It has long been suggested that if an Asian and a Caucasian have the same body mass index (BMI), then the former has higher fat mass than the latter. I have been troubled by this presumption for many years, because it does not seem consistent with my obervations. In this study (1), using Korean and Australian data, my colleagues and I challenge that presumption, by pointing out that Koreans actually have lower relative body fat than Caucasians, and this finding has huge implication for the diagnosis of obesity in Asians.
Obesity is currently defined in terms of body mass index. WHO defines obesity as “abnormal or excessive fat accumulation that presents a risk to health.” The keyword in this definition is “fat accumulation”. How can we measure fat mass of an individual? In the old days, scientists used the century-old measure called ‘body mass index’ (BMI). The index is also known as ‘Quetelet index’ because it was developed by the Belgian polymath Adolphe Quetelet in the 1800s. BMI is calculated by dividing the body weight (in kilograms) by the height (in metre squared). It turns out that BMI is correlated with mortality risk, and based on this fact, WHO recommended to classify an individual’s BMI into one of the following 4 groups:
However, BMI is only an approximate indicator of fatness; it does not measure fat mass. Body weight (a key component of BMI) is mainly made of lean mass and fat mass. So, two individuals may have the same BMI, but their fat mass can be quite different (Figure 1). Consequently, the diagnosis of obesity based on BMI can be quite misleading. For instance, Mr Arnold Schwarzenegger has a BMI of 30.8 kg/m2, which by WHO definition is obese. However, none of us would consider him obese, because his weight is mainly due to lean mass, not fat mass. People in the field have long known that BMI is not a good indicator of obesity.
Figure 1: Two men with the same BMI, but very different PBF
(Yajnik CS, Yudkin JS. The Y-Y Paradox. Lancet 10/1/2004)
Percent body fat
In recent years, dual X-ray absorptiometry (DXA) has emerged as a gold–standard for the assessment of body composition (eg fat mass and lean mass). DXA uses an X-ray tube and a filter to generate two X-ray beams with various energy levels to determine body composition. The measurement is very accurate and reliable, with the coefficient of variation being 1-2%. Therefore, it is not surprising that DXA-based PBF is considered the gold-standard method of obesity diagnosis.
However, at present, we don’t have a cut-off value of PBF for diagnosing obesity. Some investigators thought that a PBF greater than 25% (for men) or greater than 30% (for women) should be used to define obesity, and they attributed these cut-off values to a WHO Technical Report (3). But that is a misquotation, because the WHO Technical Report makes no such recommendation at all (4). It is interesting to note that until now, people still continue to misquote those cut-off values (4)!
Do Asians have higher percent body fat than Caucasian?
In the mean time, some investigators thought that for a given BMI, Asians have higher PBF than Caucasians. This presumption was based on an initial study by Wang et al (5), showing that American men and women of Asian backgrounds (eg Chinese, Japanese, Koreans, Filipinos) had higher PBF than Caucasians. The conclusion was based on a statistical analysis, not original data. In original/unadjusted analysis, the difference in PBF between Asians and Caucasians was 1.5 percentage points (95% CI, -0.2 to 3.2; P = 0.081) for women, and 2.1 percentage points (95% CI, 0.6 to 3.6; P = 0.006) for men. However, using least square means from an analysis of covariance (ANCOVA), the difference in PBF between Asians and Caucasians increased to 2.6-4.6 percentage points for women, and 3.5-5.3 percent points (ppt) for men! In other words, the ANCOVA has turned a 1.5 ppt difference into a 2.6-4.6 ppt! I cannot understand how such a difference occurred.
Remarkably, based on this flimsy evidence, many investigators and professional societies proposed that the diagnostic criteria for BMI in Asians should be lowered (ie <30 kg/m2). An Indian consensus group and the Japan Society for the Study of Obesity classify obesity as having BMI ≥ 25 kg/m2 (6-7). Still, the China Obesity Task Force recommends that BMI ≥ 28 kg/m2 be the criterion for obesity in Chinese individuals (8).
The lowering BMI criteria has major impact on the prevalence of obesity. For example, using the criterion of BMI ≥ 25 kg/m2, the prevalence of obesity in Koreans was 38% in men and 43% in women. However, using the criterion of BMI ≥ 30 kg/m2, the prevalence of obesity in Koreans was only 2% (in men) and 6% (in women).
I have always been troubled with the assumption that Asians have higher PBF than Caucasians. I have tried to verify that assumption in a previous study that compared PBF between Vietnamese women and American women (9). Based on DXA data asertained from 210 Vietnamese women and 419 American women, we found no statistically significant difference in PBF between the two groups (Figure 2). Actually, we found that Vietnamese women actually had lower PBF than American women! So, that study was the first ‘falsification’ of the common assumption that has existed for more than a decade.
Figure 2: Percent body fat and BMI in US white (open circles) and Vietnamese women (closed circles). Data are from the age and BMI matched sample. VN, Vietnam (source: Ho-Pham et al, Obesity 2010).
Then, in 2016 I was invited to give a lecture in the Seoul International Congress of Endocrinology and Metabolism, and I had a chance to meet Dr. Duong Pham and his colleague, Professor Chae Hun Leem. We had a long chat about our research interests, and specifically, my interest in testing the hypothesis that Asians have higher PBF than Caucasians. The ‘outcome’ of that informal meeting was a collaborative project. In this project, we compared PBF (measured by DXA, Lunar GE) of 1211 Koreans and 1006 Australians. This was probably the largest study to-date in this area of research.
What did we find? Well, we found that Koreans had lower PBF than Australians. More specifically, after adjusting for age and BMI, Korean women had a statistically lower PBF (-2.1 percentage points; 95% CI, -2.6 to -1.6; P < 0.0001) than Australian women (Figure 3). The same trend was also observed in men, but the difference did not reach a conventional statistical significance: -0.5 percentage point, with 95% CI being -1.2 to 0.1 (P = 0.12). So, this study again falsifies the assumption that Asians have higher fat mass than Caucasians.
Figure 3: Relationship between percent body fat (y‐axis) and BMI (x‐axis) in Australian (Au; grey) and Korean (Ko; black) for all data for (A) men and (B) women (Source: Pham et al Obesity 2019).
Back to the question, “Do Asians have higher percent body fat than Caucasians”, I have to say that the answer is NO. Actually, careful analyses of high quality data suggest that for a given BMI, East Asians actually have lower PBF than Caucasians. I don’t think there is a need to lower the current BMI cut-off (30 kgm2) for defining obesity in East Asian populations. In the absence of sound evidence, lowering the BMI cutoff will result in a large increase in the obesity prevalence in Asian populations, and that is, to me, irresponsible.
(3) Deurenberg P, Yap M, van Staveren WA. Body mass index and percent body fat: a meta analysis among different ethnic groups. Int J Obes Relat Metab Disord. 1998;22:1164-1171.
(6) Misra A, Chowbey P, Makkar BM, et al. Consensus statement for diagnosis of obesity, abdominal obesity and the metabolic syndrome for Asian Indians and recommendations for physical activity, medical and surgical management. J Assoc Physicians India 2009;57:163‐170.
(9) Ho‐Pham LT, Lai TQ, Nguyen ND, Barrett‐Connor E, Nguyen TV. Similarity in percent body fat between white and Vietnamese women: implication for a universal definition of obesity. Obesity 2010;18:1242‐1246.