One index that doesn't fit all

Hong Kong defines overweight and obese people as those with a Body Mass Index (BMI) of more than 23 and 25, respectively. A recent study published in the Canadian Medical Association Journal, however, questioned the validity of the recommended Asian thresholds.

Researchers from Taiwan and the US found that Chinese men and women with a BMI of 24 to 25.9 - overweight by Hong Kong's definitions - had the lowest risk of death. For Caucasians, it was 22.5 to 24.9, which is well within their 'healthy' threshold.

Various other studies are making it clear that defining the weight and health status across different populations isn't as straightforward as taking weight in kilograms and dividing it by the square of height in metres.

The fact is, BMI is almost 200 years old and was never designed to gauge fatness.

Belgian polymath Adolphe Quetelet, who invented the formula in the 1800s, found it was a reliable statistical tool to study the physical characteristics of the normal or average man.

When American scientist Ancel Keys revived the formula's use for population studies in 1972 and coined its modern-day moniker, he thought it useful for widespread studies of diet but not for diagnosis of obesity.

BMI's simplicity and relative reliability in gauging total body fat have boosted its popularity as an assessment tool in recent decades.

While there are other methods to more accurately measure fatness, they require the use of specialised equipment operated by trained professionals. In the 1980s, the World Health Organisation (WHO) adopted BMI as the standard index of obesity.

But it has been a subject of disagreement and confusion. For instance, the National Institutes of Health in the US first defined obesity as 27.8 for men and 27.3 for women in 1985. But in 1998, the thresholds were consolidated and changed to 25 for overweight and 30 for obesity to align them with international standards, and millions of formerly 'healthy' Americans suddenly found themselves booted into the unhealthy range.

Yet a 2005 study found that BMI-defined 'overweight' people had a mortality risk similar to that of 'healthy' people. In 2000, the WHO, the International Association for the Study of Obesity and the International Obesity Task Force jointly proposed to revise the definition for non-Caucasian populations to 23 for overweight and 25 for obesity, based on observations that risk of obesity-related diseases occurred at lower BMIs in some Asian populations than in Caucasian populations. But the revision was criticised for being too radical.

Then in 2002, the WHO released a report recommending that the original standards of 25 and 30 be retained as international definitions, but with 23, 27.5, 32.5 and 37.5 added as additional points of concern for Asian populations. Each country could decide what cut-off point it wanted to use to define obesity for its population.

While it may indicate total body fat, BMI does not measure body fat percentages and fails to make important distinctions between muscle and fat. Hence, a fit but extremely muscular athlete could be categorised as obese. Yet for the elderly, who have lost muscle and bone mass, their body fat percentage - and associated health risks - could have crept up without any significant weight gain.

BMI also does not tell you anything about the distribution of fat. Central fat, or fat concentrated around the abdomen, is a proven indicator of increased risk of diabetes, cardiovascular disease and even death.

Experts say waist circumference, which measures central fat, should be used with BMI to better assess one's obesity-related health risks. An Asian man's waist should be less than 90cm around, and an Asian woman's less than 80cm.

However, this also has its critics. Dr Daniel Ho Sai-yin, assistant professor at University of Hong Kong's School of Public Health, disagrees that a single waist circumference measurement should apply to people of different heights.

Studies have found that waist-to-stature ratio, which divides a person's waist circumference by height, better predicts one's risk of cardiovascular and metabolic diseases. It's even simpler to measure and calculate than BMI, and its recommended threshold of 0.5 (waist not more than half height) is easier to remember than the many - and confusing - figures associated with BMI.

Even the WHO does not claim BMI to be a definitive measure of obesity in individual assessments. It recommends doctors consider other measurements such as waist circumference, the patient's medical history and presence of other risk factors.

But, regardless of the methods and numbers bandied about to define obesity, the harder task is figuring out how to help individuals bring down the number on the weighing scale.