Long-term yogurt intake and colorectal cancer incidence subclassified by Bifidobacterium abundance in tumor.

Abstract

Evidence suggests a tumor-suppressive effect of the intake of yogurt, which typically contains Bifidobacterium. We hypothesized that long-term yogurt intake might be associated with colorectal cancer incidence differentially by tumor subgroups according to the amount of tissue Bifidobacterium. We utilized the prospective cohort incident-tumor biobank method and resources of two prospective cohort studies. Inverse probability weighted multivariable Cox proportional hazards regression was used to assess differential associations of yogurt intake with the incidence of colorectal carcinomas subclassified by the abundance of tumor tissue Bifidobacterium. During follow-up of 132,056 individuals, we documented 3,079 incident colorectal cancer cases, including 1,121 with available tissue Bifidobacterium data. The association between long-term yogurt intake and colorectal cancer incidence differed by Bifidobacterium abundance (P heterogeneity = 0.0002). Multivariable-adjusted hazard ratios (HRs) (with 95% confidence intervals) in individuals who consumed ≥2 servings/week (vs. <1 serving/month) of yogurt were 0.80 (0.50-1.28) for Bifidobacterium-positive tumor and 1.09 (0.81-1.46) for Bifidobacterium-negative tumor. This differential association was also observed in a subgroup analysis of proximal colon cancer (P heterogeneity = 0.018). Long-term yogurt intake may be differentially associated with the incidence of proximal colon cancer according to Bifidobacterium abundance, suggesting the antitumor effect of yogurt intake on the specific tumor subgroup.