Garlic
Also called: allicin, aged garlic extract, garlic supplements.
Last reviewed 2026-09-16. How this works.
Preventing gastric cancer Grade C
Human evidence exists and does not settle it.
In a region where stomach cancer is common, garlic supplements were linked to fewer deaths from it.
A randomised trial: fewer deaths, no fewer cancers, in a very high-risk population.
The claim
The claim is that garlic prevents cancer, usually argued from studies of populations who eat a lot of it. Unlike most of this chapter, the claim has been tested.
What it does
Sulphur compounds in garlic are active in the laboratory. This entry is graded on the trial rather than on the mechanism, and the trial is the reason it is not a D.
The evidence
The Shandong Intervention Trial randomly assigned 3,365 residents of a high-risk county in China, blinded and placebo-controlled in a factorial design, to garlic supplementation, vitamin supplementation, treatment for H. pylori, or their placebos. Garlic was given for over seven years and participants were followed for more than twenty-two.1
Death from gastric cancer was less frequent in the garlic group, with a hazard ratio of 0.66 and an interval running from 0.43 to 1.00. The upper bound touches one exactly, which is as close to the line as a result can sit while still being reported as a reduction.1
The incidence of gastric cancer was not reduced. Nor, at an earlier readout, was the prevalence of the precancerous lesions that precede it. So the trial found fewer deaths from a cancer that occurred just as often, which is an odd combination and one the register records rather than explains.1
Three things limit what this transfers to. The population was at exceptionally high risk of gastric cancer, which is not most readers. What was given was a garlic extract and an oil, not garlic as food. And it was given for seven years. None of that is a general claim about cooking with garlic.1
One more thing in the same trial is worth seeing. Treating H. pylori and giving vitamins each reduced both mortality and incidence, and by more than garlic did. The strongest result in a garlic trial was not the garlic.1
The studies this rests on
-
Shandong Intervention Trial, garlic supplementation and gastric cancer (NCT00339768)1
A randomised trial, double blind, measured survival, 3365 participants.
gastric cancer mortality lower in the garlic group; incidence not reduced, and precancerous lesions not reduced at the earlier readout.
hazard ratio 0.66, interval 0.43 to 1. gastric cancer mortality, fully adjusted; in the same trial H. pylori treatment gave 0.62, interval 0.39 to 0.99, and vitamin supplementation 0.48, interval 0.31 to 0.75, both of which also reduced incidence
What is absent
No trial has tested garlic in a population at ordinary risk of gastric cancer, and none has tested garlic as food rather than as an extract and an oil. This result belongs to the population and the preparation that were studied.
Scope: populations at ordinary gastric cancer risk, and garlic as food.
Established by trial scope, checked 2026-09-16. Result: the only trial in this record enrolled residents of a high-risk county and supplied a garlic extract and oil.1
The finding rests on one trial in one county. Nobody has repeated it elsewhere.
Scope: any population.
Established by trial scope, checked 2026-09-16. Result: a single randomised trial in Linqu County is the whole of the randomised evidence in this record.1
What to ask
Does a garlic trial run in a population at very high risk of stomach cancer apply to me?
The dish-to-person gap
Garlic is eaten as food rather than taken as a measured compound, and the register holds no human pharmacokinetic measurement for it.
When to call your team — no warning signs are recorded
If you have changed how you eat because of something you have read, or are thinking about it, the people treating you can only weigh that if they know about it. That is true when nothing has been written down here, and it is most true then.
The register records no warning signs for this entry. That is not a statement that nothing can go wrong.
Interactions
The register holds no measured interaction between this food and a cancer treatment class. An empty list means nobody has measured it, not that nothing would happen. Concentrated extracts are a different question from the food.
Nobody having looked is not the same as nothing being there. That is what this line says and all it says.
The whole interaction list, and how to turn it into a pharmacy review request.
Find a trial instead
The trial behind this entry is complete and reported. What is not settled is whether it applies to anyone outside the population it studied.
Search ClinicalTrials.gov for recruiting studies of garlic OR allicin
A trial is the same decision with monitoring, a known preparation and a doctor who knows what else you are taking.
References
Effects of Helicobacter pylori treatment and vitamin and garlic supplementation on gastric cancer incidence and mortality: follow-up of a randomized intervention trial.
How this was reached: transcribed from chapter note, via docs/CHAPTER-NOTES.json, chapter 10 note 9.