ABSTRACT
PURPOSE: We investigated whether the plasma level of 25-hydroxyvitamin D (25-OHD) after a diagnosis of colorectal cancer (CRC) influences survival outcome. PATIENTS AND METHODS: We prospectively studied 1,598 patients with stage I to III CRC. We sought association between plasma 25-OHD and stage-specific survival and tested for interaction between 25-OHD level and variation at the vitamin D receptor (VDR) gene locus. Blood was sampled postoperatively, and plasma was assayed for 25-OHD by liquid chromatography-tandem mass spectrometry. VDR polymorphisms (rs1544410, rs10735810, rs7975232, rs11568820) were genotyped, and haplotypes were inferred by using BEAGLE software. We tested for association between survival and 25-OHD, VDR genotype/haplotype, and after applying a VDR genotype-25-OHD interaction term. We conducted Kaplan-Meier survival analysis and used Cox proportional hazards models to estimate adjusted hazard ratios (HRs). RESULTS: We found strong associations between plasma 25-OHD concentration and CRC-specific (P = .008) and all-cause mortality (P = .003). Adjusted HRs were 0.68 (95% CI, 0.50 to 0.90) and 0.70 (95% CI, 0.55 to 0.89), respectively (highest v lowest 25-OHD tertile), particularly in stage II disease (HR, 0.44; P = .004 for CRC-specific mortality). We detected gene-environment interactions between 25-OHD concentration and rs11568820 genotype for CRC-specific (P = .008) and all-cause (P = .022) mortality, number of protective alleles (P = .004 and P = .018, respectively), and GAGC haplotype at the VDR locus for all-cause mortality (P = .008). CONCLUSION: In patients with stage I to III CRC, postoperative plasma vitamin D is associated with clinically important differences in survival outcome, higher levels being associated with better outcome. We observed interactions between 25-OHD level and VDR genotype, suggesting a causal relationship between vitamin D and survival. The influence of vitamin D supplementation on CRC outcome will require further investigation.