Skip to content
This GLP
← Research
Evidence

What BMI Do You Need for a GLP-1? The Cut-Off Is a Convention

A trial in China and Taiwan enrolled people at a BMI of 24 to 30, below where the Western trials start. The cut-off is a guideline, not a boundary in a body.

Glenn Torres9 min read
Body mass index at trial entrySTEP 12 entry (China, Taiwan)typical Western trial entry242730

The number depends on which guideline your prescriber uses. A phase 3b trial in China and Taiwan enrolled adults at a BMI of 24 to under 28 [1]. That band required a weight-related condition, and a second ran 28 to under 30. Weight fell 12.1% against 2.2% on placebo over 44 weeks. East Asian guidelines set the cut-points lower because risk rises at lower BMI there. A cut-off is a convention adopted by a committee. It is not a line where a body changes. Nothing on this page tells you that you qualify.

Every intake form asks for a height and a weight. Behind it sits a number that decides whether anything is offered. That number is not a fact about bodies. It is a guideline, and guidelines differ. It is a different kind of gate from the state lists sellers publish, and it works the same way.

A trial run at lower numbers

STEP 12 enrolled adults across 19 sites in mainland China and Taiwan [1]. It used locally defined BMI thresholds. Entry was 24 to under 28 with at least one weight-related condition. The second band was 28 to under 30, with or without type 2 diabetes. Of 242 participants, 161 received semaglutide 2.4 mg weekly and 81 placebo. The trial ran 44 weeks alongside lifestyle intervention. Half were women and 19.4% had type 2 diabetes.

Those entry criteria sit below where the well-known trials began. The rationale is not novel. East Asian guidelines set overweight and obesity cut-points lower, because cardiometabolic risk rises at lower BMI in these populations. The trial was designed around the local definition rather than the American one.

What it found

Body weight fell 12.1% on semaglutide against 2.2% on placebo. The estimated treatment difference was 9.9 percentage points, 95% CI −11.8 to −8.0, p<0.0001. The proportion reaching at least 5% weight loss was 80.5% against 24.4%, an odds ratio of 14.8, 95% CI 7.4 to 29.6.

Adverse events were reported by 141 of 161 participants on the drug, 87.6%, and 61 of 81 on placebo, 75.3%. Both figures are high. The difference between them is smaller than the drug figure alone suggests. That is the reason to ask what the placebo arm reported.

What the threshold does after you clear it

Meeting the number is not the same as being coded for it. A decade-long analysis looked at 42,243 hospitalized adults with a measured BMI of 30 or above and no prior obesity code [2]. Obesity appeared in the discharge summary for 10.6% of them. Each additional BMI unit raised the odds of being coded by 17%. The threshold behaves as a gradient rather than a gate. Documented patients had 3.44 times the odds of a GLP-1 prescription. That is an association, not a cause, and the full reading is in measured in everyone, written down in one in ten.

Below the threshold entirely

One study went under it. A target trial emulation took adults with type 2 diabetes and a BMI below 27, the group the weight trials exclude [3]. Against DPP-4 inhibitors, all-cause mortality was 39% lower, hazard ratio 0.61, 95% CI 0.52 to 0.71. Against SGLT2 inhibitors it did not differ: 0.89, 95% CI 0.76 to 1.05. The authors say the mortality figure likely reflects residual confounding. Coded heart failure was lower across all three comparisons, which was their most consistent result. The reading is in the authors who doubt their own hazard ratio.

How it compares

The 12.1% reduction is smaller than the headline figures from the large Western trials. The comparison is not straightforward. This cohort started at a lower BMI, so there was less weight available to lose. The trial also ran 44 weeks rather than 68. Percentage weight loss behaves differently depending on where you start. That is one reason expected-loss estimates should never be read as a promise.

What travels is the shape. Both co-primary endpoints separated clearly from placebo. That is consistent with what the wider trial evidence shows. A drug that worked only where it was first tested would be the surprising result. This is not that.

Why a single trial in one region matters here

Most readers meet this drug through trials run in the United States and Europe. The generalization outward is usually assumed rather than demonstrated. Trials like this one test the assumption. They arrive years later and attract a fraction of the attention.

The published evidence is also narrower than the marketing implies at any moment. That is the gap separating a famous trial from the literature behind it. The trial was funded by Novo Nordisk, which makes the drug.

Frequently asked

What BMI do you need for a GLP-1?
It depends on the guideline your prescriber follows. STEP 12 enrolled at a BMI of 24 to under 28 with a weight-related condition, or 28 to under 30, under East Asian cut-points that sit below the ones used in the Western trials.
Why did this trial enroll people at a lower BMI?
It used locally defined thresholds. East Asian guidelines set overweight and obesity cut-points lower because cardiometabolic risk rises at lower BMI in these populations, and the trial was designed around that definition.
Does that mean I qualify at a lower BMI?
No. A trial run under one country's guidelines does not change the criteria a prescriber elsewhere works to, and nothing here is a reason to describe yourself differently on an intake form.
Why was the weight loss smaller than in the famous trials?
Participants started at a lower BMI, so there was less weight available to lose, and the trial ran 44 weeks rather than 68. Percentage weight loss is not comparable across different starting points.
Were side effects worse?
Adverse events were reported by 87.6% on semaglutide and 75.3% on placebo. Both are high, and the gap is narrower than the drug figure alone suggests.

Sources

  1. [1] Guo L, et al. (2026). Efficacy and safety of once-weekly semaglutide 2·4 mg in Chinese adults with overweight or obesity (STEP 12): a randomised, double-blind, placebo-controlled, multicentre, phase 3b trial The Lancet Diabetes & Endocrinology. PMID 42575111
  2. [2] Avivi I, et al. (2026). From measured BMI to documented obesity in hospitalized adults: a decade-long analysis and implications for cardiometabolic care International Journal of Obesity. PMID 42687002
  3. [3] Chen SC, et al. (2026). Cardiorenal Mortality and Safety Outcomes of GLP-1 Receptor Agonists in Type 2 Diabetes With BMI Below 27 kg/m(2): A Target Trial Emulation Diabetes, Obesity and Metabolism. PMID 42642357

Where to get it

Best GLP-1 injections

Every injectable seller we can verify, with the price each one publishes and an honest read of what the trials measured.

Compare providers →

More in Evidence