Further Analysis of INTERHEART: What Do the Data Tell Clinicians?

Arya Sharma, MD


January 19, 2006

In This Article

Abdominal Obesity in INTERHEART

On the simplest level, most clinicians would no doubt say that obesity is a risk factor for CVD, although many may say that they are not really worried about "a little extra weight." Indeed, as recently as only a few years ago, the American Heart Association thought there was a need to release a statement about obesity being a risk factor for future CVD events, even though it is clear that insurance companies had known this for years, as indicated by the fact that they measured blood pressure and weight on a continual basis.

The problem no doubt arises partly because the definition of obesity is not straightforward, and of late it has been the topic of some debate. The usual measure of obesity has been the body mass index (BMI), which represents the ratio of body weight divided by height, mathematically squared (kg/m2). As a calculated parameter it is intuitively attractive because it adjusts for the apparent truth that a taller person with the same weight as a shorter person should have a different relation of weight to future health status. Whether the BMI ratio is valid across different ethnic populations, however, was not clear. Therefore, we aimed to assess whether other markers of obesity, especially waist-to-hip ratio, would be stronger indicators of MI than BMI.

To do this, we utilized the INTERHEART enrollment database and retrospectively assessed the relation between BMI, waist and hip circumferences, and waist-to-hip ratio vs risk of MI in the overall group and for each subgroup. The results were published in The Lancet,[3] and I will discuss our findings and their implications briefly below.

As expected, we found that BMI showed a modest and graded association with risk of MI (OR 1.44, 95% CI 1.32-1.57 top quintile vs bottom quintile before adjustment), which was substantially reduced after adjustment for waist-to-hip ratio (OR 1.12, 95% CI 1.03-1.22) and nonsignificant after adjustment for other risk factors (OR 0.98, 95% CI 0.88-1.09).

By contrast, after adjustment for BMI, waist (adjusted OR 1.77; 95% CI 1.59-1.97) and hip (OR 0.73; 95% CI 0.66-0.80) circumferences were both highly significantly associated with risk of MI (P < .0001), even after adjustment for other risk factors. Likewise, for an increased waist-to-hip ratio, the population-attributable risks for future MI in the top 2 quintiles was 24.3% (95% CI 22.5-26.2) compared with only 7.7% (95% CI 6.0-10.0) for the top 2 quintiles of BMI.

The relatively weak association between BMI and MI risk was seen in all ethnic groups, and no significant relationship was observed in South Asians, Arabs, or mixed-race Africans.[3] In addition, BMI was not as powerful a predictor in those with a history of hypertension or a raised ApoB/ApoA-1 ratio. By contrast, an increased waist-to-hip ratio correlated with a substantial 3-fold increase in the population-attributable risk of MI compared with BMI.[3]

Looking at the statistics in the reverse way, waist-to-hip ratio was the strongest anthropomorphic predictor of MI in men and women, across all age and ethnic groups, in smokers and in nonsmokers, and in those with or without dyslipidemia, diabetes, or hypertension.[3] Increasing waist-to-hip ratio was a predictor of MI even in those regarded as very lean (BMI < 20 kg/m2) and in those regarded as being of ideal weight (> 20 to < 25 kg/m2), overweight (> 25 kg/m2), and obese (> 30 kg/m2). Moreover, the waist-to-hip ratio was a graded measure, with the odds ratios increasing for each successive quintile ( OR =1.15 for 2nd quintile, 1.39 for 3rd quintile, 1.9 for 4th quintile, and 2.52 for 5th quintile; P < .0001).

Waist-to-hip ratio was a better measure of risk than waist circumference alone. This finding could partly relate to adjustment of measures of abdominal circumference for pelvic girth (by measurement of the hip), but might also be due to a protective effect associated with larger hip circumference, since there was a significant inverse relation with risk of MI (after adjustment for BMI). Loss of fat in the hips and limbs during weight reduction has been correlated with increases in blood pressure and worsening of metabolic risk factors.[4,5]

The biologic mechanisms underlying the relation between obesity and cardiovascular risk are unclear, but may relate to the definition of obesity. For example, the waist-to-hip ratio is a simple measure for visceral obesity, and visceral adiposity is hypothesized to be a key manifestation of the metabolic abnormalities that underlie risk of future CVD events. It may also be that the ratio of fat to muscle (sarcopenic adiposity) can be a measure of risk of CVD, which is best estimated by waist-to-hip ratio.

Visceral adiposity is the fat that is integrated within the abdominal organs, as opposed to subcutaneous fat, as measured by the "skin-fold" test. Other studies have found that visceral fat, not body fat, is the principal determinant of metabolic risk. Thorne and colleagues[6] studied 50 patients with baseline BMI > 35 kg/m2 and found that after a 2-year follow-up, there was no difference in weight loss or BMI in patients who underwent gastric banding alone vs those who underwent gastric banding with the removal of omentum (visceral fat), but in those patients with removal of visceral fat, the result was a 2- to 3-fold improvement in glucose tolerance and insulin sensitivity. These patients lost minimal amounts of overall weight (no subcutaneous or body fat was involved), but nevertheless experienced considerable metabolic improvement.

Several factors have been proposed as explanations for the opposing effects of abdominal and lower body fat on cardiovascular risk. Hormonal factors may have different effects on waist, thigh, and hip circumference and insulin resistance. Glucocorticoid excess, growth hormone deficiency, and high androgen concentrations in women and low testosterone in men are associated with increased visceral fat, reduced skeletal muscle mass, and insulin resistance.[6,7] Elevated angiotensin plasma levels, elevated renin activity, elevated aldosterone, and elevated angiotensin-converting enzyme (ACE) were seen in a study of obese women and may illuminate the mechanism by which blood pressure and metabolic dysregulation may occur.[5] By contrast, endogenous estrogens may stimulate accumulation of subcutaneous gluteal and femoral fat, which are considered cardioprotective.[7] Increased hip circumference might also indicate increased gluteal muscle and could be a marker of overall skeletal muscle mass.

Put more simply, these results suggest that it is not how much fat a patient has, but where it is stored that contributes to overall cardiovascular risk, insulin resistance, and glucose tolerance.[5] This is good news, in that it means that simple linear measurement of waist and hip circumferences provides a predictive power greater than that provided by BMI for estimating the risk of MI that is attributable to obesity in most ethnic groups.

A final point is that, with the redefinition of abdominal obesity based on waist-to-hip ratio instead of BMI, the estimated risk of MI attributable to obesity was higher than previously assumed in most ethnic groups, and the risk of MI rose progressively with increasing values for waist-to-hip ratios with no evidence of a threshold. This suggests that the global burden of obesity has been substantially underestimated by the reliance on BMI in previous studies.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.