Berberine is sold online as a natural alternative to prescription weight-loss drugs, usually on evidence far thinner than the marketing. A randomised, double-blind, placebo-controlled trial published in Nutrients tested a fixed combination of berberine, white mulberry leaf extract and chromium picolinate in 93 Greek adults with obesity over 12 weeks, and found real differences against placebo: a lower BMI, less body fat, and meaningfully better insulin sensitivity. It is a small trial with a measurement problem, and it cannot tell you which of the three ingredients did anything.
What the trial did and what it found
Ninety-three community-dwelling adults with obesity and no diagnosed chronic disease were randomised, 46 to the active supplement and 47 to a matching placebo, for 12 weeks. Neither the participants nor the researchers knew who was getting which. The trial protocol lists the daily amounts as 1,000 mg of berberine, 400 mg of white mulberry leaf extract standardised to 5% 1-deoxynojirimycin, and 0.644 mg of chromium picolinate. The primary outcome was insulin sensitivity, measured with the Matsuda index derived from a 75 g oral glucose tolerance test — a reasonable choice, since that is a physiological measurement rather than a questionnaire.
On that primary outcome the supplement group improved 28% more than placebo (adjusted geometric mean ratio 1.28, 95% CI 1.14 to 1.44). The secondary outcomes moved too. BMI fell by 1.24 kg/m² more than placebo (95% CI 0.33 to 2.15), total body fat mass by 2.35 kg more (95% CI 0.58 to 4.12), and a visceral fat rating by 0.92 units more (95% CI 0.10 to 1.74). For a person 1.7 m tall, 1.24 kg/m² is roughly 3.6 kg of body weight. That is a genuine effect over three months, and it is also a fraction of what the trials of prescription drugs report — 15% to 20% of body weight in the strongest of them. The two figures belong on the same page.
Where the evidence runs out
Three limits matter more than the headline numbers. The first is the design: there were no single-ingredient arms. Berberine gets the attention and the search traffic, but this trial compared a three-part combination against placebo, so nothing here attributes the result to berberine rather than to the mulberry extract, the chromium, or the combination. The authors say so themselves.
The second is measurement. Body composition was assessed by bioelectrical impedance, not by DXA or imaging, and the “visceral fat rating” is an index a bioimpedance device produces, not a measured volume of visceral fat. Bioimpedance readings shift with hydration, which is exactly the kind of thing a supplement affecting glucose handling might also shift. The authors call for confirmatory trials using direct metabolic and imaging methods, which is the right call and an admission of where this one sits.
The third is size and duration. Ninety-three people over 12 weeks is enough to detect a moderate difference and not enough to characterise it well; the confidence intervals for body fat mass and the visceral rating come close to zero at the lower end, meaning the true effect could be considerably smaller than the point estimate. Participants had no diagnosed chronic disease, so this does not describe people with type 2 diabetes. There is no safety signal here, but 12 weeks is also short for one to appear, and supplement quality varies between products in a way it does not for a licensed medicine — what was tested was one fixed formulation, not the category. For how this compares with the rest of the evidence on over-the-counter products sold for weight loss, see our overview of weight-loss supplements.
