Singapore classifies overweight at a BMI of 23 and obesity at 27.5, where the international figures are 25 and 30. It follows from body-composition research conducted substantially in Singapore, which found that the same BMI corresponds to a different amount of body fat, and to different risk, in this population. This article covers that evidence and what it does and does not imply.

BMI: weight divided by height squared. A ratio designed as a population screening tool, which assumes a consistent relationship between that ratio and body fat. That assumption is where the problem sits.

The measurement that started it

BMI does not measure fat. It measures the relationship between mass and height, and it is used as a proxy for fat on the assumption that the relationship holds. In Singapore, researchers tested that assumption directly.

Using a four-compartment reference model — the most rigorous body-composition method available — in 291 Singaporean Chinese, Malay and Indian adults, a prediction equation derived from Caucasian populations under-predicted body-fat percentage by 2.7 to 5.6 percentage points (Deurenberg-Yap 2000).

Put the other way round: to reach the body-fat percentage that a Caucasian carries at a BMI of 30, the equivalent BMI was about 27 kg/m² for Chinese and Malay participants and about 26 for Indian participants.

The differences ran by ethnicity within Singapore too. Indian participants showed the highest body-fat percentage at equivalent BMI, Chinese participants the lowest. So this is not one adjustment applied to "Asians" as a bloc — it is a set of different relationships, which is part of why the guideline settles on a single practical threshold rather than trying to encode all of them.

A more recent DXA-based study of Singaporean adults reached the same conclusion by a different route: body-fat percentage was higher than the same BMI, age and sex would predict in Caucasian reference populations (Chen 2021).

The risk that follows

More fat at a given BMI would be an academic finding if it did not track with anything. It does.

In a Singapore population study of 4,723 adults, at a BMI of 22–24 and waist circumferences of 75–80 cm in women and 80–85 cm in men, cardiovascular risk-factor prevalence ran between 41% and 81%, with odds ratios of 1.97 to 4.38 against reference (Deurenberg-Yap 2002). At equivalent body fat, Singaporean BMI ran about 3 kg/m² lower than Caucasian.

A BMI of 22 to 24 sits comfortably inside the international "normal" band. In this population, a substantial share of people in that band already had cardiovascular risk factors.

That is the finding the thresholds exist to catch.

What WHO actually did

This gets misreported often enough to be worth stating precisely.

A WHO expert consultation reviewed the Asian data and found that health risk rises at BMI levels below the standard international overweight threshold — an observed-risk range across Asian populations of 22–25 kg/m², and a high-risk range of 26–31 (WHO 2004).

What it did not do was lower the international cut-offs. It explicitly retained the existing global classification and proposed additional public-health action points alongside it, at 23.0, 27.5, 32.5 and 37.5 kg/m².

So "WHO changed BMI for Asians" is wrong. WHO kept the international classification and added a second set of trigger points for countries to use. Singapore adopted the relevant ones: its clinical practice guidelines set overweight at 23.0 and obesity at 27.5 (HPB-MOH 2016), alongside waist-circumference action points of above 90 cm for men and 80 cm for women.

What it does not tell you

Three things, and the third is the one people most often get wrong.

It is still BMI. A lower threshold does not fix the underlying limitation — BMI still cannot distinguish muscle from fat, and it still says nothing about where fat is carried. Lowering the number makes the screen more sensitive in this population; it does not make it a measure of composition. We cover the alternatives in body fat vs BMI vs weight.

There is no matching body-fat threshold. Singapore publishes BMI and waist-circumference figures. Neither HPB nor MOH publishes a body-fat-percentage cut-off, and a 2021 Singapore study reports the same absence of an Asian consensus cut-off (Chen 2021). If you are shown a body-fat percentage presented as a Singapore standard, it is not one.

A threshold is not an eligibility rule. Whether any particular treatment is appropriate for a particular person is a clinical assessment, and it takes account of considerably more than one ratio. A number crossing a line is a reason for a conversation, not the outcome of one.

At GetLean, our philosophy is that the composition is the goal and the weight is a by-product. The Asian threshold research is, in a sense, the same argument made at national scale: the number on its own was hiding something, and looking at what bodies are actually made of changed the answer.

Individual circumstances vary, and what any of this means for you is a question for a doctor who has assessed you.

Common questions

Why does Singapore use a BMI of 27.5 instead of 30?

Because risk appears at lower body sizes in this population. A WHO expert consultation found risk rising below the standard threshold in Asian populations and proposed additional action points including 23.0 and 27.5 kg/m² (WHO 2004), which Singapore's guidelines adopt (HPB-MOH 2016).

Do Asians have more body fat at the same BMI?

In Singaporeans, yes. A Caucasian-derived equation under-predicted body-fat percentage by 2.7 to 5.6 percentage points in a four-compartment study of 291 adults (Deurenberg-Yap 2000), and a later DXA study found the same direction (Chen 2021).

Is the difference the same for all ethnic groups in Singapore?

No. Indian participants showed the highest body-fat percentage at equivalent BMI and Chinese participants the lowest, with matching thresholds of about 26 and 27 kg/m² respectively (Deurenberg-Yap 2000).

Did WHO lower the international BMI cut-offs?

No. It explicitly retained the existing international classification and added supplementary public-health action points alongside it (WHO 2004).

Does this mean a BMI of 24 is a problem in Singapore?

It means it is worth looking at rather than dismissing — cardiovascular risk-factor prevalence ran 41–81% at BMI 22–24 in a Singapore population study (Deurenberg-Yap 2002). What it means for an individual is a clinical question.