Prospective Cohort Studies of Vitamin B6 Intake and Colorectal Cancer Incidence: Modification by Time?

Abstract

Abstract Background: Vitamin B6 may influence colorectal carcinogenesis through its role in one-carbon metabolism related DNA synthesis and methylation. However, observational studies have been inconclusive and no studies have investigated when in the natural history vitamin B6 intake may prevent colorectal cancer. Method: We followed 86,440 women in the Nurses' Health Study and 44,410 men in the Health Professionals Follow-up Study for up to 28 years. We assessed vitamin B6 intake every 4 years using validated food frequency questionnaires. We evaluated whether higher vitamin B6 intake in the remote past is strongly associated with a lower risk of colorectal cancer than intake in the recent past. Cox proportional hazards regression models were used to estimate multivariable relative risks (MV RRs, 95%CIs). Results: Comparing top with bottom quintiles of total vitamin B6 intake, the mean plasma pyridoxal 5–phosphate (PLP, the active form of vitamin B6) levels were 98.3 pmol/mL and 38.9 pmol/mL in women and were 183.2 pmol/mL and 66.0 pmol/mL in men. Total vitamin B6 intake was significantly associated with an approximately 20–30% lower risk of colorectal cancer in age-adjusted results but these significant associations became attenuated and non-significant after adjustment for other colorectal cancer risk factors. Compared extreme quintiles of cumulative intake of total vitamin B6, the MV RRs (95%CIs) for colorectal cancer were 0.98 (0.80, 1.22; P trend = 0.79) in women and 0.98 (0.76, 1.26; P trend = 0.60) in men. For the same comparison, the MV RRs were 0.92 (0.73, 1.16) for total vitamin B6 intake 0–4 year before diagnosis, 0.99 (0.78, 1.26) for intake 4–8 year before diagnosis, 0.93 (0.71, 1.21) for intake 8–12 year before diagnosis, and 0.93 (0.69, 1.26) for intake 12–16 years before diagnosis. The corresponding MV RRs for men were 0.85 (0.63, 1.16), 0.98 (0.70, 1.37), 0.90 (0.63, 1.28), and 1.19 (0.78, 1.83), respectively. Additionally, results did not differ by cancer sub-site, sources of vitamin B6 (food or supplement), or intake of alcohol and folate. Conclusion: Although a small effect cannot be excluded, our results do not support a strong role of vitamin B6 intake in adulthood in colorectal carcinogenesis among middle-aged U.S. health professionals.