Abstract
The influence of polyunsaturated fatty acids (PUFAs) on risk of colorectal cancer precursors remains largely unknown. We examined the associations of erythrocyte PUFAs, including n-3 and n-6 PUFAs, with risk of colorectal conventional adenomas and serrated polyps in 4517 participants from three US prospective cohorts who had provided a blood sample and undergone at least one endoscopic examination. We calculated the multivariable odds ratios (ORs) per 1 SD increment in individual PUFAs and the ratio of n-6/n-3 PUFAs. We considered P < .005 statistically significant to account for multiple testing. During a median of 20 years of follow-up, we documented 493 conventional adenomas and 316 serrated polyps. After adjusting for various CRC risk factors, no associations for PUFAs achieved the stringent statistical significance for either conventional adenomas or serrated polyps (ORs per 1 SD ranged from 0.90 to 1.14). Some associations achieved nominal significance (P < .05), including the association of dihomogammalinolenic acid (DGLA) (20:3, n-6) with lower risk of conventional adenomas (OR = 0.91; 95% confidence interval [CI] = 0.83-1.00), total n-6 PUFAs with higher risk of proximal serrated polyps (OR = 1.32; 95% CI = 1.01-1.74) and eicosadienoic acid (20:2, n-6) and DGLA with lower risk of advanced adenomas (OR = 0.83; 95% CI = 0.71-0.97 and OR = 0.84; 95% CI = 0.72-0.98, respectively). Our findings indicate that erythrocyte PUFAs in a typical American diet are unlikely to have a substantial influence on risk of colorectal cancer precursors. The subgroup associations require further confirmation.