The Metabolism Supplement Guide: What the Ingredients Were Tested At

An essential trace mineral involved in insulin signalling. It is the ingredient most often framed around sweet cravings.

the Cochrane review of chromium picolinate pooled arms at 200, 400, 500 and 1,000 mcg a day were the arms pooled; no dose gradient emerged. Onakpoya's 2013 meta-analysis and a 2019 meta-analysis in Clinical Obesity reached modest positive figures, with a 2019 meta-analysis in Clinical Obesity finding improvements concentrated in trials of 12 weeks or less at doses up to 400 mcg a day. a 2024 dose-response meta-analysis pooled fourteen randomised trials; no significant change in weight, BMI, fat mass or waist.

What to read on a label: the microgram figure. Below about 200 mcg a day it sits under every arm the pooled reviews looked at.

Essential, involved in hundreds of enzyme reactions including several on the metabolic side.

Capdor's zinc meta-analysis pooled fourteen reports and 3,978 people; fasting glucose fell 0.19 mmol/L overall, with a larger effect in people who already had a metabolic condition. a 2024 GRADE-assessed review covered twenty-three trials; the authors rate the certainty of most outcomes low or very low. Maret and Sandstead on zinc requirements is the source to read on why more is not better.

What to read on a label: the milligram figure and the percentage beside it. A formula at half a reference intake is a top-up; one at several times it is a different decision and should be treated as one.

The best-evidenced thermogenic row in the category. Dulloo's 24-hour chamber study used 90 mg EGCG plus 50 mg caffeine, three times a day, with meals and measured 24-hour energy expenditure directly. a 2011 energy-expenditure meta-analysis pooled it: energy expenditure rose in proportion to the milligrams given, 0.4 to 0.5 kJ per mg. Hursel's 2009 meta-analysis covered weight, eleven studies; catechins moved body weight by 1.31 kg, and habitual caffeine intake above 300 mg a day blunted it.

It also carries the category's only widely published safety caution. the United States Pharmacopeia safety review reported case reports span EGCG intakes from 140 mg to about 1,000 mg a day; USP added a label caution to take the extract with food and not on an empty stomach, and the DILIN green tea analysis explains why the risk is unpredictable.

What to read on a label: the EGCG standardisation, and whether the product tells you to take it with food.

The largest randomised human literature of any botanical in this category. Yin's 2008 trial used 0.5 g three times a day, so 1,500 mg daily, over three months; Zhang's 2008 trial used 1.0 g a day against placebo for three months; Lan's 2015 meta-analysis found the pooled trials cluster around 0.9 to 1.5 g a day in divided doses; Ju's 2018 lipid meta-analysis reported gram-scale daily dosing across the included trials.

It is also the row with the clearest interaction question. Hermann and von Richter's interaction review names the herbs that matter at commonly recommended doses, and Qiu's pharmacokinetic study found 30 and 100 mg/kg in rats raised digoxin exposure by 23 to 70 per cent.

What to read on a label: the milligram figure. Gram-scale trials and a blend total in the low hundreds are not the same product.

Mousavi's dose-response meta-analysis pooled 28 trials; weight fell 0.51 kg, and the effect sat in trials under 500 mg a day running three months or longer. a 2021 umbrella review sits above those trials and reviews the meta-analyses rather than the trials, and grades the certainty low.

What to read on a label: the duration you are prepared to commit to. This is a row where three months is the shortest honest test.

Voroneanu's 2016 meta-analysis found five trials, 270 patients; fasting glucose fell 26.86 mg/dL, on evidence the authors themselves call low quality. a 2025 insulin-sensitivity meta-analysis required six studies, 673 participants, at least four weeks of supplementation each.

What to read on a label: the standardisation, usually 80 per cent. It is also named as a CYP2C9 inhibitor in Hermann and von Richter's interaction review, which matters if you take anything regularly.

Ludy and Mattes on red pepper served 1 g of red pepper served AT a meal; the effect was larger eaten than swallowed in a capsule. Janssens on capsaicin and fullness gave 2.56 mg of capsaicin, about 1.03 g of red chilli, with every meal.

What to read on a label: whether the product is designed around meals. This row's entire literature is about eating, and a once-daily capsule is not that design.

Genuinely split evidence. Vuksan's Korean red ginseng trial used 2 g at each meal, so 6 g a day, taken before eating and found a benefit. Reeds' negative ginseng trial used 8 g a day of root extract, or 250 to 500 mg a day of ginsenoside Re, for 30 days and found none.

What to read on a label: the gram figure, and whether both halves of this literature are acknowledged anywhere on the page selling it.

Stohs' banaba review is a narrative review across animal, human and in-vitro work rather than a pooled human dose.

What to read on a label: the standardisation percentage, usually 1 to 2 per cent. A percentage of an unstated amount is not a dose.

Kucukgoncu's weight meta-analysis pooled ten double-blind trials; 1.27 kg more weight lost than placebo, with no dose relationship detectable. a 2021 dose-response meta-analysis models the response against both the daily amount and the length of the trial.

What to read on a label: this is the one row where a smaller amount is not automatically worse, because no relationship between dose and result has been demonstrated in either direction.