Table 2.
OR and 95% CI for metabolites associated with colorectal cancer risk at FDR < 0.2 in the Cancer Prevention Study II Nutrition Cohort (n = 517 matched cases and controls).
| Guanidinoacetate | Vanillylmandelate (VMA) | 3-methylxanthine | 2’-O-methylcytidine | Bilirubin (E,E) | N-palmitoylglycine | |
|---|---|---|---|---|---|---|
| Multivariable Adjusted Continuous Model a | ||||||
| Per SD | 1.32 (1.14, 1.52) | 1.29 (1.12, 1.49) | 0.79 (0.69, 0.89) | 1.27 (1.11, 1.46) | 1.29 (1.11, 1.50) | 1.27 (1.11, 1.45) |
| p | <0.001 | <0.001 | <0.001 | <0.001 | 0.001 | 0.001 |
| FDR | 0.090 | 0.090 | 0.090 | 0.090 | 0.121 | 0.106 |
| Continuous, Mutually Adjusted Model b | ||||||
| Per SD | 1.24 (1.07, 1.45) | 1.28 (1.09, 1.49) | 0.74 (0.64, 0.85) | 1.18 (1.02, 1.36) | 1.15 (0.97, 1.36) | 1.12 (0.96, 1.30) |
| p | 0.005 | 0.002 | <0.001 | 0.028 | 0.099 | 0.139 |
| Multivariable Adjusted Quartiles a | ||||||
| Q1 (ref) | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) |
| Q2 | 1.26 (0.86, 1.83) | 1.64 (1.13, 2.40) | 0.63 (0.44, 0.91) | 1.26 (0.89, 1.80) | 1.10 (0.76, 1.60) | 1.23 (0.85, 1.78) |
| Q3 | 1.42 (0.98, 2.06) | 1.51 (1.04, 2.18) | 0.65 (0.46, 0.93) | 1.71 (1.18, 2.47) | 1.51 (1.02, 2.24) | 1.40 (0.97, 2.01) |
| Q4 | 1.87 (1.26, 2.77) | 1.94 (1.31, 2.89) | 0.51 (0.35, 0.73) | 1.73 (1.20, 2.50) | 1.64 (1.10, 2.45) | 1.97 (1.36, 2.87) |
| p trend c | 0.001 | 0.003 | <0.001 | 0.001 | 0.008 | <0.001 |
| FDR | 0.220 | 0.236 | 0.220 | 0.220 | 0.272 | 0.220 |
Note: CI, confidence interval; FDR, false discovery rate-adjusted p values; OR, odds ratio; SD, standard deviation. a Odds ratios (95% confidence intervals) were estimated from conditional logistic regression models, matched on sex, race, age, and date of blood draw. Models were adjusted for hours since last meal (continuous), body mass index at blood draw (continuous, kg/m2), smoking status in 1999 (never, former, current, unknown), recreational physical activity in 1999 in metabolic equivalent (MET)-h/week (<8.75; 8.75–<17; 17+, unknown), alcohol consumption in 1999 (nondrinker, <1 drink/day, 1+ drinks/day, unknown), current NSAID use in 1999 (yes, no), ACS diet score (tertiles, comprised of scores for red and processed meat, proportion of whole vs. refined grains consumed and fruit and vegetable consumption plus a missing category), and total calories (continuous). Individuals with missing continuous variables (BMI, calories) were assigned the study median value. b Multivariable model, additionally adjusted for the other five metabolites. c p for trend based on median values in each quartile.