Phyto-oestrogens and colorectal cancer risk: a systematic review and dose-response meta-analysis of observational studies

Epidemiological studies suggest that soya consumption as a source of phyto-oestrogens and isoflavones may be associated with a reduced risk of colorectal cancer. However, findings have not yet been synthesised for all groups of phyto-oestrogens. A meta-analysis was conducted to quantify the association between phyto-oestrogens and colorectal cancer risk. Relevant observational studies published up to June 2016 were identified by searching MEDLINE, EMBASE and Cochrane Library databases. Study-specific relative risks (RR) were pooled in both categorical and dose-response meta-analyses. Out of seventeen identified studies, sixteen were included in the meta-analysis. Comparing the highest with the lowest intake category, inverse associations for phyto-oestrogens overall and by subgroup were observed but were statistically significant in case-controls studies and not in cohort studies. The pooled RR in case-control studies were 0.76 (95% CI 0.69, 0.84), 0.77 (95% CI 0.69, 0.85) and 0.70 (95% CI 0.56, 0.89) for phyto-oestrogens, isoflavones and lignans, respectively, whereas the corresponding pooled RR were 0.95 (95% CI 0.85, 1.06), 0.94 (95% CI 0.84, 1.05) and 1.00 (95% CI 0.64, 1.57) in cohort studies. Dose-response analysis yielded an 8% reduced risk of colorectal neoplasms for every 20 mg/d increase in isoflavones intake in Asians (pooled RR 0.92; 95% CI 0.86, 0.97). A non-linear inverse association with colorectal cancer risk was found for lignans intake, but no association for circulating enterolactone concentrations was observed. Thus, study heterogeneity precludes a rigorous conclusion regarding an effect of high exposure to isoflavones on risk of colorectal cancer. Current evidence for an association with lignans exposure is limited. Further prospective studies, particularly evaluating lignans, are warranted to clarify the association between different phyto-oestrogens and colorectal cancer risk.