🔬 India has long had relatively little colorectal cancer, and for decades people have wondered whether curry has something to do with it. A team in Japan spent two years testing that idea on nearly 600 patients. Overall, the capsules made no difference to how often polyps came back. Fewer of them, though, grew large.

First, this is not about curry

The trial was run by Tokushima University and Kyoto Prefectural University of Medicine across 26 sites in Japan. The paper appeared in The American Journal of Gastroenterology on August 28, 2026, as an accepted manuscript, and Tokushima University announced it on September 25, 2026.

The compound under test was curcumin, the yellow pigment in turmeric and therefore in curry powder. But nobody in the trial was eating curry, and nobody was taking an ordinary turmeric supplement. Participants swallowed Theracurmin, a submicron-powdered curcumin designed for better absorption, at 180 mg twice a day.

According to the Linus Pauling Institute at Oregon State University, curcumin taken by mouth is poorly absorbed and quickly metabolized and eliminated. The team used a formulation with greatly improved absorption, so the results apply to that product at that dose, not to a plate of curry.

Why "India eats curry" proves nothing yet

The starting observation holds up. In figures for 2022 compiled by the World Cancer Research Fund (WCRF), India's age-standardized colorectal cancer rate was 4.9 per 100,000 people, against 18.4 worldwide and 36.6 in Japan.

A 2017 paper in the Indian Journal of Surgical Oncology lists reasons that have nothing to do with spices. India's population is younger, and colorectal cancer mostly strikes older people. At the time, the country's cancer registries covered only 7.45% of the population, so some cases may simply go uncounted. Diet is one suspect among several.

Numbers like these show that two things go together, not that one causes the other. The US National Cancer Institute (NCI) notes that observational studies are exposed to biases they cannot fully correct for, while randomized controlled trials avoid most of them.

So the Japanese team ran a randomized controlled trial. It enrolled patients aged 30 to 75 whose colorectal adenomas or early tumors had already been removed by endoscopy. An adenoma is a benign growth, often called a polyp, in the lining of the colon. Most colorectal cancers are thought to develop from adenomas, which is why doctors remove them. The catch is that new ones often appear afterward.

Between April 2016 and March 2020, 577 patients were assigned by chance to curcumin (287) or to identical-looking placebo capsules (290). The trial was double-blind: neither the patients nor their doctors knew who was getting which. After two years, participants had another colonoscopy, and the analysis included only those who completed it.

What the trial actually found

The trial's primary end point, the one main question it was designed to answer, was whether curcumin reduced recurrent adenomas of any size at the two-year colonoscopy. It did not. The risk ratio was 0.98, with a 95% confidence interval of 0.82 to 1.16. In plain terms, the two groups looked almost the same.

Tokushima University's release does report this, in its third section. Its headline says curcumin suppressed the recurrence of precancerous lesions. The paper's own title is more precise: curcumin "suppresses growth" of recurrent adenomas.

The difference showed up in a category that, according to the paper, the team had set in advance: adenomas 6 mm or larger, the size at which Japanese guidelines recommend endoscopic removal. Within two years, 13.3% of the placebo group developed one, compared with 7.5% of the curcumin group. That is a risk ratio of 0.56 (95% confidence interval 0.33 to 0.95, P = 0.04), or about 44% lower.

Other measures pointed the same way. Patients on curcumin had fewer adenomas of 6 mm or larger (P = 0.03), and judged by median size, the adenomas found in their colons were smaller overall (P = 0.003). Among patients who had 6 or more adenomas removed in the past, the risk of developing 4 or more new ones was lower, with a risk ratio of 0.40. Exploratory analyses also hinted at a stronger effect in men.

On safety, adverse events were no more common in the curcumin group, and treatment-related events were infrequent and mild.

The team's own interpretation is that curcumin does not so much prevent adenomas from forming as slow their growth or progression. The large-adenoma result clears the bar for statistical significance, but not by much. The upper end of the confidence interval sits at 0.95, close to "no effect," and the findings for high-risk patients and for men are the kind that suggest the next study rather than settle anything.

What shrank was not cancer

The trial showed a possible effect on large adenomas over two years. It did not show that curcumin prevents colorectal cancer.

The researchers describe the result as the first evidence of its kind worldwide pointing to a possible colorectal cancer prevention effect, and they say longer follow-up is needed to test whether curcumin prevents the cancer itself. The team also notes that removing adenomas by endoscopy already reduces deaths from colorectal cancer. Curcumin was tested on top of that, not instead of it.

Japan's own rate is about twice the world average. For one line of research into why, see our earlier report on the gut-bacteria toxin linked to young-onset colorectal cancer in Japan.

Promising molecules that fizzled

NCI summarizes two large trials of beta-carotene. In Finland's ATBC study, middle-aged male smokers took 20 mg of beta-carotene a day for 5 to 8 years, and the beta-carotene group ended up with more lung cancer, not less. In the US CARET trial, smokers and people exposed to asbestos took 15 mg of beta-carotene plus retinol. Lung cancer and overall deaths rose, and the trial was stopped early.

Curcumin has had its own disappointment. In 2018, a trial in the US and Puerto Rico gave 44 patients with familial adenomatous polyposis, an inherited condition that causes hundreds of polyps, 1,500 mg of pure curcumin twice a day for 12 months. It found no significant difference in polyp number or size. That was a different disease, a smaller group and a different form of curcumin, so the two trials are not directly comparable.

Against that background, the Japanese trial's strength is its design: far larger and longer than the 2018 trial, double-blind and publicly registered.

Not the bottle on the shelf

Turmeric and curcumin supplements are easy to find in pharmacies and online, but the trial did not test any of them. In Japan, a public health and environment laboratory run by Shizuoka Prefecture notes that turmeric health foods are marketed as good for the stomach and liver, especially for people who drink a lot. When it tested 20 such products in fiscal 2013, the amount of curcuminoids differed by a factor of about 10 between brands.

The maker of the trial product, Theravalues, supplied the capsules and placebos free of charge and is also listed in Japan's trial registry as a source of funding. The study was additionally funded by a commissioned research project of Japan's Ministry of Agriculture, Forestry and Fisheries. Neither fact invalidates the result on its own, and both are disclosed in the registry.

The US National Center for Complementary and Integrative Health (NCCIH) warns that some people have developed liver damage after taking curcumin products formulated for higher absorption. It considers conventional turmeric or curcumin probably safe for 2 or 3 months at the recommended amounts. The Japanese trial saw no increase in side effects over two years, but that was in monitored patients, which is not the same as people dosing themselves.

In Japan, turmeric already sells on claims aimed at people who drink, and now its name is attached to a colon-polyp study as well. Where you live, how much do people trust the idea that a single ingredient in food can keep a serious disease away?

References