHtR had almost similar discrimination, but on the basis of fitting characteristics, WHR was the better-fitted model with the highest likelihood ratio (45.69) and smallest AIC (1,162). In the third model (adjusted for confounders and mediators), only WHR remained significant as a predictor of incident CVD, and a 1 SD increase in WHR resulted in 29% increase in risk of CVD.

In diabetic men, WC, WHR, and WHtR were associated with CVD risk in an age-adjusted model (model1), with similar discriminatory power; also, according to likelihood ratio and AIC, all three models had almost similar fitness (Table 3). After further adjustment (FPG and aspirin use), only WHR showed significant risk in prediction of CVD [HR 1.21 (1.00–1.48), P=0.05)]; however, a 1 SD increase in WHtR caused 19% increase in risk of CVD, which was marginally significant (P=0.08). There were significant interactions (P<0.001) between HTN and anthropometric measures in diabetic men, hence HTN-stratified analyses were done in this group (Fig. 1). In HTN-stratified analyses, all of the anthropometric variables were associated with incident CVD only in diabetic men without HTN (Fig. 1).

FIG. 1.

Hazard ratio (95% confidence interval) of anthropometric measures for incident cardiovascular disease in diabetic men stratified by hypertension status and adjusted with age, fasting plasma glucose, aspirin use, and hypercholesterolemia. BMI, body mass index; WC, waist circumference; WHR, waist-to-hip ratio; WHtR, waist-to-height ratio; HTN, hypertension.

Table 3.

Hazard Ratios and Discrimination of Anthropometric Measures for Predicting Cardiovascular Events in Diabetic Men

	Hazard ratio (95% confidence interval)	P value	C index (95% confidence interval)	Likelihood ratio chi-squared	AIC
Model 1
BMI	1.18 (0.95–1.46)	0.11	0.60 (0.54–0.66)	12.78	1035
WC	1.24 (1.00–1.55)	0.04	0.61 (0.55–0.66)	14.43	1034
WHR	1.24 (1.02–1.50)	0.02	0.61 (0.55–0.67)	15.21	1033
WHtR	1.23 (1.01–1.50)	0.03	0.61 (0.55–0.67)	14.85	1033
Model 2
BMI	1.15 (0.93–1.43)	0.19	0.64 (0.58–0.69)	21.73	982
WC	1.19 (0.95–1.48)	0.11	0.64 (0.58–0.70)	22.49	981
WHR	1.21 (1.00–1.48)	0.05	0.64 (0.58–0.70)	23.65	980
WHtR	1.19 (0.97–1.45)	0.08	0.64 (0.58–0.70)	23.00	981

Model 1 was adjusted for age. Model 2 was adjusted for age, fasting plasma glucose, and aspirin. Hazard ratio indicates the increased risk for a 1 standard deviation increase in each anthropometric variable. A higher value of likelihood ratio chi-squared indicates a better fitness. AIC is a statistical estimate of the trade-off between the likelihood of a model against its complexity; a lower value of AIC indicates a better model fit. The discrimination ability of the models was calculated using the C index. A value of 1 denotes perfect discrimination and a value of 0.5 is no better than chance.

AIC, Akaike information criterion; BMI, body mass index; WC, waist circumference; HC, hip circumference; WHR, waist-to-hip ratio; WHtR, waist-to-height ratio.

We further adjusted LDL-C and lipid-lowering drug instead of hypercholesterolemia in model 3. In women, the HRs of anthropometric measures for incident CVD were as follows: BMI [0.91(0.72–1.15), P=0.4], WC [ 1.05(0.83–1.32), P=0.7], WHR [1.26 (1.00–1.61), P=0.06], and WHtR [1.03 (0.83–1.26), P=0.7]. Corresponding risks in men without HTN were as follows: BMI [1.7(1.16–2.49), P=0.003], WC [1.48 (1.00–2.19), P=0.05], WHR [1.32 (0.93–1.88), P=0.1], and WHtR [1.56 (1.07–2.27), P=0.02]. Corresponding risks in men with HTH were as follows: BMI [0.68 (0.47–1.01), P=0.06], WC [0.79 (0.55–1.14), P=0.2], WHR [0.99 (0.7–1.38), P=0.9], and WHtR [0.80 (0.58–1.10), P=0.2].

Discussion

In this population-based prospective cohort of an Iranian diabetic population, WHR, in the presence of confounding variables, was a strong predictor of incident CVD among diabetic men and women and also predicted CVD in diabetic women, independent of HTN and hypercholesterolemia. WC in diabetic women and WHtR in both sexes, independent of other confounders predicted CVD, marginally. In this study, BMI was not associated with CVD in men and women.

Although, in several studies among overall populations, it has been shown that WHR is a much better independent predictor for incident CVD than the other anthropometric measures,^16,19,20 there are few studies about this relation in diabetic populations.^21
–23 The United Kingdom prospective diabetes study (UKPDS) showed that high levels of WHR and BMI were not major risk factors for coronary artery disease (CAD) in white patients with non-insulin-dependent diabetes mellitus.²¹ Similar to the UKPDS results, Tseng in a cross- sectional study conducted on Chinese type 2 diabetic patients in Taiwan, mentioned that WHR was not associated with CAD.²² On the other hand, Yusuf et al. studied 27,000 participants from 52 countries in 2005 and demonstrated that WHR is the strongest anthropometric measure for predicting the risk of MI in men and women, in all ages and ethnic groups, and in subjects with and without the mediators of obesity such as diabetes, HTN, and hyperlipidemia.¹⁶

In the current study, independent of the confounders, WHR was a strong predictor for CVD, and a 1 SD increase in WHR caused a 21% and 32% increase in CVD risk in diabetic men and women, respectively. Also, in line with our study findings, Le et al., in a general population, showed that WHR is associated with CVD, and this association is stronger in women than men; they also reported that the risk of CVD doubles in women with higher WHR than those with lower WHR. However, the increased risk for the men with high WHR was 50% compared with men without this condition.²⁰ Some studies mentioned that WHR is a superior predictor of CVD than WC because it includes HC, which has been shown in many studies in general populations to have an inverse association with CVD.²⁴ This protective effect of HC might be related to its correlation with other anthropometric features, higher HC being associated with increased gluteal muscle and total leg muscle. Higher leg muscle is related to more physical activity, which is inversely associated to cardiometabolic risk.²⁴ The strong association of WHR with CVD in the current study, in a diabetic population, and in an earlier study on the general population of TLGS¹⁹ might be attributed to the high prevalence of a sedentary lifestyle in this population, especially in women.

Waist circumference is one of the measures of abdominal obesity that is used widely in predicting the risk of CVD.¹⁴ Sone et al., in a prospective study of Japanese patients with type 2 diabetes,¹⁴ showed that WC per se did not increase the risk of CVD in both genders and showed there was a discrepancy between cross-sectional and longitudinal results. In the current study, WC was not an independent predictor of CVD in diabetic men, but it was associated with CVD marginally in a confounder-adjusted model in diabetic women, although, after adjustment for mediators, WC did not remain as an independent predictor of CVD. The World Health Organization has advised that adjustment with obesity mediators such as HTN and hyperlipidemia could underestimate the true estimated risk of obesity measures for cardiovascular events.²⁵ Similar to our results, in many other studies, WHR was stronger than WC for predicting CVD,²⁴ but it should be considered that patients might not allow physicians to measure HC. Thus, the choice to use WC or WHR depends on the setting where the measures are taken.²⁴

WHtR is a measure of abdominal obesity manifesting an enlarged abdomen with a short stature.²⁶ Wu et al. indicated that WHtR had a higher correlation with cardiovascular risk factors than WC, WHR, or BMI in newly diagnosed type 2 diabetes.²⁷ To our best knowledge, only a few studies have evaluated the association between WHtR with other anthropometric measures for prediction of CVD in a diabetic population.²³ In the current study, WHtR independent of confounders was associated with incident CVD marginally in both genders, and a 1 SD increase in WHtR caused a 19% and 18% increase in risk of CVD in men and women, respectively. Khalangot et al. in a study of 30,534 Ukrainian men and 58,909 women with type 2 diabetes, after 7.2 years follow-up, indicated that BMI had a U-shaped association with CVD mortality and, in contrast to extreme obesity, being overweight did not increase CVD mortality compared to normal weight.⁶ In line with the UKPDS study,²¹ our findings showed no association between BMI and CVD in a diabetic population. It seems that further studies are needed to clarify the exact relationship between BMI and CVD among diabetic populations.

There was a significant interaction between anthropometric measures and HTN in diabetic men for incident CVD in this study. We showed that all obesity measures are powerful predictors for CVD in diabetic men without HTN, in contrast to those with HTN, findings similar to those of other studies in overall populations.^19,28 In contrast to our findings, Sone et al. reported WC was not a predictor for cardiovascular events both in hypertensive and nonhypertensive diabetic men.¹⁴ Further research is needed to evaluate whether the degree of hypertension control impacts this interaction.

There are potential limitations regarding the interpretation of our results. The major limitation of this study was the modest numbers of events that threaten the statistical power of more detailed analyses. We could not evaluate the association of anthropometric measures with a specific CVD event. The most prevalent CVD event in this study was angiographic-proven CHD, but none of the anthropometric measures showed a significant association with angiographic-proven CHD in men and women. This finding was unstable because of very limited statistical power.

Furthermore, we did not have data regarding glycosylated hemoglobin (HbA1c), albuminuria, and duration of diabetes, the important predictors of CVD in diabetic patients.²⁹ Among these predictors, HbA1c has an important role. An observational study from the Swedish National Diabetes Register showed progressively increasing risks of CVD with higher HbA1c, and no risk increase at low HbA1c levels even with longer diabetes duration or previous CVD.³⁰ However, based on several studies, in patients with poor glycemic control and those with HbA1c values over 8%, FPG is an indicator of glycemic control, which we adjusted in the multivariable models.^31
–33 Accordingly, FPG is applied in place of HbA1c as a surrogate of diabetes control in prediction of CVD outcomes among diabetic patients in other studies.³⁴

Third, we had no data regarding type of diabetes, although, according to the low prevalence of type 1 diabetes in the overall TLGS population³⁵ and by including patients aged ≥30 years, we identified subjects with higher probability of type 2 than type 1 diabetes.

Fourth, some misclassification of diabetes might have occurred due to defining newly diagnosed diabetes with a single OGTT at baseline and a partial reliance on self-reported history of diabetes. Misclassification, if any, could have led to an underestimation of any association between anthropometric measures with incident CVD in diabetic patients. However, the extent of misclassification would be extensive to change the results of this study.³⁶

Finally, low-density lipoprotein cholesterol (LDL-C) was not measured directly in the Tehran Lipid and Glucose Study. If we used the Friedewald formula for calculation of LDL-C, we would miss the subjects having triglycerides more than 400 mg/dL (10.7% of the patients in this study). But, we repeated the multivariate analyses with LDL-C and lipid-lowering drugs as covariates and the results were the same.

To the best of our knowledge, this study was the first prospective and population-based study with continuous monitoring for CVD events among type 2 diabetic men and women, which examined BMI, WC, WHR, and WHtR simultaneously for prediction of CVD. In conclusion, this study showed WHR was the most powerful predictor of CVD among anthropometric measures, followed by WHtR, in a diabetic population.

Footnotes

Acknowledgments

This study was supported by Grant No. 121 from the National Research Council of Iran. We express appreciation to the participants of district 13, Tehran, for their enthusiastic support in this study. We would like to thank Dr. D. Khalili for his participation in study design and Ms. N. Shiva for her assistance in the English editing of the manuscript.

Author Disclosure Statement

The authors declare that they have no conflicts of interest.

References

Eeg-Olofsson

, Cederholm

, Nilsson

et al. Risk of cardiovascular disease and mortality in overweight and obese patients with type 2 diabetes: An observational study in 13,087 patients. Diabetologia, 2009; 52:65–73.

Hossain

, Kawar

, El Nahas

. Obesity and diabetes in the developing world—a growing challenge. N Engl J Med, 2007; 356:213–215.

Wild

, Roglic

, Green

et al. Global prevalence of diabetes: Estimates for the year 2000 and projections for 2030. Diabetes Care, 2004; 27:1047–1053.

, Cowie

, Harris

. Mortality in adults with and without diabetes in a national cohort of the U.S. population, 1971–1993. Diabetes Care, 1998; 21:1138–1145.

Wilson

, D'Agostino

, Levy

et al. Prediction of coronary heart disease using risk factor categories. Circulation, 1998; 97:1837–1847.

Khalangot

, Tronko

, Kravchenko

et al. Body mass index and the risk of total and cardiovascular mortality among patients with type 2 diabetes: A large prospective study in Ukraine. Heart, 2009; 95:454–460.

Gaede

, Vedel

, Larsen

et al. Multifactorial intervention and cardiovascular disease in patients with type 2 diabetes. N Engl J Med, 2003; 348:383–393.

Haffner

, Lehto

, Ronnemaa

et al. Mortality from coronary heart disease in subjects with type 2 diabetes and in nondiabetic subjects with and without prior myocardial infarction. N Engl J Med, 1998; 339:229–234.

Donnelly

, Doney

, Hattersley

et al. The effect of obesity on glycaemic response to metformin or sulphonylureas in type 2 diabetes. Diabet Med, 2006; 23:128–133.

10.

Adams

, Schatzkin

, Harris

et al. Overweight, obesity, and mortality in a large prospective cohort of persons 50 to 71 years old. N Engl J Med, 2006; 355:763–778.

11.

Bogers

, Bemelmans

, Hoogenveen

et al. Association of overweight with increased risk of coronary heart disease partly independent of blood pressure and cholesterol levels: a meta-analysis of 21 cohort studies including more than 300 000 persons. Arch Intern Med, 2007; 167:1720–1728.

12.

Chaturvedi

, Fuller

. Mortality risk by body weight and weight change in people with NIDDM. The WHO Multinational Study of Vascular Disease in Diabetes. Diabetes Care, 1995; 18:766–774.

13.

Cho

, Manson

, Stampfer

et al. A prospective study of obesity and risk of coronary heart disease among diabetic women. Diabetes Care, 2002; 25:1142–1148.

14.

Sone

, Tanaka

, Iimuro

et al. Waist circumference as a cardiovascular and metabolic risk in Japanese patients with type 2 diabetes. Obesity (Silver Spring), 2009; 17:585–592.

15.

Pristipino

, Beltrame

, Finocchiaro

et al. Major racial differences in coronary constrictor response between japanese and caucasians with recent myocardial infarction. Circulation, 2000; 101:1102–1108.

16.

Yusuf

, Hawken

, Ounpuu

et al. Obesity and the risk of myocardial infarction in 27,000 participants from 52 countries: A case-control study. Lancet, 2005; 366:1640–1649.

17.

Yusuf

, Reddy

, Ounpuu

, Anand

. Global burden of cardiovascular diseases: Part I: General considerations, the epidemiologic transition, risk factors, and impact of urbanization. Circulation, 2001; 104:2746–2753.

18.

Hadaegh

, Harati

, Ghanbarian

et al. Association of total cholesterol versus other serum lipid parameters with the short-term prediction of cardiovascular outcomes: Tehran Lipid and Glucose Study. Eur J Cardiovasc Prev Rehabil, 2006; 13:571–577.

19.

Hadaegh

, Zabetian

, Sarbakhsh

et al. Appropriate cutoff values of anthropometric variables to predict cardiovascular outcomes: 7.6 years follow-up in an Iranian population. Int J Obes (Lond), 2009; 33:1437–1445.

20.

, Engstrom

, Hedblad

et al. Sex differences in the relationships between BMI, WHR and incidence of cardiovascular disease: A population-based cohort study. Int J Obes (Lond), 2006; 30:1775–1781.

21.

Turner

, Millns

, Neil

et al.

Risk factors for coronary artery disease in non-insulin dependent diabetes mellitus: United Kingdom Prospective Diabetes Study (UKPDS: 23)

BMJ, 1998; 316:823–828.

22.

Tseng

. Body composition as a risk factor for coronary artery disease in Chinese type 2 diabetic patients in Taiwan. Circ J, 2003; 67:479–484.

23.

Tseng

. Waist-to-height ratio and coronary artery disease in Taiwanese type 2 diabetic patients. Obesity (Silver Spring), 2008; 16:2754–2759.

24.

de Koning

, Merchant

, Pogue

et al. Waist circumference and waist-to-hip ratio as predictors of cardiovascular events: Meta-regression analysis of prospective studies. Eur Heart J, 2007; 28:850–856.

25.

World.Health.Organization. Obesity: Preventing and managing the global epidemic. Report of a WHO consultation. World Health Organ Tech Rep Ser, 2000; 894:1–253.

26.

Park

, Choi

, Lee

et al. Waist circumference and waist-to-height ratio as predictors of cardiovascular disease risk in Korean adults. Circ J, 2009; 73:1643–1650.

27.

, Chen

, Zheng

et al. Simple anthropometric indices in relation to cardiovascular risk factors in Chinese type 2 diabetic patients. Chin J Physiol, 2007; 50:135–142.

28.

Silventoinen

, Magnusson

, Neovius

et al. Does obesity modify the effect of blood pressure on the risk of cardiovascular disease? A population-based cohort study of more than one million Swedish men. Circulation, 2008; 118:1637–1642.

29.

American Diabetes Association. Standards of medical care in diabetes—2009. Diabetes Care, 2009; 32,Suppl 1:S13–S61.

30.

Eeg-Olofsson

, Cederholm

, Nilsson

et al.

New aspects of HbA1c as a risk factor for cardiovascular diseases in type 2 diabetes: An observational study from the Swedish National Diabetes Register (NDR)

J Intern Med, 2010; 268:471–482.

31.

Monnier

, Lapinski

, Colette

. Contributions of fasting and postprandial plasma glucose increments to the overall diurnal hyperglycemia of type 2 diabetic patients: Variations with increasing levels of HbA(1c) Diabetes Care, 2003; 26:881–885.

32.

Yamamoto-Honda

, Kitazato

, Hashimoto

et al. Distribution of blood glucose and the correlation between blood glucose and hemoglobin A1c levels in diabetic outpatients. Endocr J, 2008; 55:913–923.

33.

Kikuchi

, Nezu

, Shirakawa

et al. Correlations of fasting and postprandial blood glucose increments to the overall diurnal hyperglycemic status in type 2 diabetic patients: Variations with evels of HbA1c. Endocr J, 2010; 19.

34.

Abu-Lebdeh

, Hodge

, Nguyen

. Predictors of macrovascular disease in patients with type 2 diabetes mellitus. Mayo Clin Proc, 2001; 76:707–712.

35.

Sanjari

, Hedayati

, Azizi

. Prevalence of type 2 diabetes mellitus in 3–19 age group in East of Tehran in 2001. Iranian J Endocrinol Metab, 2007; 9:119–126.

36.

Asia Pacific Cohort Studies Collaboration. Cholesterol, diabetes and major cardiovascular diseases in the Asia-Pacific region. Diabetologia, 2007; 50:2289–2297.