The authors of a recent major European study state that alcohol consumption may be associated with risk of colorectal cancer (CRC), but the epidemiological evidence for an association with specific anatomical subsites, types of alcoholic beverages and current vs. lifetime alcohol intake is inconsistent.
Within the European Prospective Investigation into Cancer and Nutrition (EPIC), 478,732 study subjects free of cancer at enrolment between 1992 and 2000 were followed up for an average of 6.2 years, during which 1,833 CRC cases were observed. Detailed information on consumption of alcoholic beverages at baseline (all cases) and during lifetime (1,447 CRC cases, 69% of the cohort) was collected from questionnaires. Cox proportional hazard models were used to examine the alcohol-CRC association. After adjustment for potential confounding factors, lifetime alcohol intake was significantly positively associated with CRC risk (hazard ratio, HR = 1.08, for 15 g/day increase), with higher cancer risks observed in the rectum (HR = 1.12) than distal colon (HR = 1.08), and proximal colon (HR = 1.02). Similar results were observed for baseline alcohol intake. When assessed by alcoholic beverages at baseline, the CRC risk for beer (HR = 1.38) was higher than wine (HR = 1.21), although the two risk estimates were not significantly different from each other. Higher HRs for baseline alcohol were observed for low levels of folate intake (1.13, for 15 g/day increase) compared with high folate intake (1.03, ). In this large European cohort, both lifetime and baseline alcohol consumption increase colon and rectum cancer risk, with more apparent risk increases for alcohol intakes greater than 30 g/day.
R. Curtis Ellison comments: Colo-rectal cancer is a very common type of cancer, with striking differences among populations. Because of its high occurrence, any putative risk factors, even those with a modest effect, could be important in terms of public health in many parts of the world.
This large, multi-country study confirmed very large differences in rates of CRC among countries. And some of the individual-country relations are interesting. For example, Greece had the second highest average lifetime intake of alcohol for men (30 g/d) and the lowest intake for women (5 g/d), and had by far the lowest age-standardized rates of CRC for both men and women. Italy, on the other hand, had slightly lower intake of alcohol for men, slightly higher intake for women, yet much higher incidence rates of CRC for both men (5 times higher) and women (twice as high) as did Greece.
Other interesting data are that while subjects reporting ≥ 60 g/day of alcohol consumed more than 500 more calories than non-drinkers, their non-alcohol calories were about the same and the body mass index (BMI) values were identical for the two groups for both men and women. Again, this raises questions about the metabolism of alcohol, and we do not yet understand what happens to those increased calories from alcohol.
For the associations between alcohol intake and cancer, the investigators combined data from men and women, although men consumed higher amounts of alcohol and showed higher age-standardized incidence rates of cancer than women. Sex-specific associations are not presented, although the authors state that results were similar for men and women for certain analyses. The investigators also combined ex-drinkers (none in last year) with never drinkers, yet state that they found no “appreciable changes in risks (data not shown)” between ex-drinkers and never drinkers. Beverage-specific effects were similar, with increases above 1.0 only for intakes greater than 20 g/d for both beer and wine. Intake of spirits/liquors was too low for meaningful associations.
Of particular importance, this (as have some other studies) showed a strong effect on cancer risk from folate intake among drinkers. The hazard ratios (HR) for CRC were 1.13 (1.06 - 1.20) for subjects in the lower third of folate intake for each center, 1.09 (1.03 - 1.15) for the middle third of folate intake, and 1.03 (0.98 - 1.09) for subjects in the upper third of folate intake, even though the mean alcohol intake was higher among those with higher folate intake (13.8, 18.5, and 28.3 g/day, for the tertiles of folate intake, respectively).
The overall impression from this analysis of a large dataset is that there are minimal effects on the risk of colo-rectal cancer of alcohol intake up to 30 g/d, with statistically significant increases generally only for consumers of 60 or more g/d of alcohol. Thus, individuals drinking within current recommendations for moderate use (≤24 g/day for men and ≤ 12 g/day for women in the United States) should have at most a minimal increase in CRC risk over non-drinkers. Larger amounts of alcohol appear to increase the risk of CRC.
Of great importance is the potential role of folate intake in attenuating the increase in the risk of CRC from alcohol intake, an attenuation in risk also shown for breast cancer in many studies. In this study, the hazard ratio among drinkers in the highest third of intake of folate was only 1.03, even though subjects in this group consumed more than twice as much alcohol as subjects in the lowest-folate group.
Source: “Lifetime and baseline alcohol intake and risk of colon and rectal cancers in the European prospective investigation into cancer and nutrition (EPIC) (p NA).” International Journal of Cancer Published Online: 19 Jul 2007 DOI: 10.1002/ijc.22966