Across the studied ingredients, the supplement literature produces effects that are statistically significant and clinically trivial — and in several cases not statistically significant either. The largest review of Chinese herbal medicine for weight management found a signal only where the herbs were added to something else, in an evidence base its own authors describe as mostly at high or unclear risk of bias (Wong 2021). Cochrane's review of green tea concluded the weight loss was "small, statistically non-significant" and "not likely to be clinically important" (Jurgens 2012). Sitting behind all of it is a regulatory problem: HSA has found products sold in Singapore as "all-natural" that contained a prescription weight-loss drug banned here since 2010 (HSA 2026).
This article goes through what each of those literatures actually reports, with the sample sizes, the confidence intervals and the authors' own verdicts.
Adulterated product: a product that contains an undeclared active pharmaceutical ingredient — a medicine that is not on the label. It is the specific failure mode regulators keep finding in weight-loss products, and it is why a product being labelled "natural" carries no information about what is inside it.
What the Chinese herbal medicine trials found
A 2021 systematic review searched 18 databases in English, Chinese, Korean and Japanese to September 2019, and assessed the trials with the Cochrane risk-of-bias tool. Thirty-nine randomised controlled trials were eligible for qualitative analysis and 34 entered the meta-analyses. Controls were placebo, Western medication or lifestyle intervention, with or without co-interventions.
The finding worth reporting precisely is the shape of the effect rather than its size. Twenty-five studies involving co-interventions found Chinese herbal medicine had significant adjunct effects on body weight and BMI at the end of treatment versus control. No serious adverse events were reported in the herbal groups. The authors' own conclusion is that it "indicates a promising adjunct to facilitate WM or lifestyle change for weight management" — immediately followed by the qualification that "methodological barriers such as lack of allocation concealment and double-blinding may have led to challenges in data synthesis" (Wong 2021).
Their risk-of-bias verdict is the sentence that governs how far this can be carried: "The majority of studies had a high or unclear risk of selection, performance, and detection bias." No effect size is quoted here, because the mean differences and confidence intervals were not retrievable from the record used to verify this source, and no verified figure is available.
So the accurate summary is a positive adjunct signal, in a low-quality evidence base,. What the review does not support is a comparison against any other treatment: it did not run one.
What the green tea trials found
Cochrane pooled 14 randomised trials of at least 12 weeks in adults with overweight or obesity — six conducted outside Japan (532 participants) and eight in Japan (1,030 participants). In the non-Japan studies, the mean difference in weight loss was −0.04 kg (95% CI −0.5 to 0.4; P=0.88). BMI fell 0.2 kg/m² (95% CI −0.5 to 0.1; P=0.21) and waist circumference 0.2 cm (95% CI −1.4 to 0.9; P=0.70) — neither significant. The Japan studies ranged from −0.2 kg to −3.5 kg and were reported separately rather than pooled with the rest, because they were heterogeneous.
Two studies with 184 participants looked at weight maintenance and found changes ranging from +0.6 kg to −1.6 kg. The review's verbatim conclusion: "Green tea preparations appear to induce a small, statistically non-significant weight loss in overweight or obese adults. Because the amount of weight loss is small, it is not likely to be clinically important. Green tea had no significant effect on the maintenance of weight loss" (Jurgens 2012). This is a 2012 review, and it should be read with that date attached.
Efficacy is only half the entry. NIH's LiverTox database gives green tea extract its top likelihood score — "A (well established cause of clinically apparent liver injury)" — with more than 100 documented instances in the literature, typical onset one to six months after starting, an acute hepatitis-like presentation, and described fatal cases of acute liver failure (LiverTox). The signal attaches to concentrated extracts, especially in weight-loss products, not to drinking tea. No incidence rate can be attached to it, because this is case-report literature rather than a cohort study.
What the garcinia trials found
A 2011 systematic review and meta-analysis identified 23 randomised trials, included 12 and pooled nine. The pooled mean difference favouring hydroxycitric acid over placebo was −0.88 kg (95% CI −1.75 to −0.00) — an upper bound touching zero. Restricting the analysis to rigorous randomised trials left the effect no longer statistically significant. Gastrointestinal adverse events were twice as common in the treatment group as in placebo in one included study.
The authors' own conclusion is quotable in full: garcinia extracts "can cause short-term weight loss. The magnitude of the effect is small, and the clinical relevance is uncertain. Future trials should be more rigorous and better reported" (Onakpoya 2011). Neither "it works" nor "it is proven useless" is what that says. The effect is too small and too fragile to change anything.
"Detox" and "cleanse" products: what analysis found inside them
Detox and cleanse framing usually promises the removal of something unspecified. What regulators find when they analyse the products is more concrete.
HSA's consumer safety article describes two cases. In the first, a woman had chest pains after taking a two-part regimen sold on social media as a "health sculpting programme": the morning product was found to contain sibutramine, and the night product contained sennosides — a laxative. Both were labelled "all-natural" and carried manufacturing-standard logos they were not entitled to. In the second, a consumer reported "insomnia, illusions, heart palpitations and trembling of limbs" after a herbal product labelled "100% natural ingredients" that contained sibutramine; a companion product in the same range contained diclofenac and phenolphthalein (HSA 2026).
The regulatory facts attached to those ingredients matter. Sibutramine was, in HSA's words, "previously a prescription only weight loss medicine used in the treatment of obesity" and "has been disallowed for sale in Singapore since 2010 due to increased risk of heart attacks and strokes". Phenolphthalein has not been registered in Singapore since 2011. A laxative in a slimming product produces water loss through the bowel, which moves a scale without touching body fat.
None of this says that all supplements or all online sellers are adulterated. These are specific enforcement findings about specific products. What it does establish is that the word "natural" on a label is a marketing description, not an analytical result.
Why "statistically significant" and "clinically important" are different
This is the sentence that unlocks most supplement research. A statistically significant result means the observed difference is unlikely to be chance. It says nothing about whether the difference is large enough to matter to a person.
Garcinia's 0.88 kg is the clean example: a real pooled estimate, an interval that reaches zero, and a magnitude the authors themselves decline to call clinically relevant (Onakpoya 2011). Cochrane's green tea result is the other pattern — an effect so small it did not reach significance at all, in a review large enough to have detected a meaningful one (Jurgens 2012).
Every number in this article is a pooled average across a trial population, and individual results vary. When a supplement result is quoted next to a figure from a prescription medicine trial, the two are usually on different scales and from different study designs, and putting them side by side without saying so misleads by arithmetic rather than by assertion.
What is worth checking before buying anything
Four practical checks, none of which require judging a seller:
- Where is it sold from? In 2023 about 14% of illegal health-product listings on major local e-commerce platforms came from overseas sellers, and HSA "does not have extra-territorial powers" (MOH 2024).
- What does the law say about supply? Supplying an unregistered health product carries a fine of up to $50,000 or two years' imprisonment; supplying an adulterated, counterfeit or tampered product carries up to $100,000 or three years (Health Products Act 2007).
- Does the claim name an outcome or an ingredient? HSA's own framing is that product claims in this category "are often exaggerated or unsubstantiated by any scientific evidence and impossible to verify" (HSA 2026).
- Is a rapid result being promised? HSA's advice reads: "There is no quick and easy way to lose weight – it's about achieving and maintaining a lifestyle that incorporates the principles of healthy eating and active living. If you need help in managing your weight, please consult your doctor, dietitian or a trained healthcare professional."
That last line is a regulator's general advice about seeking professional help, and it should not be read as an endorsement of any particular treatment, including ours. If you want a clinical view on where you stand, you can check your eligibility and speak to a doctor about what is suitable.
Common questions
Does Chinese herbal medicine help with weight loss?
The largest review of the question pooled 34 randomised trials and found significant effects on body weight and BMI only where the herbal medicine was used as an adjunct alongside medication or a lifestyle programme, with no serious adverse events reported in the herbal groups. The same review states that most of the included studies carried a high or unclear risk of selection, performance and detection bias (Wong 2021). A positive signal in a low-quality evidence base is a genuine finding and a limited one.
Does green tea help you lose weight?
Barely, on the pooled evidence. A Cochrane review of 14 randomised trials found a mean difference of −0.04 kg in the non-Japan studies (95% CI −0.5 to 0.4, P=0.88), and its authors concluded green tea preparations produce a small, statistically non-significant weight loss that is not likely to be clinically important, with no effect on maintaining a loss (Jurgens 2012). That review dates from 2012.
Is green tea extract safe?
Concentrated extract carries a documented liver signal that the beverage does not. NIH's LiverTox database gives green tea extract its highest likelihood score — a well-established cause of clinically apparent liver injury — with more than 100 reported cases, typical onset one to six months after starting, and described fatal cases of acute liver failure (LiverTox). The signal is attached to concentrated extracts in supplement products, not to drinking tea.
Does garcinia cambogia work?
The pooled estimate is 0.88 kg more weight loss than placebo across nine randomised trials, with a confidence interval reaching 0.00 and no statistically significant effect left once the analysis is restricted to rigorous trials. The review's own conclusion is that the magnitude is small and the clinical relevance uncertain (Onakpoya 2011).
Are slimming supplements sold online in Singapore safe?
Some have been found to contain undeclared prescription drugs. HSA has published cases of weight-loss products sold online here, labelled all-natural and carrying false manufacturing logos, that on analysis contained sibutramine — banned in Singapore since 2010 for increased risk of heart attack and stroke — along with undeclared diclofenac, phenolphthalein and laxatives. Consumers reported chest pains, insomnia, heart palpitations and trembling limbs (HSA 2026).