Folate is important for DNA methylation and repair, and abnormalities in DNA methylation may contribute to loss of normal control of proto-oncogene expression. Folate and methionine are involved in the production of S-adenosylmethionine, the primary methyl donor in the body. Alcohol consumption may increase folate requirements and cause relative folate deficiency,1, 2 which has been related to increased breast cancer risk.3, 4, 5
Folate deficiency, or simply low levels of folate intake, may be related to the risk of colorectal cancer, possibly together with elevated alcohol drinking.2, 6, 7, 8 Thus, in the American Health Professionals cohort study, based on 205 cases of colon cancer, the relative risk (RR) was 3.3 for the combination of low folate, low methionine and high alcohol intake compared to low alcohol, high methionine, high folate intake among nonusers of aspirin.2 In a cohort study based on the Canadian National Breast Screening Study, the RR was 0.6 for the highest level of folate intake.9 A large, multicentric U.S. case-control study10 found no consistent support for an association between folate, other factors involved in DNA methylation and colon cancer risk; in the NHANES I epidemiologic follow-up study, the RR was 1.6 for subjects reporting low-folate, low-methionine and high-alcohol diets compared to those reporting low-alcohol, high-folate and high-methionine diets. The association was stronger in men.11 A multicentric Italian case-control study found RRs of 0.7–0.8 in the highest quintile of folic acid compared to the lowest.12
Regular alcohol drinking is frequent in the Italian population.13 We decided, therefore, to further consider the effect of folate consumption in the Italian multicentric study,12 using a comprehensive food-composition database including more than 60 natural sources of folate.14 We also considered the interaction of folate with alcohol and methionine on the risk of colorectal cancer in Italy.
The data were derived from a case-control study of colorectal cancer conducted between January 1992 and June 1996 in 6 Italian areas (4 in the north, 1 in the center, 1 in the south).12, 15 Cases were 1,225 subjects with histologically confirmed, incident cancer of the colon and 728 with cancer of the rectum, aged 23–74 years (median 62). Controls were patients admitted to the same hospitals for acute, non-neoplastic conditions unrelated to digestive tract diseases. A total of 4,154 control subjects, aged 20–74 years (median 58), were interviewed. Of these, 23% were admitted for traumas, 28% for other orthopedic disorders, 20% for acute surgical conditions, 19% for eye diseases and 10% for miscellaneous other illnesses, such as ear-nose-and-throat, skin and dental conditions. Fewer than 4% of subjects approached for interview (cases and controls) refused to participate.
Information was collected in hospital by trained interviewers using a validated food-frequency questionnaire,16 including 78 foods or groups of foods, various alcoholic beverages and fat-intake pattern. To compute nutrient intake, Italian food-composition databases were used, integrated by information from the manufacturers.14 The main sources of folate were green leafy vegetables, pulses, (whole) grains and liver. To control for potential confounding factors, multiple logistic regression was used, including terms for sex, age, center, education, physical activity, family history of colorectal cancer and total energy intake, using the residual method. Additional allowance for regular aspirin use and selected dietary indicators, including measures of fiber and vitamin C intake, did not materially modify any of the risk estimates.
Table I gives the distribution of cases and controls according to quintile of intake (based on the distribution of cases and controls combined) of folate in the overall data set and in strata of sex, subsite, alcohol drinking and methionine intake. In the overall data set, the odds ratio (OR) was 0.72 for the highest quintile of folate and the continuous OR per 100 μg was 0.86. The inverse relation was similar in men and women and somewhat stronger for the rectum (OR = 0.59 for the highest quintile) compared to the colon (OR = 0.81). No apparent heterogeneity was observed across colon subsites. The inverse relation was also somewhat stronger in the highest tertile of alcohol drinking (OR = 0.65), though trends were not heterogeneous across strata of alcohol, whereas no consistent difference was observed across strata of methionine intake. The correlation coefficients of folate were 0.71 with methionine and 0.14 with alcohol intake.
|Stratum||Quintile1 of folate intake||χ2trend||Per 100 μg|
|OR2 (95% CI)2||1 (referent)||0.88 (0.74–1.04)||0.85 (0.71–1.01)||0.86 (0.73–1.03)||0.72 (0.60–0.86)||10.50 (p < 0.01)||0.86 (0.78–0.95)|
|OR (95% CI)||1 (referent)||0.91 (0.73–1.16)||0.91 (0.73–1.14)||0.87 (0.69–1.09)||0.65 (0.50–0.85)||7.97 (p < 0.01)||0.84 (0.73–0.96)|
|OR (95% CI)||1 (referent)||0.83 (0.62–1.11)||0.74 (0.55–0.99)||0.81 (0.61–1.07)||0.70 (0.53–0.93)||4.84 (p < 0.03)||0.85 (0.72–0.99)|
|OR (95% CI)||1 (referent)||0.94 (0.76–1.15)||0.89 (0.72–1.10)||0.90 (0.73–1.10)||0.81 (0.66–1.01)||3.45 (p = 0.06)||0.90 (0.80–1.01)|
|OR (95% CI)||1 (referent)||0.79 (0.62–1.01)||0.79 (0.62–1.01)||0.84 (0.66–1.07)||0.59 (0.45–0.77)||10.63 (p < 0.01)||0.81 (0.70–0.93)|
|OR (95% CI)||1 (referent)||1.15 (0.81–1.62)||0.96 (0.68–1.34)||0.84 (0.60–1.19)||0.78 (0.56–1.09)||5.27 (p < 0.03)||0.73 (0.60–0.89)|
|OR (95% CI)||1 (referent)||0.80 (0.58–1.10)||0.70 (0.51–0.96)||0.86 (0.63–1.18)||0.71 (0.51–0.98)||2.32 (p = 0.13)||0.92 (0.76–1.11)|
|OR (95% CI)||1 (referent)||0.80 (0.61–1.04)||0.94 (0.72–1.24)||0.89 (0.66–1.19)||0.65 (0.46–0.93)||3.02 (p = 0.08)||0.87 (0.74–1.02)|
|OR (95% CI)||1 (referent)||0.85 (0.64–1.12)||0.92 (0.69–1.23)||0.86 (0.63–1.17)||0.69 (0.51–0.94)||4.23 (p < 0.04)||0.83 (0.67–1.02)|
|OR (95% CI)||1 (referent)||1.05 (0.77–1.43)||0.96 (0.71–1.31)||0.89 (0.64–1.22)||0.81 (0.58–1.14)||2.23 (p = 0.13)||0.90 (0.74–1.09)|
|OR (95% CI)||1 (referent)||0.68 (0.49–0.94)||0.61 (0.44–0.85)||0.72 (0.52–0.99)||0.54 (0.39–0.75)||8.36 (p < 0.01)||0.84 (0.72–0.98)|
Table II gives the ORs for the combination of folate, alcohol and methionine. Compared to subjects reporting low alcohol (tertiles), high methionine and high folate intake (quintiles), the ORs of colorectal cancer were 1.66 in the intermediate level and 1.83 for subjects reporting high alcohol, low methionine and low folate intake. The association was similar for colon and rectal cancers, even after excluding regular aspirin users (ORs for colorectal cancer were 1.67 in the intermediate and 1.86 in the high-alcohol/low-methionine and -folate level).
|Low alcohol, high methionine, high folate2||Intermediate||High alcohol-methionine-low folate||χ21 trend|
|Colorectal cancer, cases:controls||44:173||1,805:3,803||104:178|
|OR (95% CI)||1 (referent)||1.66 (1.17–2.35)||1.83 (1.19–2.81)||5.91 (p = 0.01)|
|Colon cancer, cases:controls||31:173||1,129:3,803||65:178|
|OR (95% CI)||1 (referent)||1.47 (0.89–2.19)||1.67 (1.02–2.75)||3.57 (p = 0.06)|
|OR (95% CI)||1 (referent)||2.11 (1.18–3.77)||2.12 (1.07–4.20)||2.86 (p = 0.09)|
The present findings confirm an inverse association between folate intake and colorectal cancer risk, which appears to be consistent across strata of alcohol drinking and methionine intake, and that the combination of low folate and methionine and high alcohol intake leads to an >80% excess risk. This is of particular interest, given the widespread, regular and relatively high alcohol consumption in the Italian population.13, 17 In our study, methionine was inversely related to colorectal cancer risk (OR for the highest quintile = 0.77, 95% confidence interval 0.61–0.96), but there was no consistent association between measures of alcohol intake and colorectal cancer risk.17
The inverse association with folate was seen across strata of sex and subsite and was apparently stronger for the rectum. Previous studies have reported stronger associations for the colon,10, 18, 19 while no clear difference was observed in others.9 Other B vitamins involved in the folate pathway also showed an inverse relation with colorectal cancer risk in this data set. Thus, the OR was 0.53 in the highest quintile of vitamin B6 intake.12
There is therefore a consistent body of evidence indicating that folate may have favorable effects on colorectal carcinogenesis. Some differences between studies, in subgroup analyses or the interaction with alcohol, methionine or other variables, may be related to different sources of folate in various populations, including supplementation, which is uncommon in Italy,20 or to random variation in small subgroups. Furthermore, there are uncertainties in dietary assessment of micronutrients by interview and in the validity of measurement of specific micronutrients, though the questionnaire utilized in our study has proven to be satisfactorily reliable21 and valid.15 Additional strengths of our study are its large data set, the similar catchment areas of cases and controls, the practically complete participation and the similar interview setting for cases and controls, which may reduce information bias.22 Further, allowance for major identified potential confounding factors did not materially modify any of our results.
Together with the findings for breast cancer,5 the present results add supportive evidence to the role of folate fortification in cancer prevention in the Italian population.