AIM: To examine the relationships between pre-diagnostic biomarkers and colorectal cancer risk and assess their relevance in predictive models.METHODS: A nested case-control study was designed to include all first primary incident colorectal cancer cases diagnosed between inclusion in the SUpplémentation en VItamines et Minéraux AntioXydants cohort in 1994 and the end of follow-up in 2007. Cases (n = 50) were matched with two randomly selected controls (n = 100). Conditional logistic regression models were used to investigate the associations between pre-diagnostic levels of hs-CRP, adiponectin, leptin, soluble vascular cell adhesion molecule-1 (sVCAM-1), soluble intercellular adhesion molecule-1, E-selectin, monocyte chemoattractant protein-1 and colorectal cancer risk. Area under the receiver operating curves (AUC) and relative integrated discrimination improvement (RIDI) statistics were used to assess the discriminatory potential of the models. RESULTS: Plasma adiponectin level was associated with decreased colorectal cancer risk (P for linear trend = 0.03). Quartiles of sVCAM-1 were associated with increased colorectal cancer risk (P for linear trend = 0.02). No association was observed with any of the other biomarkers. Compared to standard models with known risk factors, those including both adiponectin and sVCAM-1 had substantially improved performance for colorectal cancer risk prediction (P for AUC improvement = 0.01, RIDI = 26.5%). CONCLUSION: These results suggest that pre-diagnostic plasma adiponectin and sVCAM-1 levels are associated with decreased and increased colorectal cancer risk, respectively. These relationships must be confirmed in large validation studies. AIM: To examine the relationships between pre-diagnostic biomarkers and colorectal cancer risk and assess their relevance in predictive models. METHODS: A nested case-control study was designed to include all first primary incident colorectal cancer cases diagnosed between inclusion in the SUpplémentation en VItamines et Minéraux AntioXydants cohort in 1994 and the end of follow-up in 2007. Cases (n = 50) were matched with two randomly selected controls (n = 100). Conditional logistic regression models were used to investigate the associations between pre-diagnostic levels of hs-CRP, adiponectin, leptin, soluble vascular cell adhesion molecule-1 (sVCAM-1), soluble intercellular adhesion molecule- 1, E-selectin, monocyte chemoattractant protein- 1 and colorectal cancer risk. AUC) and relative integrated discrimination improvement (RIDI) statistics were used to assess the discriminatory potential of the models. RESULTS: Plasma adiponectin level wa s associated with decreased colorectal cancer risk (P for linear trend = 0.03). Quartiles of sVCAM-1 were associated with increased colorectal cancer risk (P for linear trend = 0.02). No association was observed with any of the other biomarkers. Compared to standard models with known risk factors, including both adiponectin and sVCAM-1 had substantially improved performance for colorectal cancer risk prediction (P for AUC improvement = 0.01, RIDI = 26.5%). CONCLUSION: These results suggest that pre-diagnostic plasma adiponectin and sVCAM-1 levels are associated with decreased and increased colorectal cancer risk, respectively. These relationships must be be in large validation studies.