NUTRITION AND CANCER, 50(2), 111119 Copyright 2004, Lawrence Erlbaum Associates, Inc.

REPORTS

Alcohol Consumption and Risk of Colon Cancer: Evidence From the National Health and Nutrition Examination Survey I Epidemiologic Follow-Up Study
Lihchyun Joseph Su and Lenore Arab

Abstract: The epidemiologic findings on the relationship between alcohol consumption and colon cancer are inconsistent. The National Health and Nutrition Examination Survey (NHANES) I Epidemiologic Follow-Up Study (NHEFS) included a prospective cohort population representative of the general U.S. population, which had not been fully utilized for examining the risk between colon cancer and alcohol drinking. The NHEFS consisted of 10,220 participants prospectively followed over a decade. Alcohol consumption, amount and type of beverage, and drinking patterns at baseline were considered in examination of the effect of alcohol consumption on the risk of colon cancer. The consumption of one or more alcoholic beverages a day at baseline was associated with approximately a 70% greater risk of colon cancer [relative risk (RR) = 1.69; 95% confidence interval (CI) = 1.03, 2.79], with a strong positive dose-response relationship (P = 0.04). This association appeared to be exclusively related to daily drinking of one or more drinks of liquor (RR = 2.48; 95% CI = 1.66, 4.53). Additionally, more than a 70% increased risk of colon cancer was observed for more than 34 yr of alcohol drinking history compared with nondrinkers (RR = 1.73; 95% CI = 1.08, 2.78). Overall, alcohol consumption was significantly associated with increased risk of colon cancer. The most important factor for colon cancer seems to be liquor consumption.

Introduction Colon cancer is the third leading cause of cancer death and the third most common type of cancer for both men and women in the United States (1). The American Cancer Society estimates that there will be 106,370 new cases (50,400

men and 55,970 women) of colon cancer diagnosed in 2004. Moreover, it is estimated that 56,730 individuals (28,320 men and 28,410 women) will die of colon cancer during that period. Many factors that increase the risk of colon cancer have been identified. Risk factors include family history of colon cancer, personal history of intestinal polyps, history of chronic inflammatory bowel disease, aging, a diet mostly from animal sources, physical inactivity, obesity, and smoking (217). Many studies have evaluated the role of alcohol in colon cancer. The review conducted by Kune and Vitetta gathered the results of 52 major scientific studies from 1957 to 1991 and revealed that a high proportion of studies showed a positive association between colon cancer and alcohol drinking (18). However, the results from epidemiological studies showed no consistent association between alcohol consumption and risk of colon cancer (1934). A prospective cohort study in Japan reported a positive dose-response relationship between alcohol consumption and colon cancer risk in both men and women followed for 8 yr (35). On the other hand, a Danish population-based cohort study showed no association between alcohol consumption and colon cancer risk; in addition, this study did not find any type of alcoholic beverage to be associated with colon cancer (36). Most recently, Cho et al., who pooled eight cohort studies from North America and Europe, showed a significant trend between increased amount of alcohol intake and the risk of colon cancer (37). Consumption of 45 g per day or more of alcohol was associated with a 45% increased risk of colon cancer (1.141.83). Several plausible mechanisms by which alcohol might influence carcinogenesis have been proposed over the past decades, including: 1) alcohol enhances gastrointestinal cell regeneration (38), 2) metabolites of ethanol enhance

L. Joseph Su is affiliated with the Stanley S. Scott Cancer Center and School of Public Health, Louisiana State University Health Sciences Center, New Orleans, LA 70112. L. Arab is affiliated with the Department of Epidemiology and Nutrition, CB#7400, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599.

tumorigenicity (39,40), 3) alcohol depresses immune responses (41), and 4) in the absence of adequate folic acid, ethanol enhances DNA methylation (4251) and alcohol suppresses folate resorption (4648). In an attempt to further the research of alcohol and colon cancer risk, we studied a sampling of both men and women in a larger and diverse representative sample of the general U.S. population with a wide range of age, ethnicity, and socioeconomic status, using the data from the First National Health and Nutrition Examination Survey (NHANES I) Epidemiologic Follow-up Study (NHEFS). This study examined the effects of alcohol consumption on the risk of colon cancer and analyzed the individual role of different types of alcoholic beverages and their effects on colon cancer. We also used a detailed alcohol consumption and drinking behavior history at a follow-up interview to assess changes in drinking patterns on colon cancer risk.

strong etiologic factor for cancer, were administered to only a portion of the 14,407 subjects in the NHEFS cohort. The alcohol drinking behavior questions were not requested of all subjects in the NHANES I. Consequently, only subjects interviewed at the first follow-up of the NHEFS in 198284 were included in this analysis, and the information obtained in the interview was treated as the baseline for the 10-yr follow-up (198284 to 1993). Subjects (n = 149) who developed cancer of any type other than skin cancer before the first follow-up interview were excluded from the analysis. Ten subjects who were pregnant during the interview period were also excluded because drinking behaviors very likely were changed as a result of health recommendations. Additionally, 1,610 subjects who died during the first follow-up interview period between 1982 and 1984 were also excluded. Finally, 33 subjects who did not report answers to the alcohol drinking questions were also excluded from the study. This left 10,418 subjects who were included in the final analyses.

Methods Detailed descriptions of the NHEFS have been published previously (5256). Briefly, the NHEFS cohort was established based on the NHANES I conducted by the National Center for Health Statistics, Centers for Disease Control and Prevention, from 1971 through 1975. A U.S. national probability sample of 20,729 noninstitutionalized persons aged 2574 yr was initially enrolled in the survey. Population subgroups thought to be at nutritional risk were oversampled. Three waves of follow-ups for the 14,407 subjects who underwent the complete medical examination were conducted in the NHEFS during the years 198284, 198687, and 1992. The first wave of data collection was conducted for all members of the NHEFS cohort from 1982 through 1984. The 1986 NHEFS was conducted for members of the NHEFS cohort who were 5574 yr of age at their baseline examination and not known to be deceased at the 198284 NHEFS (n = 3,980). The 1987 NHEFS was conducted for the entire nondeceased NHEFS cohort (n = 11,750) with the same questionnaire as the 1986 survey. The third wave of data collection, the 1992 NHEFS, included the entire nondeceased NHEFS cohort (n = 11,195). Subjects who had died were followed up by proxy interview. Of the 14,407 participants eligible for follow-up, 13,291 (92.2%) were successfully traced through 1992. Death certificates were coded according to the International Classification of Diseases, Ninth Revision (ICD-9). Study Population The first follow-up for the NHEFS was conducted in 198284. A total of 12,220 individuals were interviewed with an extensive questionnaire, including detailed information regarding drinking and smoking habits, which were not obtained for all subjects in the NHANES I. The 93-item food frequency questionnaire was also administered in this follow-up interview (57). The questions on smoking habits, a 112 Colon Cancer Both self-reported colon cancer incidence and colon cancer mortality report were examined to make sure that the analysis captured all colon cancer cases in the NHEFS. Cancer incidence was self-reported in four follow-up interviews. The ICD-9 code 153 was used to identify colon cancer morbidity as the first cancer incidence other than skin cancer. Subjects who reported having had other cancers previously were excluded from the analysis. If subjects reported multiple cancers in the same year and colon cancer was among the cancers, colon cancer was considered as the primary cancer incidence. Cancer death was ascertained through proxy interview and/or by using ICD-9 codes from National Death Index and hospital record diagnostic codes in NHEFS mortality and health care facility stay databases. The cause of death was validated by referring to the underlying cause of death listed on the death certificate. Similar to the cancer incidence, colon cancer mortality was extracted for subjects whose underlying cause of death was listed as ICD-9 code 153. Study subjects were determined as colon cancer cases primarily from the self-reported colon cancer incidence. However, 27 subjects died of colon cancer during the follow-up period, although they had never reported colon cancer incidence. These subjects were included as colon cancer cases in the analysis. Because they died of colon cancer and the time of death was between two follow-up interviews, we designated these subjects as colon cancer cases, and the year of death was identified as the year of cancer incidence. The follow-up time of cases was calculated by subtracting the month and year of the baseline interview from the midyear of the incidence of colon cancer. Similarly, for subjects lost, the follow-up time was estimated for the total number of months between baseline interview and the time of censoring. The censoring time was either the time of dropout (loss to follow-up or cancer of other sites) or the end of the study in July 1993. Nutrition and Cancer 2004

Alcohol Consumption Systematic questions regarding drinking behavior over the past 12 months, including frequency and amount of different types of alcoholic beverages, were used in the interview conducted in 198284. Nondrinkers were defined as those who reported fewer than 12 alcoholic drinks of any kind in the past 12 months. Subjects were asked how often they drank alcoholic beverages per the number of days in a week or a month. Subjects were also asked how many drinks they usually had on the days the drinking occurred. A quantity frequency measure of drinking was used. This measure was calculated by multiplying the frequency of drinking occasion, that is, drink days per week, and the average amount of consumption per occasion, that is, drinks per day, to yield the total number of drinks consumed each day. Separate measures were calculated for beer, wine, and liquor consumption. Extensive questions on drinking patterns at ages 25, 35, 45, 55, 65, and 75 yr were asked at the interview in 198284. The series of questions provided an opportunity to identify the age at which subjects began drinking and the estimates of duration. For example, if subjects reported nondrinking at age 25 yr but drinking at age 35 yr, the age they started drinking was coded as 35 yr. Length (in years) of drinking was also generated based on this set of questions. The length of continuous drinking was defined as the baseline age of the subject minus the age the subject started and continued drinking throughout the decades. Additionally, drinking patterns were determined based on reported alcohol drinking at baseline (198284) and the drinking behavior in the previous decade, based on the series of drinking pattern questionnaires. Five drinking patterns (abstainers, casual drinkers, initiators, quitters, and drinkers) were included simultaneously in the model (58). If a subject reported no drinking at baseline and the follow-ups, the subject was considered an abstainer. If a subject reported drinking fewer than 3.5 drinks per week (one drink every other day) at both periods of time, the subject was considered a casual drinker. If a subject reported no drinking initially but drank at the later period, the subject was categorized as an initiator. Subjects were considered quitters if they drank at baseline but not the later period. Drinkers were those who reported more than 3.5 drinks per week at both baseline and the follow-up periods. The information was then modeled using the abstainers as the reference group to examine whether a particular pattern of drinking was associated with elevated colon cancer risk.

self-reported height and weight data. Subjects were asked at the 198284 and 1992 interviews if they had first-degree family members (parents, siblings, children) who had been diagnosed with cancer. The information from both interviews was combined to create the variable of family colon cancer history. Dietary behavior information was abstracted from the semiquantitative food frequency questionnaire administered in 198284, and the amounts of fruit, vegetable, meat, poultry, and grain products were calculated. Statistical Analyses Cox proportional hazard models were used to estimate the effect of alcohol consumption on risk of colon cancer, adjusting for age, race, sex, and education as recommended by the National Center for Health Statistics (59). Additional potential confounding effects of smoking status; nonpoultry meat consumption; poultry and seafood consumption; regular multivitamin use; history of colonic polyps; BMI; aspirin use; physical activity; dietary fiber; dietary nutrients, such as vitamins C and E, carotenoids, and folate; fruit and vegetable consumption; and family history of colon cancer were considered for their impact on the relationship between alcohol consumption and colon cancer risk. Each of the covariates was modeled individually with the alcohol consumption variable on colon cancer risk to determine whether the specific covariate expressed a confounding effect or was an effect modifier. If the covariate changed the risk estimate by more than 10%, the covariate was later modeled simultaneously with other covariates in the final model. Variables included in the final models were: age, race, sex, education, smoking status, nonpoultry meat consumption, poultry and seafood consumption, regular multivitamin use, history of colonic polyps, and BMI. Alcohol consumption data, dichotomously (no/yes), continuously (drinks per week), and categorically (nondrinker plus three levels of drinking) were used to model the risk of colon cancer. The Wald test was used to test the linear relationship between the risk of colon cancer and the categorical alcohol consumption levels. Nonparametric analyses were used to examine the relationship between alcohol drinking patterns and colon cancer risk. Statistical models on the relationship between alcohol consumption and colon cancer risk have also been examined for a possible interactive effect from sex. The results suggest that sex does not play an interactive role with alcohol consumption in colon cancer risk. Therefore, combined findings from both men and women adjusting for sex are used to demonstrate the results of this study.

Other Covariates Sociodemographic information obtained from the survey (age, education, race, sex, and poverty index) was included as a source of possible confounders or effect modifiers. Detailed systematic questions regarding smoking behaviors and history were also solicited on all subjects at the 198284 interview because the information was available only on approximately one third of the cohort in the NHANES I survey. Body mass index (BMI; kg/m2) was calculated based on the Vol. 50, No. 2 Results Demographic characteristics of this NHEFS population stratified by sex are shown in Table 1. This study population consisted of 3,887 men and 6,531 women. In general, the mean age was higher in men than in women (t test P value < 0.01). More men drank alcohol and smoked than women in the cohort. Family histories of colon cancer and aspirin use 113

Table 1. Demographic Characteristics of the NHEFS Population by Sex

Male (n = 3,887) Colon cancer cases Age BMI (kg/m2) Education High school or less High school graduate or less More than high school graduate Race Black White Others Family colon cancer history Polyps or tumor of the colon Current smoker Aspirin use Frequency of meat intake/week Servings of fruits and vegetables/week Alcohol consumption 198284 drinkers Beer drinking (weekly mean/SD) Wine drinking (weekly mean/SD) Liquor drinkers (weekly mean/SD) 55 (1.4%) 58.5 14.7 26.5 4.1 1,612 (41.5%) 1,167 (30.0%) 1,108 (28.5%) 434 (11.2%) 3,357 (86.3%) 96 (2.5%) 181 (4.7%) 127 (3.3%) 1,105 (28.4%) 839 (21.6%) 11.6 6.1 6.4 14.9 2,239 (57.6%) 3.87/9.39 0.81/3.04 2.20/6.91 Female (n = 6,531) 56 (0.9%) 56.1 14.8 25.9 5.5 2,517 (38.5%) 2,574 (39.4%) 1,440 (22.1%) 927 (14.2%) 5,470 (83.8%) 134 (2.0%) 391 (6.0%) 201 (3.1%) 1,656 (25.4%) 1,689 (25.9%) 9.2 5.0 6.3 14.1 2,482 (38.0%) 0.68/3.17 0.63/2.10 0.76/2.91

Table 2. Alcohol Consumption and Colon Cancer Risk in the NHEFS

Cases Alcoholic beverages No Yes Nondrinkers <1 Drink/day 1 Drink/day Trend P value Beer consumption Nondrinkers <1 drink/day 1 Drink/day Trend P value Wine consumption Nondrinkers <1 Drink/day 1 Drink/day Trend P value Liquor consumption Nondrinkers <1 Drink/day 1 Drink/day Trend P value Person-Month Minimally Adjusteda (95% CI) Fully Adjustedb (95% CI)

63 48 63 22 26

45,734 40,177 45,734 23,301 16,587

1.00 1.26 (0.85, 1.89) 1.00 1.04 (0.63, 1.71) 1.67 (1.03, 2.70) 0.04 1.00 1.03 (0.65, 1.64) 1.10 (0.51, 2.35) 0.82 1.00 1.02 (0.66, 1.57) 0.83 (0.26, 2.63) 0.97 1.00 1.44 (0.93, 2.23) 2.28 (1.28, 4.06) 0.01

1.00 1.29 (0.86, 1.95) 1.00 1.08 (0.65, 1.79) 1.69 (1.03, 2.79) 0.04 1.00 1.04 (0.65, 1.66) 1.09 (0.51, 2.34) 0.84 1.00 1.04 (0.67, 1.61) 0.78 (0.24, 2.49) 0.67 1.00 1.48 (0.95, 2.31) 2.48 (1.66, 4.53) <0.01

76 27 8

53,672 25,639 6,601

77 31 3

51,292 31,585 3,035

63 33 15

50,154 30,339 5,419

a: Model adjusted for age, sex, and race. b: Models adjusted for age, sex, race, body mass index, educational level, nonpoultry meat consumption (log transformed), poultry and seafood consumption (log transformed), regular multivitamin use, history of colonic polyps, and smoking status.

were observed more frequently in women than in men (2 tests P values < 0.01). Within this population, 111 incident colon cancer cases occurred between the surveys of 198284 and the end of follow-up in 1993. Table 2 summarizes the associations between alcohol consumption and risk of colon cancer. After adjusting for age 114

at baseline, sex (male/female), race (black/white/others), education level (< high school graduate, high school graduate, > high school graduate), nonpoultry meat, poultry, and seafood consumption (log-transformed), regular multivitamin and mineral use, history of colonic polyps, and smoking status (no/yes), there was a nonsignificant 29% increased risk of Nutrition and Cancer 2004

colon cancer when comparing drinkers with nondrinkers [relative risk (RR) = 1.29; 95% confidence interval (CI) = 0.86, 1.95]. When the level of alcoholic beverage consumption was broken down into three levels (nondrinkers, drinkers but less than one drink per day, and drinkers of one or more drinks per day), the association with colon cancer risk was not statistically significant for drinkers who consumed less than one alcoholic beverage per day compared with nondrinkers. Alcohol consumption was significantly associated with colon cancer for drinkers who drank more than one drink a day (RR = 1.69; 95% CI = 1.03, 2.79) when compared with nondrinkers. The test for trend showed a significant dose-response relationship (P = 0.04) between increased alcohol consumption and colon cancer risk. As for the types of alcoholic beverage consumed, beer and wine consumption were not associated with the risk of colon cancer. In fact, the direction of association with wine was inversely related to colon cancer risk. Liquor consumption, however, was solely responsible for the significant association with colon cancer. Subjects who drank one or more drinks of liquor per day have more than a twofold risk of colon cancer (RR = 2.48; 95% CI = 1.66, 4.53) compared with nonliquor drinkers. A highly significant dose-response relationship is also seen for the increased liquor consumption (P < 0.01).

The initiation age that drinking started and the years of continuous drinking were related to colon cancer are presented in Table 3. Based on the tertile categorization of years of continuous alcohol drinking, subjects who continually drank for more than 34 yr have a 73% increased risk of colon cancer compared with nondrinkers (RR = 1.73; 95% CI = 1.08, 2.78). Those with a mean of 25 yr of drinking history had a nonsignificant 32% increased colon cancer risk. No increase was noted in those with less than a mean of 9 yr of drinking. Subjects who started drinking in their 30s had a significant 95% elevated colon cancer risk. Subjects who initiated drinking in their 20s and 40s had nonsignificant elevated colon cancer risks (RRs = 1.35 and 1.44, respectively). When these models were further adjusted for the type and amount of alcoholic beverage consumed, the point estimate did not change more than 10% so the more parsimonious models are used. This study takes advantage of the detailed information solicited at the baseline interview on alcohol consumption over the lifetime of subjects. The assessment of alcohol consumption based on approximately 5 yr prior to the starting time of the cohort survey was estimated and the drinking pattern at two time intervals was assigned (Table 4). Subjects who drank alcoholic beverages casually at both time points did not seem to have significantly elevated risk for colon cancer

Table 3. Length of Regular Alcohol Consumption and Colon Cancer Risk in the NHEFS
Cases Years of drinking 0 017 yr 1734 yr >34 yr Trend P value Age starts drinking Nondrinkers 45 yr old 35 yr old 25 yr old Person-Month Minimally Adjusteda (95% CI) Fully Adjustedb (95% CI)

52 3 17 39

551,276 176,294 188,109 115,260

1.00 0.71 (0.20, 2.46) 1.60 (0.87, 2.95) 1.66 (1.05, 2.63) 0.02 1.00 1.49 (0.75, 2.97) 2.05 (1.12, 3.77) 1.40 (0.87, 2.27)

1.00 0.65 (0.19, 2.25) 1.32 (0.73, 2.39) 1.73 (1.08, 2.78) 0.02 1.00 1.44 (0.72, 2.87) 1.95 (1.05, 3.60) 1.35 (0.83, 2.20)

52 10 14 35

550,136 55,537 93,496 331,770

a: Model adjusted for age, sex, and race. b: Models adjusted for age, sex, race, body mass index, educational level, nonpoultry meat consumption (log transformed), poultry and seafood consumption (log transformed), regular multivitamin use, history of colonic polyps, and smoking status.

Table 4. Alcohol Consumption Patterns and Colon Cancer Risk in the NHEFS
Cases Drinking patterns Abstainer Casual drinker Initiator Quitter Drinker Person-Month Minimally Adjusted (95% CI)
a

Fully Adjusted (95% CI)

55 10 23 7 16

484,454 108,832 254,598 59,223 102,239

1.00 1.14 (0.58, 2.23) 1.52 (0.90, 2.56) 1.13 (0.51, 2.47) 1.89 (1.06, 3.37)

1.00 1.14 (0.58, 2.22) 1.50 (0.89, 2.53) 1.10 (0.50, 2.41) 1.80 (1.00, 3.23)

a: Models adjusted for age, sex, and race. b: Models adjusted for age, sex, race, body mass index, educational level, nonpoultry meat consumption (log transformed), poultry and seafood consumption (log transformed), regular multivitamin use, history of colonic polyps, and smoking status.

Vol. 50, No. 2

115

when compared with nondrinkers. Nonstatistically significant elevated colon cancer risks were also observed for initiators and quitters (RRs = 1.50 and 1.10, respectively). The association increased 80% for those who consistently drank at least one alcoholic beverage every other day (RR = 1.80; 95% CI = 1.00, 3.23; P = 0.048).

Discussion Previous studies have been inconsistent in reporting the relationship between alcohol and colon cancer (18). The results of this study show an association between alcohol consumption and an increased risk of colon cancer in the U.S. population. The consumption of one drink of alcoholic beverage a day or more at baseline is associated with approximately a 70% greater risk of colon cancer. There is a strong positive dose-response relationship between the amount of alcohol consumed and the risk of colon cancer. This effect was found to be almost exclusively due to liquor. Daily drinking of one or more drinks of liquor is found to be highly significant and associated with more than a twofold risk of colon cancer. No association with wine was noted. This may explain the inconsistencies in other reports, which did not segregate alcohol sources. Examination of the pattern of alcohol consumption yields interesting relationships. Subjects who drank one drink or more of alcoholic beverages every other day at the baseline period and for 5 yr prior had a significantly increased risk of colon cancer compared with abstainers. However, persons who report drinking 5 yr earlier and who later quit, initiated drinking behavior at the baseline, or drank relatively lightly only had nonsignificantly elevated risks of colon cancer compared with nondrinkers. This is similar to the relationship found between cigarette smoking, quitting, and lung cancer risk (60). The years of drinking history are associated with elevated colon cancer risk. More than 70% increased risk of colon cancer is observed when there are more than 34 yr of alcohol drinking history. Some studies suggest that the type of alcohol (e.g., beer, wine, and liquor) may be more important to carcinogenesis than the total alcohol consumed (18,31). Other studies have observed an increased risk of colon cancer with alcohol consumption, regardless of the types of alcohol (61). In this study, we found a strong positive association between liquor consumption and colon cancer risk. Unlike the other prospective study by Goldbohm et al. (31,6264), the association with colon cancer was moderately elevated with beer consumption and even inversely associated with wine drinking. Polyphenols from tea and wine have been considered a chemoprevention agent against colon cancer. Polyphenols are powerful antioxidants and free radical scavengers (65). They have anti-inflammatory properties (66) and may modulate the activity of phase I and II enzymes, in particular glutathione-related enzymes (67). The association between tea consumption, another rich source of polyphenols, and reduced colon cancer risk has been reported in this study popu116

lation (68). Wines, in particular red wine, also contain a large array of polyphenolic constituents that have been shown to block carcinogenesis and to inhibit the growth of tumors in animals and in cell culture by altering the activity of certain enzymes or the expression of specific genes (69). The observed nonsignificantly reduced risk of colon cancer from wine drinking may be the result of interaction between these polyphenols. Change in drinking patterns has not been previously examined in relationship to colon cancer risk. Whether subjects are abstainers, constant drinkers, or sporadic drinkers may impact colonic carcinogenesis. This study assessed habitual alcohol use at baseline and 5 yr prior to the baseline to construct a model to examine the change of drinking patterns on colon cancer risk. We found that constant drinkers who drank on average one drink every other day for the past 5 yr before the baseline and continued the drinking pattern had significantly higher risk of colon cancer than abstainers. Those who drank earlier but quit drinking had a risk similar to that of casual drinkers. Nonsignificantly higher risk was observed in subjects who initiated the drinking behavior at the baseline compared with quitters and casual drinkers. This result suggests that exposure to alcohol within the recent decade may pose a higher risk for carcinogenesis than exposure at distant past. Alcohol exposure may act more as a promoter for carcinogenesis rather than an initiator, considering the latent period of cancer development. The risks observed among initiators, casual drinkers, and quitters are not statistically significant, so the findings should be interpreted with caution. However, the tendency mimics a model of waxing and waning risk as a function of exposure. The relationship between alcohol exposure and colon cancer is biologically plausible. Stimulation of cell regeneration is known to enhance carcinogenesis. Mucosal proliferative status induced by long-term alcohol consumption by itself has been proposed as an important contribution to carcinogenesis (38). An imbalance between cell proliferation and apoptosis is a prominent feature that distinguishes cancer from normal tissue. Actively proliferating cells are more susceptible to initiators of carcinogenesis (primary carcinogens) and genetic alterations. As cells migrate from lower to upper crypts, the number of proliferative cells decreases. The cells no longer divide and become differentiated when they reach the upper crypt region. However, this sequence of events is disorderly during the evolution of a neoplastic lesion in the colon. Enhancement of colonic tumor growth is often associated with epithelial hyper-regeneration. Studies investigating the influence of chronic ethanol consumption on mucosal renewal found a marked elevation of cell production rate in the distal colon in rats fed alcohol for 4 wk compared with the rats not exposed to alcohol (70,71). This increased cell production was accompanied by an extension of the proliferative compartment and reduced life span of functional epithelial cells. It appears that chronic consumption of alcohol, in particular liquor (a strong stimulus), could result in mucosal hyper-regeneration in the colorectal cells, which is associated with an expansion of the proliferative compartment, thus leading to an inNutrition and Cancer 2004

creased risk of colorectal cancer. As for the mechanism underlying an association between alcohol consumption and the hyper-regeneration, ethanol and its metabolite, acetaldehyde, are thought to play major roles in the injuries to the mucosa. Acetaldehyde adducts can interfere with normal cellular functions such as the inhibition of a human DNA-repair protein, O6-methylguanine transferase (72). Alcohol consumption has also been associated with increased bile acid excretion, increased bile acid recirculation, increased production of secondary bile acids, reduced bile cholesterol saturation, and elevation of high-density lipoprotein cholesterol levels (18). Bile acid concentrations in colon and rectum are considered risk factors for colorectal carcinogenesis (73). An increased synthesis rate of bile acid has been observed both after short-term ingestion of alcohol in humans (74) and after long-term alcohol exposure in rats (75). Although an increased synthesis of bile acids may be secondary to increased fecal losses of bile acids, a primary stimulatory effect of alcohol on bile acid synthesis after short-term alcohol use has also been suggested (76). Immune depression can enhance tumor growth (18) and, in turn, can be induced through alcohol consumption (43). The immune system may be suppressed by alcoholic liver damage or malnutrition; excessive alcohol use alone may damage the immune system (41). Heavy consumption of alcohol may alter the production and turnover rates of T and B lymphocytes, with a resultant shift in the relative concentration of the lymphocytes. Alcoholic subjects have been found to have significantly lower circulating T lymphocyte counts than nonalcoholics and reduced ability for the cells to undergo blastic transformation following mitogenic stimulation (77). In addition, nonspecific activation of B lymphocytes occurs in patients who drink alcohol regularly in excess (78). Individuals who drink alcohol may have decreased intake and/or decreased bioavailability of other nutrients, such as folate, that may prevent cancer. In addition to the hypothesized mechanisms for alcohol consumption and colonic carcinogenesis described so far, long-term ethanol (the immediate metabolite of alcohol) feeding of rats has been observed to cause significant reduction of hepatic methyltetrahydrofolate/homocysteine methyltransferase activity with compensatory increase of betaine/homocysteine methyltransferase activity (51,7981). Ethanol-induced alterations in folate and methionine metabolism could be relevant to carcinogenesis through decreased availability of methyl group for DNA methylation or through altered regulation of DNA precursor metabolism (51,82). In conclusion, this studys findings were consistent with those reviewed by Kune and Vitetta, Longnecker et al., and the pooled analysis conducted by Cho et al., that alcohol consumption is positively associated with colon cancer (18,32,37,83). With as little as one drink per day, the results of this study show an approximately 70% increased risk of colon cancer compared with nondrinkers. Liquor drinking appears to account for the majority of this. Based on the information obtained in this population, 19.0% of the surveyed population drank alcoholic beverages and 6.4% drank liquor, Vol. 50, No. 2

one or more drinks at baseline; this would translate to population attributable risks of 0.15 and 0.10 for alcoholic beverage and liquor drinking, respectively. Considering the mechanisms of action, the consistency of findings, the magnitude of the effects, and the dose effect of alcohol consumption, with a probability-sampled cohort representing the noninstitutionalized U.S. population prospectively followed for 20 yr, alcohol consumption can be considered a significant risk factor for colon cancer incidence in the United States.

Acknowledgments and Notes

We thank Christine Cox and Dr. Deborah Ingram at the National Center for Health Statistics for clarifying information regarding the NHEFS, Walter Davis at the University of North Carolina (UNC-CH) for coordinating and preparing raw data, Dr. Robert Millikan at the Epidemiology Department at UNC-CH for helpful suggestions, and Susan Watson and Soowon Kim at the Nutrition Department at UNC-CH for data coding. Address correspondence to L. Joseph Su, School of Public Health, Epidemiology Program, Louisiana State University Health Sciences Center, 1600 Canal Street, Suite 800, New Orleans, LA 70112. E-mail: [email protected]. Submitted 9 September 2003; accepted in final form 20 September 2004.

References
1. Jemal A, Tiwari RC, Murray T, Ghafoor A, Samuels A, et al.: Cancer statistics, 2004. CA Cancer J Clin 54, 829, 2004. 2. American Cancer Society: Cancer Facts and Figures 2002. Atlanta, GA: American Cancer Society, 2002. 3. Moyad MA: Is obesity a risk factor for prostate cancer, and does it even matter? A hypothesis and different perspective. Urology 59, 4150, 2002. 4. Willett WC: Diet and cancer: one view at the start of the millennium. Cancer Epidemiol Biomarkers Prev 10, 38, 2001. 5. Platz EA, Willett WC, Colditz GA, Rimm EB, Spiegelman D, et al.: Proportion of colon cancer risk that might be preventable in a cohort of middle-aged US men. Cancer Causes Control 11, 579588, 2000. 6. Weber RV, Stein DE, Scholes J, and Kral JG: Obesity potentiates AOM-induced colon cancer. Dig Dis Sci 45, 890895, 2000. 7. Giacosa A, Franceschi S, La Vecchia C, Favero A, and Andreatta R: Energy intake, overweight, physical exercise and colorectal cancer risk. Eur J Cancer Prev 8 Suppl 1, S5360, 1999. 8. Sesink AL, Termont DS, Kleibeuker JH, and Van der Meer R: Red meat and colon cancer: the cytotoxic and hyperproliferative effects of dietary heme. Cancer Res 59, 57045709, 1999. 9. Slattery ML, Edwards SL, Boucher KM, Anderson K, and Caan BJ: Lifestyle and colon cancer: an assessment of factors associated with risk. Am J Epidemiol 150, 869877, 1999. 10. Zhuo XG and Watanabe S: Factor analysis of digestive cancer mortality and food consumption in 65 Chinese counties. J Epidemiol 9, 275284, 1999. 11. Elitsur Y, Majumdar AP, Tureaud J, Dosescu J, Neace C, et al.: Tryosine kinase and ornithine decarboxylase activation in children with Helicobactor pylori gastritis. Life Sci 65, 13731380, 1999. 12. Ford ES: Body mass index and colon cancer in a national sample of adult US men and women. Am J Epidemiol 150, 390398, 1999. 13. Potter JD, Bigler J, Fosdick L, Bostick RM, Kampman E, et al.: Colorectal adenomatous and hyperplastic polyps: smoking and N-acetyltransferase 2 polymorphisms. Cancer Epidemiol Biomarkers Prev 8, 6975, 1999.

117

14. Slattery ML, Boucher KM, Caan BJ, Potter JD, and Ma KN: Eating patterns and risk of colon cancer. Am J Epidemiol 148, 416, 1998. 15. Oliveria SA and Christos PJ: The epidemiology of physical activity and cancer. Ann NY Acad Sci 833, 7990, 1997. 16. Bird CL, Frankl HD, Lee ER, and Haile RW: Obesity, weight gain, large weight changes, and adenomatous polyps of the left colon and rectum. Am J Epidemiol 147, 670680, 1998. 17. Bertone ER, Giovannucci EL, King NW Jr, Petto AJ, and Johnson LD: Family history as a risk factor for ulcerative colitis-associated colon cancer in cotton-top tamarin. Gastroenterology 114, 669674, 1998. 18. Kune G and Vitetta L: Alcohol consumption and the etiology of colorectal cancer: A review of the scientific evidence from 1957 to 1991. Nutr Cancer 18, 97111, 1992. 19. Giovannucci E, Rimm EB, Ascherio A, Stampfer MJ, Colditz GA, et al.: Alcohol, low- methioninelow-folate diets, and risk of colon cancer in men. J Natl Cancer Inst 87, 265273, 1995. 20. Goldbohm RA, Van den Brandt PA, Van t Veer P, Dorant E, Sturmans F, et al.: Prospective study on alcohol consumption and the risk of cancer of the colon and rectum in the Netherlands. Cancer Causes Control 5, 95104, 1994. 21. Gapstur SM, Potter JD, and Folsom AR: Alcohol consumption and colon and rectal cancer in postmenopausal women. Int J Epidemiol 23, 5057, 1994. 22. Gerhardsson de Verdier M, Romelsjo A, and Lundberg M: Alcohol and cancer of the colon and rectum. Eur J Cancer Prev 2, 401408, 1993. 23. Meyer F and White E: Alcohol and nutrients in relation to colon cancer in middle-aged adults. Am J Epidemiol 138, 225236, 1993. 24. Slattery ML, West DW, Robison LM, French TK, Ford MH, et al.: Tobacco, alcohol, coffee, and caffeine as risk factors for colon cancer in a low-risk population. Epidemiology 1, 141145, 1990. 25. Hirayama T: Association between alcohol consumption and cancer of the sigmoid colon: observations from a Japanese cohort study. Lancet 2, 725727, 1989. 26. Franceschi S and La Vecchia C: Alcohol and the risk of cancers of the stomach and colon-rectum. Dig Dis 12, 276289, 1994. 27. Boutron MC, Faivre J, Dop MC, Quipourt V, and Senesse P: Tobacco, alcohol, and colorectal tumors: a multistep process. Am J Epidemiol 141, 10381046, 1995. 28. Gapstur SM, Potter JD, and Folsom AR: Alcohol consumption and colon and rectal cancer in postmenopausal women. Int J Epidemio 23, 5057, 1994. 29. Gerhardsson de Verdier M, Romelsjo A, and Lundberg M: Alcohol and cancer of the colon and rectum. Eur J Cancer Prev 2, 401408, 1993. 30. Glynn SA, Albanes D, Pietinen P, Brown CC, Rautalahti M, et al.: Alcohol consumption and risk of colorectal cancer in a cohort of Finnish men. Cancer Causes Control 7, 214223, 1996. 31. Goldbohm RA, Van den Brandt PA, Van t Veer P, Dorant E, Sturmans F, et al.: Prospective study on alcohol consumption and the risk of cancer of the colon and rectum in the Netherlands. Cancer Causes Control 5, 95104, 1994. 32. Longnecker MP, Orza MJ, Adams ME, Vioque J, and Chalmers TC: A meta-analysis of alcoholic beverage consumption in relation to risk of colorectal cancer. Cancer Causes Control 1, 5968, 1990. 33. Longnecker MP: A case-control study of alcoholic beverage consumption in relation to risk of cancer of the right colon and rectum in men. Cancer Causes & Control 1, 514, 1990. 34. Yamada K, Araki S, Tamura M, Sakai I, Takahashi Y, et al.: Case-control study of colorectal carcinoma in situ and cancer in relation to cigarette smoking and alcohol use (Japan). Cancer Causes Control 8, 780785, 1997. 35. Shimizu N, Nagata C, Shimizu H, Kametani M, Takeyama N, et al.: Height, weight, and alcohol consumption in relation to the risk of colorectal cancer in Japan: a prospective study. Br J Cancer 88, 10381043, 2003. 36. Pedersen A, Johansen C, and Gronbaek M: Relations between amount and type of alcohol and colon and rectal cancer in a Danish population based cohort study. Gut 52, 861867, 2003.

37. Cho E, Smith-Warner SA, Ritz J, van den Brandt PA, Colditz GA, et al.: Alcohol intake and colorectal cancer: a pooled analysis of 8 cohort studies. Ann Intern Med 140, 603613, 2004. 38. Simanowski UA, Stickel F, Maier H, Gartner U, and Seitz HK: Effect of alcohol on gastrointestinal cell regeneration as a possible mechanism in alcohol-associated carcinogenesis. Alcohol 12, 111115, 1995. 39. Hakkak R, Korourian S, Ronis MJ, Ingelman-Sundberg M, and Badger TM: Effects of diet and ethanol on the expression and localization of cytochromes P450 2E1 and P450 2C7 in the colon of male rats. Biochem Pharmacol 51, 6169, 1996. 40. Salaspuro M: Bacteriocolonic pathway for ethanol oxidation: characteristics and implications. Ann Med 28, 195200, 1996. 41. Dunne F: Alcohol and the immune system. A causative agent in altering host defense mechanisms. Br Med J 298, 543544, 1989. 42. Hidiroglou N, Camilo ME, Beckenhauer HC, Tuma DJ, Barak AJ, et al.: Effect of chronic alcohol ingestion on hepatic folate distribution in the rat. Biochem Pharmacol 47, 15611566, 1994. 43. Muldoon RT and McMartin KE: Ethanol acutely impairs the renal conservation of 5-methyltetrahydrofolate in the isolated perfused rat kidney. Alcohol Clin Exp Res 18, 333339, 1994. 44. Barak AJ, Beckenhauer HC, Hidiroglou N, Camilo ME, Selhub J, et al.: The relationship of ethanol feeding to the methyl folate trap. Alcohol 10, 495497, 1993. 45. Ross DM and McMartin KE: Effect of ethanol on folate binding by isolated rat renal brush border membranes. Alcohol 13, 449454, 1996. 46. Cravo ML, Gloria LM, Selhub J, Nadeau MR, Camilo ME, et al.: Hyperhomocysteinemia in chronic alcoholism: correlation with folate, vitamin B-12, and vitamin B-6 status. Am J Clin Nutr 63, 220224, 1996. 47. Boutron-Ruault MC, Senesse P, Faivre J, Couillault C, and Belghiti C: Folate and alcohol intakes: related or independent roles in the adenoma-carcinoma sequence? Nutr Cancer 26, 337346, 1996. 48. Carreras O, Vazquez AL, Rubio JM, Delgado MJ, and Murillo ML: Comparative effects of intestinal absorption of folic acid and methyltetrahydrofolic acid in chronic ethanol-fed rats. Ann Nutr Metab 38, 221225, 1994. 49. Fernandez-Borrachero O, Rubio JM, Delgado MJ, Carreras Sanchez O, and Murillo ML: Folate absorption in the jejunum of chronic ethanol-fed rats: in vivo studies. Ann Nutr Metab 40, 277282, 1996. 50. Giovannucci E, Rimm EB, Ascherio A, Stampfer MJ, Colditz GA, et al.: Alcohol, low- methioninelow-folate diets, and risk of colon cancer in men. J Natl Cancer Inst 87, 26573, 1995. 51. Halsted CH, Villanueva J, Chandler CJ, Stabler SP, Allen RH, et al.: Ethanol feeding of micropigs alters methionine metabolism and increases hepatocellular apoptosis and proliferation. Hepatology 23, 497505, 1996. 52. National Center for Health Statistics: National Center for Health Statistics: Plan and operation of the Health and Nutrition Examination Survey, 19711973. Vital and Health Statistics. Rockville, MD: US Department of Health, Education, and Welfare, 1973. 53. National Center for Health Statistics: National Center for Health Statistics: Plan and operation of the NHANES I Epidemiologic Follow-up Study, 19821984. Vital and Health Statistics. Rockville, MD: US Department of Health and Human Services, 1987. 54. National Center for Health Statistics: National Center for Health Statistics: Plan and operation of the NHANES I Epidemiologic Follow-up Study, 1986. Vital and Health Statistics. Rockville, MD: US Department of Health and Human Services, 1990. 55. Madans JH, Cox CS, Kleinman JC, Makuc D, Feldman JJ, et al.: 10 years after NHANES I: mortality experience at initial followup, 198284. Public Health Rep 101, 474481, 1986. 56. Cornoni-Huntley J, Barbano HE, Brody JA, Cohen B, Feldman JJ, et al.: National health and nutrition examination Iepidemiologic follow-up survey. Public Health Rep 98, 245251, 1983. 57. Byrne C, Ursin G, and Ziegler RG: A comparison of food habit and food frequency data as predictors of breast cancer in the NHANES I/NHEFS cohort. J Nutr 126, 27572764, 1996.

118

Nutrition and Cancer 2004

58. Curtis AB, James SA, Strogatz DS, Raghunathan TE, and Harlow S: Alcohol consumption and changes in blood pressure among African Americans. The Pitt County Study. Am J Epidemiol 146, 727733, 1997. 59. National Center for Health Statistics: Plan and Operation of the NHANES I Epidemiologic Followup Study, 1992. Vital and Health Statistics, Series 1, No. 35. Washington, DC: U.S. Department of Health and Human Services, 1997. 60. Tong L, Spitz M, Fueger J, and Amos C: Lung carcinoma in former smokers. Cancer 78, 10041010, 1996. 61. Glynn S, Albanes D, Pietinen P, Brown C, Rautalahti M, et al.: Alcohol consumption and risk of colorectal cancer in a cohort of Finnish men. Cancer Causes Control 7, 214223, 1996. 62. Hsing AW, McLaughlin JK, Chow WH, Schuman LM, Co Chien HT, et al.: Risk factors for colorectal cancer in a prospective study among U.S. white men. Int J Cancer 77, 549553, 1998. 63. Glynn SA, Albanes D, Pietinen P, Brown CC, Rautalahti M, et al.: Alcohol consumption and risk of colorectal cancer in a cohort of Finnish men. Cancer Causes Control 7, 214223, 1996. 64. Todoroki I, Kono S, Shinchi K, Honjo S, Sakurai Y, et al.: Relationship of cigarette smoking, alcohol use, and dietary habits with sigmoid colon adenomas. Ann Epidemiol 5, 478483, 1995. 65. Rice-Evans CA, Miller NJ, Bolwell PG, Bramley PM, and Pridham JB: The relative antioxidant activities of plant- derived polyphenolic flavonoids. Free Radic Res 22, 375383, 1995. 66. Middleton E, Jr and Kandaswami C: Effects of flavonoids on immune and inflammatory cell functions. Biochem Pharmacol 43, 11671179, 1992. 67. Steele VE, Kelloff GJ, Balentine D, Boone CW, Mehta R, et al.: Comparative chemopreventive mechanisms of green tea, black tea and selected polyphenol extracts measured by in vitro bioassays. Carcinogenesis 21, 6367, 2000. 68. Su LJ and Arab L: Tea consumption and the reduced risk of colon cancer results from a national prospective cohort study. Public Health Nutr 5, 419425, 2002. 69. Soleas GJ, Diamandis EP, and Goldberg DM: Wine as a biological fluid: history, production, and role in disease prevention. J Clin Lab Anal 11, 287313, 1997.

70. Seitz H, Simanowski U, Garzon F, Ridcout J, Peters T, et al.: Possible role of acetaidehyde in ethanol- related rectal cocarcinogenesis in the rat. Gastroenterology 98, 406413, 1990. 71. Simanowski U, Suter P, Russell R, Heller M, Waldherr R, et al.: Enhancement of ethanol induced rectal mucosal hyperregeneration with age in F344 rats. Gut 35, 11021106, 1994. 72. Espina N, Lima V, Lieber C, and Garro A: In vitro and in vivo inhibitory effect of ethanol and acetaidehyde on O-methylguanine transferase. Carcinogenesis 9, 761766, 1988. 73. Cheah P: Hypotheses for the etiology of colorectal cancer - An overview. Nutr Cancer 8, 92105, 1990. 74. Axelson M, Mork B, and Sjovall J: Ethanol has an acute effect on bile acid biosvnthesis in man. FEBS Lett 281, 155159, 1991. 75. Vendemiale G and Lieber C: Acute and chronic effects of ethanol on biliarv secretion of bilirubin and bile acids. Subst Alcohol Actions/Misuse 5, 307317, 1985. 76. Ackehed G, Hedenborg G, Wisen O, and Norman A: Faecal bile acid excretion during detoxification in patients with alcohol abuse. Scand J Gastroenterol 31, 12051210, 1996. 77. Glassman A, Bennett C, and Randall C: Effects of ethyl alcohol on human peripheral lymphocytes. Arch Pathol Lab Med 109, 540542, 1985. 78. Drew P, Clifton P, LaBrooy J, and Sherman D: Polyclonal B cell activation in alcoholic patients with o evidence of liver dysfunction. Clin Exp Immunol 57, 479486, 1984. 79. Shaw S, Jayatilleke E, Herbert V, and Coleman N: Cleavage of folates during ethanol metabolism: role of xanthine oxidase-generated superoxide dismutase. Biochem J 257, 277280, 1989. 80. Barak AJ, Beckenhauer HC, and Tuma DJ: Effects of prolonged ethanol feeding on methionine metabolism in rat liver. Biochem Cell Biol 65, 230233, 1987. 81. Trimble KC, Molloy AM, Scott JM, and Weir DG: The effect of ethanol on one-carbon metabolism: increased methionine catabolism and lipotrope methyl-group wastage. Hepatology 18, 984989, 1993. 82. Hoffman RM: Unbalanced transmethylation and the perturbation of the diiferentiated state leading to cancer. Bioassays 12, 163166, 1990. 83. Bujanda L: The effects of alcohol consumption upon the gastrointestinal tract. Am J Gastroenterol 95, 33743382, 2000.

Vol. 50, No. 2

119