Ultra-Processed Food Consumption and Colorectal Cancer Risk: A Systematic Review and Two-Stage Mediation Meta-analysis

Objective

To evaluate the association between ultra-processed food (UPF) consumption and colorectal cancer (CRC) risk and to assess whether inflammatory bowel disease (IBD) mediates this relationship using a two-stage mediation meta-analysis.

Design

We conducted a systematic review and meta-analysis of cohort studies on (i) UPF intake and CRC, (ii) UPF intake and IBD, and (iii) IBD and CRC. Two reviewers independently screened records, extracted hazard ratios (HRs) or relative risks, and assessed study quality with the Newcastle–Ottawa Scale (NOS). Random-effects pooling (DerSimonian–Laird) generated summary HRs. A two-stage product-of-coefficients approach combined the pooled HR for UPF → IBD with that for IBD → CRC to estimate an indirect (mediated) effect; the direct effect was calculated by dividing the total UPF → CRC HR by the indirect HR. Heterogeneity, publication bias, risk of bias, and overall certainty (GRADE) were evaluated.

Results

Of 7,152 records screened, three comparisons from two prospective cohorts (≈ 404 000 participants) met criteria for the UPF → CRC analysis. High UPF consumption was not significantly associated with CRC risk (pooled HR = 1.14; 95 % CI 0.96–1.35; I²≈ 45 %; Egger p = 0.60). Five cohorts informed the UPF → IBD analysis (pooled HR = 1.33; 95 % CI 1.13– 1.57; I²≈ 76 %), and four comparisons from two registry-based studies informed the IBD → CRC analysis (pooled HR = 1.23; 95 % CI 0.96–1.57; I²≈ 96 %).

The two-stage mediation yielded an indirect HR of ≈ 1.64, but because both the IBD → CRC and the total UPF → CRC estimates included unity, mediation could not be confirmed; the implied direct effect (≈ 0.70) was also statistically uncertain. NOS scores ranged 6 – 9, indicating generally good quality. Under GRADE, certainty was low for UPF → CRC, moderate for UPF → IBD, and low for IBD → CRC; certainty for the mediation analysis was very low.

Conclusion

Current cohort evidence shows no statistically significant overall association between high ultra-processed food intake and colorectal cancer risk, although UPF consumption is consistently linked to higher IBD incidence. Any IBD-mediated CRC risk appears to be small and uncertain, and residual confounding cannot be excluded. Large prospective studies and intervention trials are needed to clarify the direct pathways through which UPFs may influence colorectal carcinogenesis.