Berberine

Claimed to lower blood glucose and HbA1c comparably to metformin, lower LDL cholesterol and triglycerides, cause weight loss and improve insulin sensitivity and the gut microbiome. It is widely marketed as "nature’s Ozempic", a comparison with no randomised trial support of that kind.

The reputation rests on a literature whose provenance decides how much weight it can carry. In a 50-trial, 4,150-participant meta-analysis, 49 of the 50 trials were conducted in China and one in Iran, only four were published in English, 31 of 50 were rated low quality on the modified Jadad scale, only four were at low risk of bias, 31 were unclear for selection bias and 44 had unclear blinding. For berberine alone the HbA1c effect was −0.24% (95% CI −0.60 to 0.11), p=0.181 — not statistically significant. The larger HbA1c reduction of −0.69% came from berberine combined with hypoglycaemic drugs and cannot be attributed to berberine independently of the co-administered drug. A separate 12-trial placebo-controlled analysis did find significant lipid and glucose effects, with BMI moving only −0.435 kg/m² (upper bound −0.013) and blood pressure not moving at all; its authors state the evidence base is insufficient to claim syndrome-level remission.

Safety knowledge is limited and the regulatory direction is restrictive. EFSA could not establish any safe intake level. Signals cited by EFSA and ANSES include genotoxicity in vitro, liver carcinogenic activity in a two-year NTP mouse study of goldenseal-derived berberine, 45 cases of liver injury associated with berberine from greater celandine described as idiosyncratic, gastrointestinal disorders, hypoglycaemia, hypotension and drug interactions with significant data gaps. Trial-level adverse events looked benign (pooled RR 1.07) but with a 95% CI of 0.15 to 6.76 that is uninformative, from trials lasting 84 to 140 days.