Plasma n-3 polyunsaturated fatty acid levels and colorectal cancer risk in the UK Biobank: Evidence of non-linearity, as well as tumour site- and sex-specificity

Abstract Background: The relationship between n-3 polyunsaturated fatty acid (PUFA) intake and colorectal cancer (CRC) risk is unclear. Blood n-3 PUFA concentration is a biomarker of dietary n-3 PUFA intake. We examined the relationship between plasma n-3 PUFA concentrations and CRC risk in UK Biobank (UKBB) participants. Methods: We analysed the relationship between tertiles (T) of plasma total n-3 PUFAs and n-3 PUFA docosahexaenoic acid (DHA) levels, and overall CRC (also stratified by tumour location and sex) risk. Cox proportional hazards regression models were adjusted for clinical co-variates. Non-linearity was tested by restricted cubic splines. Results: There were 2,602 incident CRC cases in 234,598 UKBB participants with baseline plasma fatty acid data (mean follow-up 13.4 years). There was an inverse association between the plasma total n-3 PUFA level (T2 hazard ratio [HR] 0.88[95% confidence interval 0.80-0.97] compared with the T1 reference; T3 0.91[0.83-1.00]), as well as the plasma DHA concentration (T2 0.89[0.80-0.98]; T3 0.91[0.82-1.00]), and CRC risk. The relationship was non-linear (P for non-linearity=0.14 [total n-3 PUFAs] and 0.008 [DHA]), with a plateau at the highest n-3 PUFA concentrations. The relationship was more pronounced for proximal colon cancer (T2 0.82[0.69-0.97], T3 0.76[0.64-0.90] for DHA) and was evident for males (T2 0.84[0.74-0.95], T3 0.89[0.78-1.00]), but not females. Conclusions: Higher plasma n-3 PUFAs are associated with reduced CRC risk in the UKBB. Impact: Non-linearity, tumour site- and sex-specificity of the inverse relationship between plasma n-3 PUFA levels and CRC risk, if confirmed in other diverse populations, have significant implications for nutritional prevention guidelines.