Yogurt only lowers colorectal cancer risk in tumors with one specific gut bacterium

Eating yogurt has long been pitched as a simple, low-effort way to support gut health, and now a large new study gives that habit a more specific, and more complicated, connection to cancer. Researchers from Mass General Brigham and Harvard Medical School followed more than 132,000 people for up to three decades and found that regular yogurt eaters had a lower incidence of colorectal cancer, but only when their tumors carried a particular gut bacterium called Bifidobacterium. Tumors without that bacterium showed no such benefit, and in this analysis trended in the opposite direction.

Long-term yogurt intake tied to Bifidobacterium-positive tumors

The study, published in Gut Microbes, drew on decades of dietary questionnaires from the Nurses' Health Study and the Health Professionals Follow-up Study, two of the longest-running cohorts in American medicine. Among 3,079 people who developed colorectal cancer during follow-up, researchers had tumor tissue data on Bifidobacterium status for 1,121 cases. People who ate two or more servings of yogurt per week, compared with less than one serving a month, had a 20% lower incidence of colorectal cancer specifically in tumors positive for Bifidobacterium. Tumors that tested negative for the bacterium showed no such pattern.

Senior investigators Shuji Ogino and Tomotaka Ugai, both of Brigham and Women's Hospital, described the approach in a Mass General Brigham press release as an effort to connect long-term diet to a specific biological signature inside the tumor itself, rather than treating colorectal cancer as one uniform disease. Ugai said the findings suggest yogurt's apparent protective effect may be specific to Bifidobacterium-positive tumors, not colorectal cancer broadly.

Why this matters for how we think about diet and cancer prevention

Colorectal cancer is being diagnosed more often in adults under 50, and dietary advice for lowering risk has tended to be broad and generic: eat more fiber, cut red meat, and so on. This study pushes toward something more precise. Only about 31% of the tumors with available data were Bifidobacterium-positive, concentrated in the proximal colon (the right side, farther from the rectum). If yogurt's benefit really is limited to that subgroup, then broad public health messaging like "eat more yogurt to prevent colon cancer" oversells what the data show. It also raises the possibility that a person's own gut microbiome, and the microbial makeup a tumor develops with, shapes whether diet has any protective effect at all.

Methodology: an incident-tumor biobank approach

What sets this study apart methodologically is the use of an incident-tumor biobank method layered onto two prospective cohorts that had already tracked dietary habits with repeated questionnaires since 1976 (Nurses' Health Study) and 1986 (Health Professionals Follow-up Study). Rather than simply asking whether yogurt intake correlated with any colorectal cancer, the researchers went back into banked tumor tissue and tested for Bifidobacterium DNA, then ran inverse-probability-weighted Cox regression models to compare yogurt intake against each tumor subtype separately. That is a meaningfully different design than most diet-and-cancer studies, which rarely have access to tumor-level microbial data at this scale.

The catch: neither hazard ratio reached significance on its own

Here is the detail that headlines tend to skip. Comparing people who ate yogurt at least twice a week against those who ate it less than once a month, the hazard ratio for Bifidobacterium-positive tumors was 0.80 (95% CI 0.50 to 1.28), and for Bifidobacterium-negative tumors it was 1.09 (95% CI 0.81 to 1.46). Both confidence intervals cross 1.0, meaning neither result is statistically significant by itself. What is significant is the difference between the two subgroups: a P-heterogeneity value of 0.0002 overall, and 0.018 when the analysis was narrowed to proximal colon cancers specifically. In plain terms, the study found a real, statistically meaningful split in how yogurt tracked with two different tumor types, but it did not prove that yogurt lowers cancer risk on its own.

There are other limits worth naming. This is an observational cohort study, so it cannot establish that yogurt causes any change in cancer risk, only that the two are associated. Diet was self-reported through periodic questionnaires, which introduces recall error over time. And tumor-tissue Bifidobacterium data existed for only about 36% of the colorectal cancer cases identified (1,121 of 3,079), so more than half the cases in the cohort could not be classified either way.

What happens next

The researchers frame this as a step toward tumor-microbiome subtyping in cancer prevention research, rather than a finished dietary recommendation. Future work would need to replicate the Bifidobacterium-positive/negative split in independent cohorts, test whether specific yogurt strains or fermentation methods matter, and ideally run studies that measure a person's gut microbiome directly rather than inferring bacterial exposure from food-frequency questionnaires. Until then, the honest takeaway is narrower than "yogurt fights cancer": in this large, long-running dataset, yogurt's association with lower colorectal cancer risk showed up in one biologically defined tumor subtype, and only as part of a statistically significant split between subtypes, not as a proven effect on its own.

Evidence rating

EMERGING — 86/100

Caveats:

• Observational design — association only, not proof of causation; residual confounding cannot be ruled out.
• Tissue Bifidobacterium data were available for only 1,121 of 3,079 cancer cases; inverse probability weighting was used to adjust for this but selection bias cannot be fully excluded.
• Study populations were U.S. health professionals and predominantly non-Hispanic White; findings may not generalize to other populations.
• Yogurt intake was self-reported via food frequency questionnaire, introducing measurement error despite strong validation data.
• The finding for proximal colon cancer subgroup did not meet the pre-specified stringent alpha threshold of 0.005 (P heterogeneity = 0.018) and should be considered suggestive pending replication.
• Funding source was not disclosed in the article.

Large prospective observational analysis pooling two U.S. cohort studies (n=132,056; 3,079 colorectal cancer cases). The study demonstrates a statistically significant differential association between yogurt intake and colorectal cancer incidence by intratumoral Bifidobacterium abundance, but cannot establish causation. Tissue microbial data were available for roughly one-third of cancer cases, and findings in the proximal colon subgroup are suggestive rather than conclusive at the study's own stringent significance threshold. Independent replication is needed.

References

1. Satoko Ugai, Li Liu, Keisuke Kosumi, et al. (2025). Long-term yogurt intake and colorectal cancer incidence subclassified by Bifidobacterium abundance in tumor. Gut Microbes. https://doi.org/10.1080/19490976.2025.2452237