Turmeric, curcumin
Also called: curcumin, turmeric, haldi, diferuloylmethane.
Last reviewed 2026-09-16. How this works.
Treating cancer Grade D
No human evidence of benefit, or none that survived testing.
Turmeric does not reach levels in a person anywhere near the levels used in a laboratory dish.
The bioavailability problem is the whole story.
The claim
The claim is that curcumin, the yellow compound in turmeric, kills cancer cells. It is supported by a very large amount of laboratory work, and that work is real.
What it does
Curcumin does affect many pathways in a dish. The obstacle is getting it into a person. It is poorly absorbed and heavily processed by the gut wall and liver before it reaches the bloodstream.
This entry carries the comparison rather than asserting it. In the laboratory, the concentration needed to halve the growth of triple-negative breast cancer cells was between roughly eleven and nineteen micromolar. In healthy volunteers taking a curcuminoid extract by mouth, the highest concentration anyone was measured to reach was about a seventh of one micromolar, and most volunteers had none detectable at all.21
The gap between those two numbers is around eighty-fold at its narrowest. It cannot be closed by taking more, and the trial itself shows why: the two volunteers who had any detectable were already at the highest amounts tested, and the study was designed to escalate until something appeared.1
The evidence
The register holds no citation for a trial showing that curcumin changes what a cancer does in a person. What it holds is the measurement that explains why.1
One further finding belongs here, because it applies to what is actually on sale. The investigators analysed the capsules they had used and found the curcumin activity of the capsule contents was substantially lower than that of an equivalent weight of pure curcumin powder. What is written on a supplement label is not necessarily what is in it.1
The studies this rests on
-
Dose escalation of a curcuminoid formulation in healthy volunteers1
A trial with no comparison group, open label, phase 1, measured a laboratory marker, not an outcome, 24 participants.
no curcumin detectable in serum below the two highest amounts tested; low levels in two of twenty-four volunteers.
What is absent
No trial showing that curcumin changes what a cancer does in a person is recorded in this register.
Scope: any tumour type, any endpoint.
Nobody has run a search to establish this, so it is the register’s reading of what exists rather than a gap somebody has looked for and confirmed.
What to ask
What reaches the blood is a small fraction of what the laboratory work used, so is there any version of this worth my time?
The dish-to-person gap
This is the register's worked example of the dish-to-person gap, and both sides of it are measured. In twenty-four healthy volunteers given a standardised curcuminoid extract by mouth, no curcumin at all was detectable in the blood at any of the lower amounts tested. Only two volunteers, at the two highest amounts, had any detectable, and what they reached was a small fraction of one micromolar.1
| Lowest laboratory concentration | 11.32 micromolar |
|---|---|
| Highest measured in a person | 0.137 micromolar1 |
| Gap | about 80-fold |
Where the laboratory concentrations come from
- 11.32 micromolar, curcumin, cancer cells: half-maximal inhibitory concentration at 72 hours by MTT assay, MDA-MB-231 triple-negative breast. The lowest of the three laboratory values, and around eighty times what a person was measured to reach2
- 16.23 micromolar, curcumin, cancer cells: half-maximal inhibitory concentration at 72 hours by MTT assay, BT-20 triple-negative breast2
- 18.61 micromolar, curcumin, cancer cells: half-maximal inhibitory concentration at 72 hours by MTT assay, MDA-MB-468 triple-negative breast2
The gap cannot be closed by taking more. Absorption has a ceiling, the liver and gut process most of what is swallowed before it reaches the blood, and for many compounds unpleasant effects arrive long before the laboratory concentration does.
The gap on this page must never be run in the other direction. It does not tell anyone how much to take, and the reason is on the page: the amounts that produced a detectable trace were already the highest tested and still fell far short.
The laboratory concentrations are from one cell-line study in triple-negative breast cancer. They must not be presented as the concentration required for any other cancer.
When to call your team — no warning signs are recorded
Whatever you are taking or thinking of taking, the people treating you can only weigh it 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 turmeric. That is not a statement that nothing can go wrong.
Interactions
The register holds no measured interaction between curcumin and a cancer treatment class. Chapter 16 of the book covers supplement interactions in general. An empty list means nobody has measured this one.
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
Trials of curcumin exist. What the register has no citation for is one showing an effect on a cancer.
Search ClinicalTrials.gov for recruiting studies of curcumin OR turmeric
A trial is the same decision with monitoring, a known preparation and a doctor who knows what else you are taking.
References
Lao CD, Ruffin MT 4th, Normolle D, Heath DD, Murray SI, Bailey JM, Boggs ME, Crowell J, Rock CL, Brenner DE. Dose escalation of a curcuminoid formulation. BMC Complementary and Alternative Medicine 2006;6:10.
How this was reached: transcribed from chapter note, via docs/CHAPTER-NOTES.json, chapter 2 note 1; the same paper is cited again at chapter 5 note 3 for the capsule-content finding.
Reported half-maximal inhibitory concentrations for curcumin in triple-negative breast cancer lines. Experimental and Therapeutic Medicine 2018.
How this was reached: transcribed from chapter note, via docs/CHAPTER-NOTES.json, chapter 2 note 5.