Abstract

Background: The International Agency for Research on Cancer (IARC) concluded that alcohol consumption is related to colorectal cancer (CRC). However, several issues remain unresolved, including quantification of the association for light (≤1 drink/day) and moderate (2–3 drinks/day) alcohol drinking, investigation of the dose–response relationship, and potential heterogeneity of effects by sex, colorectal site, and geographical region.

Methods: Twenty-seven cohort and 34 case–control studies presenting results for at least three categories of alcohol intake were identified from a PubMed search of articles published before May 2010. The summary relative risks (RRs) were estimated by the random effects model. Second-order fractional polynomials and random effects meta-regression models were used for modeling the dose–risk relation.

Results: The RRs were 1.21 [95% confidence interval (CI) 1.13–1.28] for moderate and 1.52 (95% CI 1.27–1.81) for heavy (≥4 drinks/day) alcohol drinking. The RR for moderate drinkers, compared with non-/occasional drinkers, was stronger for men (RR = 1.24, 95% CI 1.13–1.37) than for women (RR = 1.08, 95% CI 1.03–1.13; Pheterogeneity = 0.02). For heavy drinkers, the association was stronger in Asian studies (RR = 1.81, 95% CI 1.33–2.46; Pheterogeneity = 0.04). The dose–risk analysis estimated RRs of 1.07 (95% CI 1.04–1.10), 1.38 (95% CI 1.28–1.50), and 1.82 (95% CI 1.41–2.35) for 10, 50, and 100 g/day of alcohol, respectively.

Conclusions: This meta-analysis provides strong evidence for an association between alcohol drinking of >1 drink/day and colorectal cancer risk.

introduction

Based on the World Health Organization estimates, there are about two billion people worldwide who consume alcoholic beverages regularly [1], with an average of 6.2 l of ethanol per adult per year [2]. Alcohol consumption is one of the most important known risk factors for human cancers [3], and potentially, one of the largest avoidable factors. In has been estimated that in 2002, 5.1% and 1.3% of all cancer deaths were attributable to alcohol drinking worldwide in men and women, respectively; the corresponding figures for incidence were 5.2% of all cancers in men and 1.7% of all cancers in women [4]. Intake of alcohol is causally related to cancers of the oral cavity, pharynx, larynx, esophagus, liver, female breast, and colorectum [5, 6].

A pooled analysis of eight cohort studies from North America and Europe found a modestly increased colorectal cancer risk (45% for colon and 49% for rectal cancers) with regular high alcohol intake (≥45 g/day), compared with nondrinkers, in men and women combined [7]. Another pooled study by Mizoue et al. [8] analyzed original data from five Japanese cohort studies [9–12] and found an increased risk for colorectal cancer among men and women who regularly drink ≥23 g/day of ethanol, compared with nondrinkers. There were also several meta-analyses, and quantitative overviews [13–17], all of which have supported a positive association between alcoholic beverages consumption and colorectal cancer risk. However, several issues remained unresolved. First, the dose–risk relation of alcohol intake with colorectal cancer risk has not yet been investigated in detail. In particular, a more precise quantification of the association for light and/or moderate alcohol consumption and the identification of a possible threshold of effect remain to be determined. Secondly, it is still uncertain whether the effect of alcohol varies across colon and rectal anatomical subsites. Some studies have reported a stronger alcohol–cancer risk association in the colon than in the rectum [18–20], whereas others have found a stronger [21–25] or similar [7, 8, 11] association for the rectum. In addition, the few studies that have investigated the association between alcohol consumption and the risk for cancer in the proximal or distal colon showed a strong positive association in the latter and a weak or null association in the former [7, 11, 18, 22, 26–28]. Thirdly, the dose–response relationship is less apparent in women, probably because they tend to consume less alcohol than men. To date, the largest cohort study among women, with 6300 cases of colorectal cancer, has shown a small and statistically significantly increased risk for rectal, but not colon, cancer [23]. However, the range of alcohol consumption in this cohort was narrow. Finally, the association of alcohol drinking with colorectal cancer risk may be stronger among Asian populations as compared with Western populations, but this may also be due to random variation. Therefore, in order to address these issues we conducted a meta-analysis for any, light, moderate, and heavy alcohol drinking, and dose–risk meta-regression analysis of observational studies published before May 2010 on alcohol consumption and colorectal cancer.

methods

search strategy and inclusion criteria

Publications were identified in PubMed using the Me SH terms ‘alcohol’, ‘ethanol’, ‘alcoholic beverages’, and ‘colorectal neoplasms’ as key words, following the MOOSE (Meta-analyses Of Observational Studies) guidelines [29]. Also, reference lists of the identified articles and previous literature reviews and meta-analyses were carefully examined for additional studies. The criteria for inclusion were as follows: (i) observational epidemiological studies (case–control, case–cohort, or cohort) on total alcohol intake and colorectal cancer incidence or mortality in general population, (ii) published in English before May 2010 (except for one article by Lim and Park [30] in Korean, in which all relevant data and tables were presented in English), (iii) reporting the odds ratio (OR) or relative risk (RR) estimates with the corresponding 95% confidence intervals (CI) or sufficient information to calculate them for each alcohol exposure level, and (iv) reporting an association for at least three categories of alcohol consumption. When several reports were published on the same study, only the most recent and informative one was included.

data abstraction

Figure 1 shows the flowchart for the selection of articles. For each study, the following information was extracted: study design, country, number of patients, duration of follow-up for cohort studies and type of controls for case–control studies, sex, variables adjusted for in the analysis, risk estimates for categories of alcohol drinking and the corresponding 95% CIs, and, when available, the number of cases and noncases or person-years for each level of alcohol consumption. A quality of each study was assessed according to the predefined criteria [31], which addressed study design, assessment of alcohol drinking, and data analysis. The range of the quality score was between 0 (lowest) and 10 (highest) (Tables 1 and 2).

Table 1.

Characteristics of published case–control studies on alcohol intake and colorectal cancer risk

References Country and name of the study Sex strata explored in the analyses Sites explored in the analyses Period of enrolment No. of cases No. of controls Quality score Variables adjusted for (or matched on) in the regression models 
Potter and McMichael [32Australia, South Australian Central Cancer Registry M, W C, R 1979–1981 419 732 Matched on age and sex 
Kune et al. [33Australia, Melbourne Colorectal Cancer Study M, W C, R – 715 727 3.5 Matched on age and sex 
Peters et al. [34United States C, R, CR 1974–1982 147 147 8.5 Matched on age, sex, race, and neighborhood; adjusted for education 
Longnecker [35United States C, R 1985–1988 644 992 Age, income, and smoking 
Slattery et al. [36United States M, W C, CP, CD 1979–1983 231 391 5.5 Age, religion, BMI, and intakes of calories and fiber 
Choi and Kahyo [25Korea, Korea Cancer Center Hospital C, R – 130 390 Matched on age, sex, and admission date; adjusted for marital status, education, diet, and smoking 
Riboli et al. [37France M, W C, R 1979–1985 389 641 3.5 None 
Barra et al. [38Italy M, W, M + W C, R 1985–1990 1470 2475 5.5 Age, sex, study center, BMI, education, and intake of total energy 
Peters et al. [39United States M, W, M + W 1983–1986 746 746 6.5 Matched on age, sex, and neighborhood; adjusted for family history, activity level, weight, and intakes of fat, protein, carbohydrates, calcium, and if female, pregnancies 
Gerhardsson de Verdier et al. [40Sweden M, W C, R 1986–1988 569 512 7.5 Age, sex, BMI, physical activity, smoking, and intakes of total energy, protein, and fiber 
Hoshiyama et al. [41Japan, Saitama Prefecture M + W CR 1984–1990 181 653 Age and sex 
Newcomb et al. [42United States C, R, CR 1990–1991 779 2315 8.5 Age, BMI, screening sigmoidoscopy history, and family history 
Boutron et al. [43France M, W CR 1985 171 309 Age 
Chyou et al. [44United States, Honolulu Heart Program C, R 1965–1968 453 7945 Age 
Murata et al. [19Japan, Chiba Cancer Registry C, CP 1984–1993 887 1774 4.5 Matched on age and address code; no adjustment for other risk factors 
Slattery [45United States, Kaiser Permanente M + W C, CP 1991–1994 1993 2410 5.5 Age at diagnosis, BMI, physical activity, smoking, and intakes of total energy, fiber and calcium 
Yamada et al. [46Japan M + W CR 1991–1993 195 390 4.5 Age, sex, BMI, and smoking 
Tavani et al. [47Italy M + W C, R, CR 1991–1996 1953 4154 Age, sex, education, center, physical activity, smoking, family history, and intakes of beta-carotene, vitamin D, and total energy 
Murata et al.[48Japan M, W C, R, CR 1989–1997 429 794 Age 
Chen et al. [49United States, Physicians' Health Study CR 1982–1995 211 1113 Matched on age and smoking status; adjusted for aspirin and multivitamin use 
Ji et al. [50China M, W C, R 1990–1992 1805 1552 4.5 Matched on age and sex; adjusted for income and smoking 
Sharpe et al. [28Canada C, CP, CD, R 1979–1985 585 500 5.5 Matched on age; adjusted for respondent status, ethnicity, family income, education, marital status, and smoking 
Ho et al. [51Hong Kong M + W C, R, CR 1998–2000 822 926 4.5 None 
Kim et al. [52Korea M + W CR 1998–2000 243 225 None 
Murtaugh et al. [53United States, Kaiser Permanente M, W 1997–2001 952 1205 3.5 Age, physical activity, and intakes of energy, fiber, and calcium 
Hu et al. [27Canada, NECSS M, W C, CP, CD 1994–1997 1723 3097 Matched on age and sex; adjusted for province, education, BMI, and physical activity 
Stern et al. [54Singapore, Singapore Chinese Study M + W CR 1993–2002 310 1176 None 
Gao et al. [55China CR 2000–2002 190 223 7.5 Age and smoking 
Lightfoot et al. [56UK M + W CR 1997–2000 500 742 Matched on age and sex 
Benedetti et al. [57Canada C, R mid-1980s 666 507 8.5 Age, smoking, respondent status, ethnicity, census tract income, and education 
Kim et al. [58Korea M, W, M + W CR 2001–2004 596 509 4.5 None 
Morita et al. [59Japan, Fukuoka Colorectal Cancer Study M + W CR 2000–2003 685 778 4.5 None 
Wernli et al. [60United States C, R, CR 1998–2002 1014 1064 None 
Yamamoto et al. [61Japan, Hitachi Health Center M + W CR 2004–2007 22 66 None 
References Country and name of the study Sex strata explored in the analyses Sites explored in the analyses Period of enrolment No. of cases No. of controls Quality score Variables adjusted for (or matched on) in the regression models 
Potter and McMichael [32Australia, South Australian Central Cancer Registry M, W C, R 1979–1981 419 732 Matched on age and sex 
Kune et al. [33Australia, Melbourne Colorectal Cancer Study M, W C, R – 715 727 3.5 Matched on age and sex 
Peters et al. [34United States C, R, CR 1974–1982 147 147 8.5 Matched on age, sex, race, and neighborhood; adjusted for education 
Longnecker [35United States C, R 1985–1988 644 992 Age, income, and smoking 
Slattery et al. [36United States M, W C, CP, CD 1979–1983 231 391 5.5 Age, religion, BMI, and intakes of calories and fiber 
Choi and Kahyo [25Korea, Korea Cancer Center Hospital C, R – 130 390 Matched on age, sex, and admission date; adjusted for marital status, education, diet, and smoking 
Riboli et al. [37France M, W C, R 1979–1985 389 641 3.5 None 
Barra et al. [38Italy M, W, M + W C, R 1985–1990 1470 2475 5.5 Age, sex, study center, BMI, education, and intake of total energy 
Peters et al. [39United States M, W, M + W 1983–1986 746 746 6.5 Matched on age, sex, and neighborhood; adjusted for family history, activity level, weight, and intakes of fat, protein, carbohydrates, calcium, and if female, pregnancies 
Gerhardsson de Verdier et al. [40Sweden M, W C, R 1986–1988 569 512 7.5 Age, sex, BMI, physical activity, smoking, and intakes of total energy, protein, and fiber 
Hoshiyama et al. [41Japan, Saitama Prefecture M + W CR 1984–1990 181 653 Age and sex 
Newcomb et al. [42United States C, R, CR 1990–1991 779 2315 8.5 Age, BMI, screening sigmoidoscopy history, and family history 
Boutron et al. [43France M, W CR 1985 171 309 Age 
Chyou et al. [44United States, Honolulu Heart Program C, R 1965–1968 453 7945 Age 
Murata et al. [19Japan, Chiba Cancer Registry C, CP 1984–1993 887 1774 4.5 Matched on age and address code; no adjustment for other risk factors 
Slattery [45United States, Kaiser Permanente M + W C, CP 1991–1994 1993 2410 5.5 Age at diagnosis, BMI, physical activity, smoking, and intakes of total energy, fiber and calcium 
Yamada et al. [46Japan M + W CR 1991–1993 195 390 4.5 Age, sex, BMI, and smoking 
Tavani et al. [47Italy M + W C, R, CR 1991–1996 1953 4154 Age, sex, education, center, physical activity, smoking, family history, and intakes of beta-carotene, vitamin D, and total energy 
Murata et al.[48Japan M, W C, R, CR 1989–1997 429 794 Age 
Chen et al. [49United States, Physicians' Health Study CR 1982–1995 211 1113 Matched on age and smoking status; adjusted for aspirin and multivitamin use 
Ji et al. [50China M, W C, R 1990–1992 1805 1552 4.5 Matched on age and sex; adjusted for income and smoking 
Sharpe et al. [28Canada C, CP, CD, R 1979–1985 585 500 5.5 Matched on age; adjusted for respondent status, ethnicity, family income, education, marital status, and smoking 
Ho et al. [51Hong Kong M + W C, R, CR 1998–2000 822 926 4.5 None 
Kim et al. [52Korea M + W CR 1998–2000 243 225 None 
Murtaugh et al. [53United States, Kaiser Permanente M, W 1997–2001 952 1205 3.5 Age, physical activity, and intakes of energy, fiber, and calcium 
Hu et al. [27Canada, NECSS M, W C, CP, CD 1994–1997 1723 3097 Matched on age and sex; adjusted for province, education, BMI, and physical activity 
Stern et al. [54Singapore, Singapore Chinese Study M + W CR 1993–2002 310 1176 None 
Gao et al. [55China CR 2000–2002 190 223 7.5 Age and smoking 
Lightfoot et al. [56UK M + W CR 1997–2000 500 742 Matched on age and sex 
Benedetti et al. [57Canada C, R mid-1980s 666 507 8.5 Age, smoking, respondent status, ethnicity, census tract income, and education 
Kim et al. [58Korea M, W, M + W CR 2001–2004 596 509 4.5 None 
Morita et al. [59Japan, Fukuoka Colorectal Cancer Study M + W CR 2000–2003 685 778 4.5 None 
Wernli et al. [60United States C, R, CR 1998–2002 1014 1064 None 
Yamamoto et al. [61Japan, Hitachi Health Center M + W CR 2004–2007 22 66 None 

BMI, body mass index; C, colon; CD, distal colon; CP, proximal colon; CR, colorectal; M, men; M + W, men and women combined; R, rectal; W, women.

Table 2.

Characteristics of published cohort studies on alcohol intake and colorectal cancer risk

References Country and name of the study Sex strata explored in the analyses Sites explored in the analyses Duration of follow-up (years) No. of cases No. of noncases/person-years Quality score Variables adjusted for in the regression models 
Wu et al. [62United States M, W C, CR 126 11 888 Age 
Klatsky et al. [63United States, Kaiser Permanente M, W, M + W C, R 230 106 203 7.5 Age, smoking, sex, race, BMI, coffee, cholesterol, and education 
Stemmermann et al. [64United States, Iowa Women’s Health Study C, R – 312 – Age at exam, smoking 
Gapstur et al. [65United States C, CP, CD, R 312 41 837 Age 
Goldbohm et al. [66Netherlands M, W, M + W C, R 3.3 330 120 852 Age, smoking, BMI, history of gall bladder surgery, education, energy intake, and energy-adjusted intakes of fat, meat protein and dietary fiber 
Flood et al. [67United States CR 8.5 490 45 264 5.5 Intakes of energy, dietary folate, and methionine and smoking 
Otani et al. [9Japan, Japan Public Health Center-based Prospective Study C, R, CR 7–10 457 42 540 Age, family history, BMI, smoking, physical activity, and study center 
Pedersen et al. [68Denmark M + W C, R 15 613 29 132 5.5 Age, sex, smoking, BMI, and study 
Shimizu et al. [12Japan, Takayama study M, W C, R 295 29 051 Age, height, BMI, smoking, and education 
Sanjoaquin et al. [69UK M, W, M + W CR 17 95 10 998 Age, sex, and smoking 
Su and Arab [70United States, NHANES, NHEFS M + W 111 10 418 7.5 Age, sex, race, BMI, education, history of colonic polyps, smoking, multivitamins, and intakes of non-poultry meat, poultry meat, and seafood 
Wei et al. [71United States, Nurses’ Health Study (NHS) C, R 14–20 1478 134 365 5.5 Age, sex, family history, BMI, physical activity, height, smoking, history of endoscopy, and consumption of beef, pork or lamb, processed meat, calcium, and folate 
Chen et al. [24China, Jiashan County M, W, M + W CR, C, R 11 242 64 343 Age, sex, smoking, occupation, education, and marital status 
Wakai et al. [10Japan, Japan Collaborative Cohort Study M, W R, C 7.69 629 57 736 Age, area, education, family history of colorectal cancer, BMI, smoking, walking time, sedentary work, and consumption of green leafy vegetables and beef 
Akhter et al. [11Japan, Miyagi cohort study C, CP, CD, R, CR 11 307 21 199 Age, family history, education, BMI, walking time, smoking, and intakes of meat, green and yellow vegetables, and fruits 
Ferrari et al. [26Europe, EPIC M + W C, CP, CD, R, CR 6.2 1833 478 732 Age, sex, center, physical activity, smoking, education, weight, height, and intake of energy from nonalcohol sources 
Tsong et al. [72China, Singapore Chinese Study M + W C, R, CR 11 845 63 257 Age, sex, year of recruitment, education, BMI, history of diabetes, family history, smoking, and physical exercise 
Thygesen et al. [18United States, Health Professionals Follow-up study (HPFS) C, CP, CD, R, CR 16 868 47 432 Stratified by age in 1-year groups; adjusted for family history, aspirin use, smoking, physical activity, BMI, colonoscopy, sigmoidoscopy, and intakes of folate, methionine, vitamin D, calcium, total calories, multivitamins, and processed and red meat 
Toriola et al. [73Finland, Findrink study CR 16.7 59 2682 7.5 Age, examination year, vegetable consumption, fiber intake, family history of cancer, smoking, socioeconomic status, and leisure time physical activity 
Bongaerts et al. [22The Netherlands, the Netherlands Cohort Study M + W C, CP, CD, R, CR 13.3 2323 120 852 Age, sex, family history, BMI, physical activity, and intakes of total energy, energy-adjusted fat, fiber, and calcium 
Kabat et al. [74Canada, Canadian National Breast Screening Study CR 16.4 617 89 835 6.5 Age, BMI, smoking, education, menopausal status, oral contraceptive use, hormone replacement therapy, and total calories 
Lim and Park [30Korea, Korea Elderly Pharmacoepidemiologic Cohort (KEPEC) M + W CR 4.8 112 14 304 Age and sex 
Allen et al. [23UK, Million Women Study C, R 10 6298 1 280 296 6.5 Age, region of residence, socioeconomic status, BMI, smoking, physical activity, oral contraceptives, and hormone replacement therapy 
Mortality 
    Kono et al. [75Japan, Male Japanese Physician’s study CR 19 39 5135 Age and smoking 
    Camargo et al. [76United States, US Male Physicians CR 10.7 80 22, 071 Age, smoking, and treatment groups 
    Ozasa [21Japan, Japan Collaborative Cohort Study for Evaluation of Cancer (JACC) M, W C, R 13–15 692 109 778 Age and area of study 
    Yi et al. [20Korea, Kangwha Cohort Study C, R, CR 20.8 26 6291 6.5 Age, smoking, ginseng use, education, and pesticide use 
References Country and name of the study Sex strata explored in the analyses Sites explored in the analyses Duration of follow-up (years) No. of cases No. of noncases/person-years Quality score Variables adjusted for in the regression models 
Wu et al. [62United States M, W C, CR 126 11 888 Age 
Klatsky et al. [63United States, Kaiser Permanente M, W, M + W C, R 230 106 203 7.5 Age, smoking, sex, race, BMI, coffee, cholesterol, and education 
Stemmermann et al. [64United States, Iowa Women’s Health Study C, R – 312 – Age at exam, smoking 
Gapstur et al. [65United States C, CP, CD, R 312 41 837 Age 
Goldbohm et al. [66Netherlands M, W, M + W C, R 3.3 330 120 852 Age, smoking, BMI, history of gall bladder surgery, education, energy intake, and energy-adjusted intakes of fat, meat protein and dietary fiber 
Flood et al. [67United States CR 8.5 490 45 264 5.5 Intakes of energy, dietary folate, and methionine and smoking 
Otani et al. [9Japan, Japan Public Health Center-based Prospective Study C, R, CR 7–10 457 42 540 Age, family history, BMI, smoking, physical activity, and study center 
Pedersen et al. [68Denmark M + W C, R 15 613 29 132 5.5 Age, sex, smoking, BMI, and study 
Shimizu et al. [12Japan, Takayama study M, W C, R 295 29 051 Age, height, BMI, smoking, and education 
Sanjoaquin et al. [69UK M, W, M + W CR 17 95 10 998 Age, sex, and smoking 
Su and Arab [70United States, NHANES, NHEFS M + W 111 10 418 7.5 Age, sex, race, BMI, education, history of colonic polyps, smoking, multivitamins, and intakes of non-poultry meat, poultry meat, and seafood 
Wei et al. [71United States, Nurses’ Health Study (NHS) C, R 14–20 1478 134 365 5.5 Age, sex, family history, BMI, physical activity, height, smoking, history of endoscopy, and consumption of beef, pork or lamb, processed meat, calcium, and folate 
Chen et al. [24China, Jiashan County M, W, M + W CR, C, R 11 242 64 343 Age, sex, smoking, occupation, education, and marital status 
Wakai et al. [10Japan, Japan Collaborative Cohort Study M, W R, C 7.69 629 57 736 Age, area, education, family history of colorectal cancer, BMI, smoking, walking time, sedentary work, and consumption of green leafy vegetables and beef 
Akhter et al. [11Japan, Miyagi cohort study C, CP, CD, R, CR 11 307 21 199 Age, family history, education, BMI, walking time, smoking, and intakes of meat, green and yellow vegetables, and fruits 
Ferrari et al. [26Europe, EPIC M + W C, CP, CD, R, CR 6.2 1833 478 732 Age, sex, center, physical activity, smoking, education, weight, height, and intake of energy from nonalcohol sources 
Tsong et al. [72China, Singapore Chinese Study M + W C, R, CR 11 845 63 257 Age, sex, year of recruitment, education, BMI, history of diabetes, family history, smoking, and physical exercise 
Thygesen et al. [18United States, Health Professionals Follow-up study (HPFS) C, CP, CD, R, CR 16 868 47 432 Stratified by age in 1-year groups; adjusted for family history, aspirin use, smoking, physical activity, BMI, colonoscopy, sigmoidoscopy, and intakes of folate, methionine, vitamin D, calcium, total calories, multivitamins, and processed and red meat 
Toriola et al. [73Finland, Findrink study CR 16.7 59 2682 7.5 Age, examination year, vegetable consumption, fiber intake, family history of cancer, smoking, socioeconomic status, and leisure time physical activity 
Bongaerts et al. [22The Netherlands, the Netherlands Cohort Study M + W C, CP, CD, R, CR 13.3 2323 120 852 Age, sex, family history, BMI, physical activity, and intakes of total energy, energy-adjusted fat, fiber, and calcium 
Kabat et al. [74Canada, Canadian National Breast Screening Study CR 16.4 617 89 835 6.5 Age, BMI, smoking, education, menopausal status, oral contraceptive use, hormone replacement therapy, and total calories 
Lim and Park [30Korea, Korea Elderly Pharmacoepidemiologic Cohort (KEPEC) M + W CR 4.8 112 14 304 Age and sex 
Allen et al. [23UK, Million Women Study C, R 10 6298 1 280 296 6.5 Age, region of residence, socioeconomic status, BMI, smoking, physical activity, oral contraceptives, and hormone replacement therapy 
Mortality 
    Kono et al. [75Japan, Male Japanese Physician’s study CR 19 39 5135 Age and smoking 
    Camargo et al. [76United States, US Male Physicians CR 10.7 80 22, 071 Age, smoking, and treatment groups 
    Ozasa [21Japan, Japan Collaborative Cohort Study for Evaluation of Cancer (JACC) M, W C, R 13–15 692 109 778 Age and area of study 
    Yi et al. [20Korea, Kangwha Cohort Study C, R, CR 20.8 26 6291 6.5 Age, smoking, ginseng use, education, and pesticide use 

BMI, body mass index; C, colon; CD, distal colon; CP, proximal colon; CR, colorectal; M, men; M + W, men and women combined; R, rectal; W, women.

Figure 1.

Flowchart of publication selection for the meta-analysis.

Figure 1.

Flowchart of publication selection for the meta-analysis.

statistical methods

The multivariate-adjusted risk estimates were included in the meta-analyses; however, when unavailable, unadjusted RRs were computed from the exposure distributions for cases and controls as reported in the published article. When studies reported adjusted RR estimates without CIs, the 95% CI for the unadjusted RR estimate penalized by a factor of 1.5 was computed.

Different studies used different units to express alcohol intake. Therefore, alcohol consumption was converted into grams of ethanol per day using the following conversion factors: 1 drink = 12.5 g; 1 ounce = 28.35 g; and 1 ml = 0.8 g. The dose associated with each RR estimate was computed as the midpoint of the corresponding exposure category. When the highest category was open ended, the midpoint was calculated as 1.2 times its lower bound [77]. Nondrinkers or occasional alcohol drinkers were the reference category. Light alcohol drinking was defined as consumption of ≤1 drink/day (≤12.5 g/day of ethanol), moderate as 2–3 drinks/day (12.6–49.9 g/day of ethanol), and heavy as consumption of ≥4 drinks/day (≥50 g/day of ethanol). When more than one study category fell in the range considered for light, moderate, or heavy drinking, or when the same set of controls was used for colorectal cancer subsites (colon and rectum, proximal and distal colon), we combined the corresponding risk estimates using the method by Hamling et al. [78]. When a study reported risk estimates and 95% CI relative to a reference category other than nondrinkers or occasional drinkers, with available data for nondrinkers, the RRs were recalculated using the nondrinkers or occasional drinkers as reference by the method proposed by Greenland and Longnecker [79].

A random effects model was used to estimate pooled RRs in order to take into account the heterogeneity of the risk estimates and to provide more conservative estimates compared with the fixed effects model [80]. Forest plots were done for any, light, moderate, and heavy versus nonconsumption and occasional alcohol consumption. However, only two forest plots for moderate and heavy alcohol consumption are presented. Statistical heterogeneity between studies was assessed with the chi-square statistic and quantified by I2, a statistic that represents the percentage of total variation contributed by between-study variation [80, 81]. A significant heterogeneity was defined as a P value <0.10. To investigate potential sources of between-study heterogeneity, subgroup analyses and meta-regression models were conducted. Also, sensitivity analyses were carried out to assess whether the summary estimates are robust to inclusion of studies (i) with a reference category for alcohol exposure different from nondrinkers, (ii) reporting estimates not adjusted for the main risk factors (age, sex, body fatness, smoking, and physical activity), and (iii) not reporting 95% CI for adjusted risk estimates. Publication bias was assessed using the tests by Egger [82], Begg and Mazumdar [83], the trim and fill method [84], and the contour-enhanced funnel plots [85].

A dose–response analysis was carried out using both linear and nonlinear random effects models on the natural logarithm of the RR using the method by van Houwelingen [86], which was modified by our group [87]. This method accounts for correlation between reported risk estimates within the same study, heterogeneity between the studies, and nonlinear dose–risk relation. Thirty-six second-order fractional polynomial random effects models and linear random effect models were tested. The best-fitting model, defined as the one with the lowest Akaike’s information criterion, a model fit statistic, was selected as the final dose–risk relation model.

All statistical tests were two-sided, and all statistical analyses were carried out with SAS (version 9.2; SAS Institute Inc., Cary, NC) and Stata Statistical Software (version 10; StataCorp LP, College Station, TX).

results

alcohol intake and CRC incidence

A total of 57 studies on colorectal cancer incidence and alcohol intake published between 1986 and 2010 were identified, among which 22 studies were from Asia (Japan, Korea, China, Hong Kong, and Singapore), 2 from Australia, 13 from Western Europe, and 24 from North America (Canada and United States). Of all these studies, 22 reported fully adjusted risk estimates and 36 reported risk estimates adjusted for tobacco smoking (Tables 1 and 2).

The pooled random effects RRs for comparison with nondrinkers were as follows: any drinkers, 1.12 (95% CI 1.06–1.19); light drinkers, 1.00 (95% CI 0.95–1.05); moderate drinkers, 1.21 (95% CI 1.13–1.28); and heavy drinkers, 1.52 (95% CI 1.27–1.81) (Table 3). The relative risks were higher for rectal than for colon cancer among any drinkers (P = 0.03) and light drinkers (P = 0.05), but about the same among moderate and heavy drinkers. There was no significant heterogeneity of effect estimates by colon subsites among any and light drinkers. However, there was a nonsignificant increased risk for cancer of the distal colon compared with the proximal colon among moderate (P = 0.12) and heavy (P = 0.18) drinkers. Men had statistically significantly higher risk than women among any drinkers (P = 0.001) and moderate drinkers (P = 0.02). Geographical region, type of study, study quality, adjustment for main confounders (age, sex, smoking, body mass index, and physical activity), and year of publication were not significant sources of heterogeneity. For colorectal cancer, a potential heterogeneity by geographical location was observed only among heavy drinkers (P = 0.04), with the highest risk summary estimate of 1.81 (95% CI 1.33–2.46) for studies conducted in Asia and the lowest risk summary estimate of 1.16 (95% CI 0.95–1.43) for studies conducted in Europe (supplemental Figure S1, available at Annals of Oncology online). RRs were systematically higher in hospital-based case–control studies than in population-based case–controls; however, the difference was not statistically significant.

Table 3.

Pooled RR estimates for colorectal cancer incidence stratified by colon site, sex, geographical region, and potential modifying factors

Factors stratified Drinkers versus non-/occasional drinkersa
 
Light versus non-/occasional drinkersa
 
Moderate versus non-/occasional drinkersa
 
Heavy versus non-/occasional drinkersa
 
No. of studiesb RR LCI UCI P valuec No. of studiesb RR LCI UCI P valuec No. of studiesb RR LCI UCI P valuec No. of studiesb RR LCI UCI P valuec 
All studies 57 1.12 1.06 1.19  49 1.00 0.95 1.05  53 1.21 1.13 1.28  19 1.52 1.27 1.81  
Site                     
    Colon 42 1.05 0.99 1.12 0.03 36 0.96 0.90 1.02 0.05 39 1.15 1.06 1.24 0.27 16 1.43 1.23 1.67 0.56 
    Rectum 38 1.19 1.09 1.31  32 1.06 0.98 1.14  35 1.23 1.13 1.35  15 1.59 1.18 2.15  
Colon site                     
    Proximal 10 1.02 0.91 1.14 0.66 1.01 0.88 1.16 0.30 1.01 0.86 1.17 0.12 1.38 0.96 1.98 0.18 
    Distal 1.07 0.90 1.28  0.91 0.80 1.05  1.22 1.02 1.47  2.46 1.38 4.40  
Sexd                     
    Female 26 1.00 0.94 1.07 0.001 25 0.95 0.89 1.01 0.27 21 1.08 1.03 1.13 0.02 1.54 1.04 2.29 0.82 
    Male 33 1.25 1.13 1.39  27 1.02 0.92 1.14  32 1.24 1.13 1.37  15 1.62 1.31 2.01  
Geographical region                    
    Asia 19 1.21 1.03 1.43 0.67 14 0.97 0.83 1.14 0.82 19 1.27 1.09 1.49 0.80 1.81 1.33 2.46 0.04 
    Australia 1.04 0.76 1.44  0.98 0.69 1.38  1.10 0.82 1.49    N/A   
    Europe 14 1.09 1.01 1.18  12 1.03 0.97 1.11  13 1.17 1.06 1.29  1.16 0.95 1.43  
    North America 22 1.08 1.01 1.15  21 0.99 0.92 1.05  19 1.18 1.08 1.30  1.59 1.25 2.01  
Type of study                     
    Cohort 23 1.12 1.03 1.22 0.87 23 1.02 0.96 1.08 0.43 22 1.24 1.13 1.28 0.38 1.57 1.38 1.80 0.74 
    Case–control 34 1.11 1.04 1.19  26 0.98 0.90 1.06  31 1.18 1.07 1.29  12 1.49 1.13 1.96  
Source of controlse                     
    Population based 25 1.08 0.99 1.17 0.24 20 0.98 0.90 1.07 0.85 23 1.15 1.03 1.29 0.15 1.43 1.15 1.79 0.82 
    Hospital based 1.26 1.01 1.58  0.96 0.78 1.17  1.29 1.16 1.44  1.54 0.89 2.67  
Quality score                     
    Above median 29 1.08 1.02 1.14 0.31 25 0.99 0.95 1.04 0.71 27 1.21 1.08 1.35 0.91 10 1.42 1.15 1.75 0.46 
    Below median 28 1.15 1.04 1.28  24 1.01 0.92 1.11  26 1.20 1.10 1.29  1.65 1.20 2.26  
Adjustment for main confoundersf                   
    Adjusted 22 1.08 1.02 1.18 0.39 20 1.01 0.97 1.05 0.69 22 1.20 1.11 1.30 0.90 1.42 1.13 1.80 0.54 
    Unadjusted 35 1.14 1.04 1.26  29 0.99 0.91 1.09  31 1.21 1.09 1.34  12 1.59 1.21 2.08  
Publication year                     
    <2000 24 1.10 0.99 1.23 0.67 20 0.97 0.88 1.07 0.46 22 1.17 1.05 1.30 0.45 10 1.49 1.06 2.09 0.89 
    ≥2000 33 1.13 1.07 1.20  29 1.01 0.96 1.05  31 1.23 1.14 1.33  1.53 1.33 1.76  
Factors stratified Drinkers versus non-/occasional drinkersa
 
Light versus non-/occasional drinkersa
 
Moderate versus non-/occasional drinkersa
 
Heavy versus non-/occasional drinkersa
 
No. of studiesb RR LCI UCI P valuec No. of studiesb RR LCI UCI P valuec No. of studiesb RR LCI UCI P valuec No. of studiesb RR LCI UCI P valuec 
All studies 57 1.12 1.06 1.19  49 1.00 0.95 1.05  53 1.21 1.13 1.28  19 1.52 1.27 1.81  
Site                     
    Colon 42 1.05 0.99 1.12 0.03 36 0.96 0.90 1.02 0.05 39 1.15 1.06 1.24 0.27 16 1.43 1.23 1.67 0.56 
    Rectum 38 1.19 1.09 1.31  32 1.06 0.98 1.14  35 1.23 1.13 1.35  15 1.59 1.18 2.15  
Colon site                     
    Proximal 10 1.02 0.91 1.14 0.66 1.01 0.88 1.16 0.30 1.01 0.86 1.17 0.12 1.38 0.96 1.98 0.18 
    Distal 1.07 0.90 1.28  0.91 0.80 1.05  1.22 1.02 1.47  2.46 1.38 4.40  
Sexd                     
    Female 26 1.00 0.94 1.07 0.001 25 0.95 0.89 1.01 0.27 21 1.08 1.03 1.13 0.02 1.54 1.04 2.29 0.82 
    Male 33 1.25 1.13 1.39  27 1.02 0.92 1.14  32 1.24 1.13 1.37  15 1.62 1.31 2.01  
Geographical region                    
    Asia 19 1.21 1.03 1.43 0.67 14 0.97 0.83 1.14 0.82 19 1.27 1.09 1.49 0.80 1.81 1.33 2.46 0.04 
    Australia 1.04 0.76 1.44  0.98 0.69 1.38  1.10 0.82 1.49    N/A   
    Europe 14 1.09 1.01 1.18  12 1.03 0.97 1.11  13 1.17 1.06 1.29  1.16 0.95 1.43  
    North America 22 1.08 1.01 1.15  21 0.99 0.92 1.05  19 1.18 1.08 1.30  1.59 1.25 2.01  
Type of study                     
    Cohort 23 1.12 1.03 1.22 0.87 23 1.02 0.96 1.08 0.43 22 1.24 1.13 1.28 0.38 1.57 1.38 1.80 0.74 
    Case–control 34 1.11 1.04 1.19  26 0.98 0.90 1.06  31 1.18 1.07 1.29  12 1.49 1.13 1.96  
Source of controlse                     
    Population based 25 1.08 0.99 1.17 0.24 20 0.98 0.90 1.07 0.85 23 1.15 1.03 1.29 0.15 1.43 1.15 1.79 0.82 
    Hospital based 1.26 1.01 1.58  0.96 0.78 1.17  1.29 1.16 1.44  1.54 0.89 2.67  
Quality score                     
    Above median 29 1.08 1.02 1.14 0.31 25 0.99 0.95 1.04 0.71 27 1.21 1.08 1.35 0.91 10 1.42 1.15 1.75 0.46 
    Below median 28 1.15 1.04 1.28  24 1.01 0.92 1.11  26 1.20 1.10 1.29  1.65 1.20 2.26  
Adjustment for main confoundersf                   
    Adjusted 22 1.08 1.02 1.18 0.39 20 1.01 0.97 1.05 0.69 22 1.20 1.11 1.30 0.90 1.42 1.13 1.80 0.54 
    Unadjusted 35 1.14 1.04 1.26  29 0.99 0.91 1.09  31 1.21 1.09 1.34  12 1.59 1.21 2.08  
Publication year                     
    <2000 24 1.10 0.99 1.23 0.67 20 0.97 0.88 1.07 0.46 22 1.17 1.05 1.30 0.45 10 1.49 1.06 2.09 0.89 
    ≥2000 33 1.13 1.07 1.20  29 1.01 0.96 1.05  31 1.23 1.14 1.33  1.53 1.33 1.76  
a

Nondrinkers category included nondrinkers and occasional drinkers; light drinking was defined as ≤12.5 g/day of alcohol (≤1 drink/day), moderate drinking as 12.6-49.9 g/day (2–3 drinks/day), and heavy drinking as ≥50 g/day (≥4 drinks/day).

b

Strata-specific results from the same study were counted as one study.

c

P values from the test of homogeneity between strata.

d

Studies reporting estimates separately for men and women were selected.

e

Among case–control studies only.

f

Age, sex, body mass index, and/or physical activity.

LCI, lower confidence interval; RR, relative risk; UCI, upper confidence interval.

Figure 2A presents RRs for colorectal cancer incidence and moderate alcohol intake, compared with no alcohol intake in men and women from 31 case–control and 22 cohort studies. Combined, the 53 studies included more than 20 700 colorectal cancer cases. There was a statistically significant heterogeneity among studies (I2 = 60%, P < 0.001). Summary results did not materially change when studies with no adjustment for potential confounders were excluded (Table 3). Because there was a significant heterogeneity by sex (P = 0.02), the forest plots are also presented by sex (Figure 2B and C). The nine cohort and 12 case–control studies that investigated the association between moderate alcohol intake and colorectal cancer risk among women (involving 6084 cases) did not show heterogeneity (I2 = 0%, P = 0.50; Figure 2B), whereas 11 cohort and 21 case–control studies among men showed substantial heterogeneity (I2 = 55%, P < 0.001; Figure 2C). The summary RRs of colorectal cancer were 1.08 (95% CI 1.03–1.13) and 1.24 (95% CI 1.13–1.37) for women and men, respectively, for moderate alcohol consumption, compared with nondrinkers.

Figure 2.

Pooled risk estimates for colorectal cancer incidence for moderate alcohol drinkers versus nondrinkers or occasional drinkers from case–control and cohort studies reporting estimates for men and women (A), for women (B), and for men (C). Moderate alcohol consumption was defined as 12.6-49.9 g of alcohol per day (>1–3 drinks/day).

Figure 2.

Pooled risk estimates for colorectal cancer incidence for moderate alcohol drinkers versus nondrinkers or occasional drinkers from case–control and cohort studies reporting estimates for men and women (A), for women (B), and for men (C). Moderate alcohol consumption was defined as 12.6-49.9 g of alcohol per day (>1–3 drinks/day).

Figure 3 presents RR estimates for colorectal cancer incidence for heavy alcohol drinkers, compared with nondrinkers or occasional drinkers from seven cohort and 12 case–control studies involving 6653 colorectal cancer cases (I2 = 76%, P < 0.001). The summary RR for heavy drinking was 1.52 (95% CI 1.27–1.81), compared with nondrinkers or occasional drinkers. The majority of studies reported results for men or for men and women combined. Only two studies reported results for women (summary RR = 1.54, 95% CI 1.04–2.29; Table 3). Exclusion of studies with no adjustment for potential confounders (N = 12) slightly attenuated the summary RR (1.42, 95% CI 1.13–1.80; Table 3).

Figure 3.

Pooled risk estimates for colorectal cancer incidence for heavy alcohol drinkers versus nondrinkers or occasional drinkers from case–control and cohort studies reporting estimates for men and women. Heavy alcohol consumption was defined as ≥50 g of alcohol per day (≥4 drinks/day).

Figure 3.

Pooled risk estimates for colorectal cancer incidence for heavy alcohol drinkers versus nondrinkers or occasional drinkers from case–control and cohort studies reporting estimates for men and women. Heavy alcohol consumption was defined as ≥50 g of alcohol per day (≥4 drinks/day).

Detailed evaluation of publication bias suggested that the presence of publication bias is unlikely (supplemental Figures S2 and S3, available at Annals of Oncology online). Furthermore, several sensitivity analyses showed that the summary estimates are robust to inclusion of studies with certain methodological limitations and are not substantially influenced by definition of the highest alcohol intake category (supplemental material, available at Annals of Oncology online). Results for alcohol intake and CRC mortality were consistent with the results for CRC incidence and are presented in the supplemental material (available at Annals of Oncology online).

dose–response meta-analyses

Among the second-order fractional polynomial random effects models, the best-fitting dose–response relationship between alcohol intake and colorectal cancer risk was ln(RR) = 0.006992 × dose − 0.00001 × dose2 (Figure 4). Compared with nondrinkers, the fractional polynomial model estimates of the RR were 1.07 (95% CI 1.04–1.10), 1.18 (95% CI 1.12–1.25), 1.38 (95% CI 1.28–1.50), and 1.82 (95% CI 1.41–2.35) for 10, 25, 50, and 100 g/day of alcohol, respectively.

Figure 4.

Relative risk function and the corresponding 95% confidence interval estimated by van Houwelingen approach, describing the best-fitting dose–response association of alcohol drinking (in grams per day) and colorectal cancer risk.

Figure 4.

Relative risk function and the corresponding 95% confidence interval estimated by van Houwelingen approach, describing the best-fitting dose–response association of alcohol drinking (in grams per day) and colorectal cancer risk.

discussion

The results of this meta-analysis support the evidence for a causal relation between high intakes of alcohol and increased risk for colorectal cancer, and provide additional evidence of an association for moderate intakes of alcohol and a shape for the dose–risk relationship. Compared with nondrinkers or occasional alcohol drinkers, moderate drinking (>1–4 drinks/day, equivalent to 12.6–49.9 g/day of ethanol) was associated with a 21% and heavy drinking (≥4 drinks/day, equivalent to ≥50 g/day of ethanol) with a 52% increased risk for colorectal cancer, whereas light alcohol consumption (≤1 drink/day, equivalent to ≤12.5 g/day of ethanol) was not associated with an increased risk. However, results of the dose–risk analysis showed a statistically significant 7% increased colorectal cancer risk for 10 g/day of alcohol intake, which includes light alcohol consumers.

The results for heavy and moderate drinking are consistent with previous pooled [7, 8] and meta-analyses [14, 15, 17]; however, the results for light drinking in these studies were either not reported or statistically nonsignificant. In our categorical meta-analysis, there was no association between light alcohol intake and colorectal cancer risk; however, the dose–response analysis found a 7% increase in colorectal cancer risk for low doses. The differences between the dose–response analysis and meta-analysis for light drinkers may likely be explained by the different methods used. The dose–response analysis of aggregate data with the use of fractional polynomial allows investigation of functional relations but does not overcome the general limitations of modeling because the risk estimates for low alcohol doses may be influenced by the function used and affected by observations in high-dose categories and by exposure misclassification in general [87].

The association of alcohol drinking with colorectal cancer risk did not differ by colon and rectal anatomic subsites, consistent with previous meta-analysis [13–15] and pooled analysis [7, 8]. The findings according to proximal and distal colon subsites were consistent with the previous observational studies and one pooled analysis [7, 11, 18, 22, 26–28]. Our results suggested a stronger positive association of moderate and heavy alcohol drinking with cancer in the distal colon compared with cancer in the proximal colon, but the difference was not statistically significant.

The results for alcohol drinking and colorectal cancer risk appeared to be similar between men and women for any and light drinkers. There was a suggestion that the colorectal cancer–moderate alcohol drinking association is stronger among men than among women. This can be explained by the limited number of studies reporting data on high alcohol intake among women, by lower average alcohol consumption in women as compared with men, and/or by possible effect modification of the association by sex.

A large number of studies in our meta-analysis allowed us to investigate whether the association between alcohol drinking and colorectal cancer risk is stronger among Asian populations. Consistent with the previous pooled analyses of prospective studies from North America and Europe [7] and Japan [8], our study has found a slightly stronger association between alcohol drinking and colorectal cancer risk among studies from Asia when compared with studies from other geographical regions. Potential explanations for these findings include (i) a high prevalence (up to 30%) of the slow-metabolizing variant of aldehyde dehydrogenase enzyme, which is associated with increased blood levels of acetaldehyde after alcohol ingestion [88], and (ii) other nongenetic factors, e.g. body composition [8]. No studies were published on colorectal cancer–alcohol intake association among South American and African populations; therefore, further research in these populations is required.

Our meta-analysis had several strengths, including an extensive search of literature on the association between colorectal cancer risk and alcohol drinking that was conducted to identify all published articles before May 2010. Furthermore, the associations for colon and rectal cancers were evaluated separately, as well as the associations by sex, geographical region, and other factors. Finally, two different methods were used to investigate the association between colorectal cancer risk and alcohol consumption, which allowed us to conduct traditional meta-analysis by categories of alcohol drinking and dose–response analysis.

Our meta-analysis also had some limitations. A statistically significant heterogeneity between the studies for moderate and high alcohol doses, including open-ended categories, was observed, which was likely to be attributed to the variation in study design and quality. The type of alcoholic beverage, as well as lifetime exposure to alcohol, and drinking patterns were not included in the meta-analysis because very few studies investigated them. Furthermore, high alcohol intake may be associated with behaviors that predispose to colorectal cancer, such as smoking, unhealthy diet, and low physical activity [89–92]; however, exclusion of studies with no adjustment for main risk factors resulted in no substantial change of summary estimates. Another limitation was that we did not examine whether the association of alcohol with colorectal cancer risk varied by folate status, smoking, or other potential modifying factors because very few studies investigated these associations. Furthermore, our results are likely to be affected by some degree of alcohol exposure misclassification. However, studies with a high-quality score, which have a better collection of alcohol exposure data, found results similar to those reported by the studies with low-quality score. Finally, the evaluation of contour-enhanced funnel plots and other methods suggested minor evidence of publication bias.

The results from this large meta-analysis have important public health implications, given the large number of women and, especially, men consuming alcohol and the high incidence of colorectal cancer worldwide and in developed countries in particular. Our results have shown that alcohol consumption was associated with an increase in risk for colorectal cancer, for intakes of >1 drink/day (>12.5 g/day of ethanol). Thus, public health recommendations for colorectal cancer prevention should consider limiting intake of alcoholic beverages.

disclosure

The authors declare no conflict of interest. The funding sources had no influence on the design of the study; the collection, analysis, and interpretation of data; the writing of the report; or the decision to submit the paper for publication.

PB, CLV, and MJ conceived and coordinated the study; VF and LS carried out literature search, selected the articles for this meta-analysis, and extracted the data; VB and MR developed the statistical analyses methods; IT, VB, and MR provided assistance in data analyses; and VF conducted the statistical analyses and drafted the paper. All authors contributed substantially to interpreting the data, writing of the manuscript, and critically reviewing the manuscript.

funding

This work was supported by International Agency for Research on Cancer (IARC, Lyon, France). Fellowship from International Agency for Research on Cancer (VF); Italian Association for Research on Cancer (IT, EN, and CLV); fellowship from Italian Foundation for Cancer Research (IT); PhD fellowship from International Agency for Research on Cancer (FI). International Agency for Research on Cancer (IARC, Lyon, France).

References

1.
World Health Organization, Department of Mental Health and Substance Abuse
Global status report on alcohol 2004
2004
Geneva
World Health Organization, Department of Mental Health and Substance Abuse
2.
Rehm
J
Mathers
C
Popova
S
, et al.  . 
Global burden of disease and injury and economic cost attributable to alcohol use and alcohol-use disorders
Lancet
 , 
2009
, vol. 
373
 
9682
(pg. 
2223
-
2233
)
3.
World Cancer Research Fund/American Institute for Cancer Research
Food, Nutrition, Physical Activity, and the Prevention of Cancer: A Global Perspective
 , 
2007
Washington, DC
AICR
4.
Boffetta
P
Hashibe
M
La Vecchia
C
, et al.  . 
The burden of cancer attributable to alcohol drinking
Int J Cancer
 , 
2006
, vol. 
119
 
4
(pg. 
884
-
887
)
5.
Baan
R
Straif
K
Grosse
Y
, et al.  . 
Carcinogenicity of alcoholic beverages
Lancet Oncol
 , 
2007
, vol. 
8
 
4
(pg. 
292
-
293
)
6.
Secretan
B
Straif
K
Baan
R
, et al.  . 
A review of human carcinogens—Part E: tobacco, areca nut, alcohol, coal smoke, and salted fish
Lancet Oncol
 , 
2009
, vol. 
10
 
11
(pg. 
1033
-
1034
)
7.
Cho
E
Smith-Warner
SA
Ritz
J
, et al.  . 
Alcohol intake and colorectal cancer: a pooled analysis of 8 cohort studies
Ann Intern Med
 , 
2004
, vol. 
140
 
8
(pg. 
603
-
613
)
8.
Mizoue
T
Inoue
M
Wakai
K
, et al.  . 
Alcohol drinking and colorectal cancer in Japanese: a pooled analysis of results from five cohort studies
Am J Epidemiol
 , 
2008
, vol. 
167
 
12
(pg. 
1397
-
1406
)
9.
Otani
T
Iwasaki
M
Yamamoto
S
, et al.  . 
Alcohol consumption, smoking, and subsequent risk of colorectal cancer in middle-aged and elderly Japanese men and women: Japan Public Health Center-based prospective study
Cancer Epidemiol Biomarkers Prev
 , 
2003
, vol. 
12
 
12
(pg. 
1492
-
1500
)
10.
Wakai
K
Kojima
M
Tamakoshi
K
, et al.  . 
Alcohol consumption and colorectal cancer risk: findings from the JACC Study
J Epidemiol
 , 
2005
, vol. 
15
 
Suppl 2
(pg. 
S173
-
S179
)
11.
Akhter
M
Kuriyama
S
Nakaya
N
, et al.  . 
Alcohol consumption is associated with an increased risk of distal colon and rectal cancer in Japanese men: the Miyagi Cohort Study
Eur J Cancer
 , 
2007
, vol. 
43
 
2
(pg. 
383
-
390
)
12.
Shimizu
N
Nagata
C
Shimizu
H
, et al.  . 
Height, weight, and alcohol consumption in relation to the risk of colorectal cancer in Japan: a prospective study
Br J Cancer
 , 
2003
, vol. 
88
 
7
(pg. 
1038
-
1043
)
13.
Corrao
G
Bagnardi
V
Zambon
A
Arico
S
Exploring the dose-response relationship between alcohol consumption and the risk of several alcohol-related conditions: a meta-analysis
Addiction
 , 
1999
, vol. 
94
 
10
(pg. 
1551
-
1573
)
14.
Huxley
RR
Ansary-Moghaddam
A
Clifton
P
, et al.  . 
The impact of dietary and lifestyle risk factors on risk of colorectal cancer: a quantitative overview of the epidemiological evidence
Int J Cancer
 , 
2009
, vol. 
125
 
1
(pg. 
171
-
180
)
15.
Moskal
A
Norat
T
Ferrari
P
Riboli
E
Alcohol intake and colorectal cancer risk: a dose-response meta-analysis of published cohort studies
Int J Cancer
 , 
2007
, vol. 
120
 
3
(pg. 
664
-
671
)
16.
Mizoue
T
Tanaka
K
Tsuji
I
, et al.  . 
Alcohol drinking and colorectal cancer risk: an evaluation based on a systematic review of epidemiologic evidence among the Japanese population
Jpn J Clin Oncol
 , 
2006
, vol. 
36
 
9
(pg. 
582
-
597
)
17.
Bagnardi
V
Blangiardo
M
La Vecchia
C
Corrao
G
A meta-analysis of alcohol drinking and cancer risk
Br J Cancer
 , 
2001
, vol. 
85
 
11
(pg. 
1700
-
1705
)
18.
Thygesen
LC
Wu
K
Gronbaek
M
, et al.  . 
Alcohol intake and colorectal cancer: a comparison of approaches for including repeated measures of alcohol consumption
Epidemiology
 , 
2008
, vol. 
19
 
2
(pg. 
258
-
264
)
19.
Murata
M
Takayama
K
Choi
BC
Pak
AW
A nested case-control study on alcohol drinking, tobacco smoking, and cancer
Cancer Detect Prev
 , 
1996
, vol. 
20
 
6
(pg. 
557
-
565
)
20.
Yi
SW
Sull
JW
Linton
JA
, et al.  . 
Alcohol consumption and digestive cancer mortality in Koreans: the Kangwha Cohort Study
J Epidemiol
 , 
2010
, vol. 
20
 
3
(pg. 
204
-
211
)
21.
Ozasa
K
Alcohol use and mortality in the Japan Collaborative Cohort Study for Evaluation of Cancer (JACC)
Asian Pac J Cancer Prev
 , 
2007
, vol. 
8
 
Suppl
(pg. 
81
-
88
)
22.
Bongaerts
BW
van den Brandt
PA
Goldbohm
RA
, et al.  . 
Alcohol consumption, type of alcoholic beverage and risk of colorectal cancer at specific subsites
Int J Cancer
 , 
2008
, vol. 
123
 
10
(pg. 
2411
-
2417
)
23.
Allen
NE
Beral
V
Casabonne
D
, et al.  . 
Moderate alcohol intake and cancer incidence in women
J Natl Cancer Inst
 , 
2009
, vol. 
101
 
5
(pg. 
296
-
305
)
24.
Chen
K
Jiang
Q
Ma
X
, et al.  . 
Alcohol drinking and colorectal cancer: a population-based prospective cohort study in China
Eur J Epidemiol
 , 
2005
, vol. 
20
 
2
(pg. 
149
-
154
)
25.
Choi
SY
Kahyo
H
Effect of cigarette smoking and alcohol consumption in the etiology of cancers of the digestive tract
Int J Cancer
 , 
1991
, vol. 
49
 
3
(pg. 
381
-
386
)
26.
Ferrari
P
Jenab
M
Norat
T
, et al.  . 
Lifetime and baseline alcohol intake and risk of colon and rectal cancers in the European prospective investigation into cancer and nutrition (EPIC)
Int J Cancer
 , 
2007
, vol. 
121
 
9
(pg. 
2065
-
2072
)
27.
Hu
J
Morrison
H
Mery
L
, et al.  . 
Diet and vitamin or mineral supplementation and risk of colon cancer by subsite in Canada
Eur J Cancer Prev
 , 
2007
, vol. 
16
 
4
(pg. 
275
-
291
)
28.
Sharpe
CR
Siemiatycki
J
Rachet
B
Effects of alcohol consumption on the risk of colorectal cancer among men by anatomical subsite (Canada)
Cancer Causes Control
 , 
2002
, vol. 
13
 
5
(pg. 
483
-
491
)
29.
Stroup
DF
Berlin
JA
Morton
SC
, et al.  . 
Meta-analysis of observational studies in epidemiology: a proposal for reporting. Meta-analysis Of Observational Studies in Epidemiology (MOOSE) group
JAMA
 , 
2000
, vol. 
283
 
15
(pg. 
2008
-
2012
)
30.
Lim
HJ
Park
BJ
Cohort study on the association between alcohol consumption and the risk of colorectal cancer in the Korean elderly
J Prev Med Public Health
 , 
2008
, vol. 
41
 
1
(pg. 
23
-
29
)
31.
Tramacere
I
Scotti
L
Jenab
M
, et al.  . 
Alcohol drinking and pancreatic cancer risk: a meta-analysis of the dose-risk relation
Int J Cancer
 , 
2010
, vol. 
126
 
6
(pg. 
1474
-
1486
)
32.
Potter
JD
McMichael
AJ
Diet and cancer of the colon and rectum: a case-control study
J Natl Cancer Inst
 , 
1986
, vol. 
76
 
4
(pg. 
557
-
569
)
33.
Kune
S
Kune
GA
Watson
LF
Case-control study of alcoholic beverages as etiological factors: the Melbourne Colorectal Cancer Study
Nutr Cancer
 , 
1987
, vol. 
9
 
1
(pg. 
43
-
56
)
34.
Peters
RK
Garabrant
DH
Yu
MC
Mack
TM
A case-control study of occupational and dietary factors in colorectal cancer in young men by subsite
Cancer Res
 , 
1989
, vol. 
49
 
19
(pg. 
5459
-
5468
)
35.
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
 , 
1990
, vol. 
1
 
1
(pg. 
5
-
14
)
36.
Slattery
ML
West
DW
Robison
LM
, et al.  . 
Tobacco, alcohol, coffee, and caffeine as risk factors for colon cancer in a low-risk population
Epidemiology
 , 
1990
, vol. 
1
 
2
(pg. 
141
-
145
)
37.
Riboli
E
Cornee
J
Macquart-Moulin
G
, et al.  . 
Cancer and polyps of the colorectum and lifetime consumption of beer and other alcoholic beverages
Am J Epidemiol
 , 
1991
, vol. 
134
 
2
(pg. 
157
-
166
)
38.
Barra
S
Negri
E
Franceschi
S
, et al.  . 
Alcohol and colorectal cancer: a case-control study from northern Italy
Cancer Causes Control
 , 
1992
, vol. 
3
 
2
(pg. 
153
-
159
)
39.
Peters
RK
Pike
MC
Garabrant
D
Mack
TM
Diet and colon cancer in Los Angeles County, California
Cancer Causes Control
 , 
1992
, vol. 
3
 
5
(pg. 
457
-
473
)
40.
Gerhardsson de Verdier
M
Romelsjo
A
Lundberg
M
Alcohol and cancer of the colon and rectum
Eur J Cancer Prev
 , 
1993
, vol. 
2
 
5
(pg. 
401
-
408
)
41.
Hoshiyama
Y
Sekine
T
Sasaba
T
A case-control study of colorectal cancer and its relation to diet, cigarettes, and alcohol consumption in Saitama Prefecture, Japan
Tohoku J Exp Med
 , 
1993
, vol. 
171
 
2
(pg. 
153
-
165
)
42.
Newcomb
PA
Storer
BE
Marcus
PM
Cancer of the large bowel in women in relation to alcohol consumption: a case-control study in Wisconsin (United States)
Cancer Causes Control
 , 
1993
, vol. 
4
 
5
(pg. 
405
-
411
)
43.
Boutron
MC
Faivre
J
Dop
MC
, et al.  . 
Tobacco, alcohol, and colorectal tumors: a multistep process
Am J Epidemiol
 , 
1995
, vol. 
141
 
11
(pg. 
1038
-
1046
)
44.
Chyou
PH
Nomura
AM
Stemmermann
GN
A prospective study of colon and rectal cancer among Hawaii Japanese men
Ann Epidemiol
 , 
1996
, vol. 
6
 
4
(pg. 
276
-
282
)
45.
Slattery
ML
Schaffer
D
Edwards
SL
, et al.  . 
Are dietary factors involved in DNA methylation associated with colon cancer?
Nutr Cancer
 , 
1997
, vol. 
28
 
1
(pg. 
52
-
62
)
46.
Yamada
K
Araki
S
Tamura
M
, et al.  . 
Case-control study of colorectal carcinoma in situ and cancer in relation to cigarette smoking and alcohol use (Japan)
Cancer Causes Control
 , 
1997
, vol. 
8
 
5
(pg. 
780
-
785
)
47.
Tavani
A
Ferraroni
M
Mezzetti
M
, et al.  . 
Alcohol intake and risk of cancers of the colon and rectum
Nutr Cancer
 , 
1998
, vol. 
30
 
3
(pg. 
213
-
219
)
48.
Murata
M
Tagawa
M
Watanabe
S
, et al.  . 
Genotype difference of aldehyde dehydrogenase 2 gene in alcohol drinkers influences the incidence of Japanese colorectal cancer patients
Jpn J Cancer Res
 , 
1999
, vol. 
90
 
7
(pg. 
711
-
719
)
49.
Chen
J
Ma
J
Stampfer
MJ
, et al.  . 
Alcohol dehydrogenase 3 genotype is not predictive for risk of colorectal cancer
Cancer Epidemiol Biomarkers Prev
 , 
2001
, vol. 
10
 
12
(pg. 
1303
-
1304
)
50.
Ji
BT
Dai
Q
Gao
YT
, et al.  . 
Cigarette and alcohol consumption and the risk of colorectal cancer in Shanghai, China
Eur J Cancer Prev
 , 
2002
, vol. 
11
 
3
(pg. 
237
-
244
)
51.
Ho
JW
Lam
TH
Tse
CW
, et al.  . 
Smoking, drinking and colorectal cancer in Hong Kong Chinese: a case-control study
Int J Cancer
 , 
2004
, vol. 
109
 
4
(pg. 
587
-
597
)
52.
Kim
DH
Ahn
YO
Lee
BH
, et al.  . 
Methylenetetrahydrofolate reductase polymorphism, alcohol intake, and risks of colon and rectal cancers in Korea
Cancer Lett
 , 
2004
, vol. 
216
 
2
(pg. 
199
-
205
)
53.
Murtaugh
MA
Ma
KN
Caan
BJ
Slattery
ML
Association of fluids from beverages with risk of rectal cancer
Nutr Cancer
 , 
2004
, vol. 
49
 
1
(pg. 
25
-
31
)
54.
Stern
MC
Conti
DV
Siegmund
KD
, et al.  . 
DNA repair single-nucleotide polymorphisms in colorectal cancer and their role as modifiers of the effect of cigarette smoking and alcohol in the Singapore Chinese Health Study
Cancer Epidemiol Biomarkers Prev
 , 
2007
, vol. 
16
 
11
(pg. 
2363
-
2372
)
55.
Gao
CM
Takezaki
T
Wu
JZ
, et al.  . 
Polymorphisms of alcohol dehydrogenase 2 and aldehyde dehydrogenase 2 and colorectal cancer risk in Chinese males
World J Gastroenterol
 , 
2008
, vol. 
14
 
32
(pg. 
5078
-
5083
)
56.
Lightfoot
TJ
Barrett
JH
Bishop
T
, et al.  . 
Methylene tetrahydrofolate reductase genotype modifies the chemopreventive effect of folate in colorectal adenoma, but not colorectal cancer
Cancer Epidemiol Biomarkers Prev
 , 
2008
, vol. 
17
 
9
(pg. 
2421
-
2430
)
57.
Benedetti
A
Parent
ME
Siemiatycki
J
Lifetime consumption of alcoholic beverages and risk of 13 types of cancer in men: results from a case-control study in Montreal
Cancer Detect Prev
 , 
2009
, vol. 
32
 
5–6
(pg. 
352
-
362
)
58.
Kim
J
Kim
DH
Lee
BH
, et al.  . 
Folate intake and the risk of colorectal cancer in a Korean population
Eur J Clin Nutr
 , 
2009
, vol. 
63
 
9
(pg. 
1057
-
1064
)
59.
Morita
M
Le Marchand
L
Kono
S
, et al.  . 
Genetic polymorphisms of CYP2E1 and risk of colorectal cancer: the Fukuoka Colorectal Cancer Study
Cancer Epidemiol Biomarkers Prev
 , 
2009
, vol. 
18
 
1
(pg. 
235
-
241
)
60.
Wernli
KJ
Wang
Y
Zheng
Y
, et al.  . 
The relationship between gravidity and parity and colorectal cancer risk
J Womens Health (Larchmt)
 , 
2009
, vol. 
18
 
7
(pg. 
995
-
1001
)
61.
Yamamoto
S
Nakagawa
T
Matsushita
Y
, et al.  . 
Visceral fat area and markers of insulin resistance in relation to colorectal neoplasia
Diabetes Care
 , 
2010
, vol. 
33
 
1
(pg. 
184
-
189
)
62.
Wu
AH
Paganini-Hill
A
Ross
RK
Henderson
BE
Alcohol, physical activity and other risk factors for colorectal cancer: a prospective study
Br J Cancer
 , 
1987
, vol. 
55
 
6
(pg. 
687
-
694
)
63.
Klatsky
AL
Armstrong
MA
Friedman
GD
Hiatt
RA
The relations of alcoholic beverage use to colon and rectal cancer
Am J Epidemiol
 , 
1988
, vol. 
128
 
5
(pg. 
1007
-
1015
)
64.
Stemmermann
GN
Nomura
AM
Chyou
PH
Yoshizawa
C
Prospective study of alcohol intake and large bowel cancer
Dig Dis Sci
 , 
1990
, vol. 
35
 
11
(pg. 
1414
-
1420
)
65.
Gapstur
SM
Potter
JD
Folsom
AR
Alcohol consumption and colon and rectal cancer in postmenopausal women
Int J Epidemiol
 , 
1994
, vol. 
23
 
1
(pg. 
50
-
57
)
66.
Goldbohm
RA
Van den Brandt
PA
Van 't Veer
P
, et al.  . 
Prospective study on alcohol consumption and the risk of cancer of the colon and rectum in the Netherlands
Cancer Causes Control
 , 
1994
, vol. 
5
 
2
(pg. 
95
-
104
)
67.
Flood
A
Caprario
L
Chaterjee
N
, et al.  . 
Folate, methionine, alcohol, and colorectal cancer in a prospective study of women in the United States
Cancer Causes Control
 , 
2002
, vol. 
13
 
6
(pg. 
551
-
561
)
68.
Pedersen
A
Johansen
C
Gronbaek
M
Relations between amount and type of alcohol and colon and rectal cancer in a Danish population based cohort study
Gut
 , 
2003
, vol. 
52
 
6
(pg. 
861
-
867
)
69.
Sanjoaquin
MA
Appleby
PN
Thorogood
M
, et al.  . 
Nutrition, lifestyle and colorectal cancer incidence: a prospective investigation of 10998 vegetarians and non-vegetarians in the United Kingdom
Br J Cancer
 , 
2004
, vol. 
90
 
1
(pg. 
118
-
121
)
70.
Su
LJ
Arab
L
Alcohol consumption and risk of colon cancer: evidence from the national health and nutrition examination survey I epidemiologic follow-up study
Nutr Cancer
 , 
2004
, vol. 
50
 
2
(pg. 
111
-
119
)
71.
Wei
EK
Giovannucci
E
Wu
K
, et al.  . 
Comparison of risk factors for colon and rectal cancer
Int J Cancer
 , 
2004
, vol. 
108
 
3
(pg. 
433
-
442
)
72.
Tsong
WH
Koh
WP
Yuan
JM
, et al.  . 
Cigarettes and alcohol in relation to colorectal cancer: the Singapore Chinese Health Study
Br J Cancer
 , 
2007
, vol. 
96
 
5
(pg. 
821
-
827
)
73.
Toriola
AT
Kurl
S
Laukanen
JA
, et al.  . 
Alcohol consumption and risk of colorectal cancer: the Findrink study
Eur J Epidemiol
 , 
2008
, vol. 
23
 
6
(pg. 
395
-
401
)
74.
Kabat
GC
Miller
AB
Jain
M
Rohan
TE
Dietary intake of selected B vitamins in relation to risk of major cancers in women
Br J Cancer
 , 
2008
, vol. 
99
 
5
(pg. 
816
-
821
)
75.
Kono
S
Ikeda
M
Tokudome
S
, et al.  . 
Alcohol and mortality: a cohort study of male Japanese physicians
Int J Epidemiol
 , 
1986
, vol. 
15
 
4
(pg. 
527
-
532
)
76.
Camargo
CA
Jr
Hennekens
CH
Gaziano
JM
, et al.  . 
Prospective study of moderate alcohol consumption and mortality in US male physicians
Arch Intern Med
 , 
1997
, vol. 
157
 
1
(pg. 
79
-
85
)
77.
Berlin
JA
Longnecker
MP
Greenland
S
Meta-analysis of epidemiologic dose-response data
Epidemiology
 , 
1993
, vol. 
4
 
3
(pg. 
218
-
228
)
78.
Hamling
J
Lee
P
Weitkunat
R
Ambuhl
M
Facilitating meta-analyses by deriving relative effect and precision estimates for alternative comparisons from a set of estimates presented by exposure level or disease category
Stat Med
 , 
2008
, vol. 
27
 
7
(pg. 
954
-
970
)
79.
Greenland
S
Longnecker
MP
Methods for trend estimation from summarized dose-response data, with applications to meta-analysis
Am J Epidemiol
 , 
1992
, vol. 
135
 
11
(pg. 
1301
-
1309
)
80.
Greenland
S
Quantitative methods in the review of epidemiologic literature
Epidemiol Rev
 , 
1987
, vol. 
9
 (pg. 
1
-
30
)
81.
Higgins
JP
Thompson
SG
Quantifying heterogeneity in a meta-analysis
Stat Med
 , 
2002
, vol. 
21
 
11
(pg. 
1539
-
1558
)
82.
Egger
M
Davey Smith
G
Schneider
M
Minder
C
Bias in meta-analysis detected by a simple, graphical test
BMJ
 , 
1997
, vol. 
315
 
7109
(pg. 
629
-
634
)
83.
Begg
CB
Mazumdar
M
Operating characteristics of a rank correlation test for publication bias
Biometrics
 , 
1994
, vol. 
50
 
4
(pg. 
1088
-
1101
)
84.
Duval
S
Tweedie
R
Trim and fill: a simple funnel-plot-based method of testing and adjusting for publication bias in meta-analysis
Biometrics
 , 
2000
, vol. 
56
 
2
(pg. 
455
-
463
)
85.
Peters
JL
Sutton
AJ
Jones
DR
, et al.  . 
Contour-enhanced meta-analysis funnel plots help distinguish publication bias from other causes of asymmetry
J Clin Epidemiol
 , 
2008
, vol. 
61
 
10
(pg. 
991
-
996
)
86.
van Houwelingen
HC
Arends
LR
Stijnen
T
Advanced methods in meta-analysis: multivariate approach and meta-regression
Stat Med
 , 
2002
, vol. 
21
 
4
(pg. 
589
-
624
)
87.
Bagnardi
V
Zambon
A
Quatto
P
Corrao
G
Flexible meta-regression functions for modeling aggregate dose-response data, with an application to alcohol and mortality
Am J Epidemiol
 , 
2004
, vol. 
159
 
11
(pg. 
1077
-
1086
)
88.
Eriksson
CJ
The role of acetaldehyde in the actions of alcohol (update 2000)
Alcohol Clin Exp Res
 , 
2001
, vol. 
25
 
5 Suppl ISBRA
(pg. 
15S
-
32S
)
89.
Breslow
RA
Guenther
PM
Juan
W
Graubard
BI
Alcoholic beverage consumption, nutrient intakes, and diet quality in the US adult population, 1999–2006
J Am Diet Assoc Apr
 , 
2010
, vol. 
110
 
4
(pg. 
551
-
562
)
90.
Burger
M
Mensink
GB
High alcohol consumption in Germany: results of the German National Health Interview and Examination Survey 1998
Public Health Nutr
 , 
2004
, vol. 
7
 
7
(pg. 
879
-
884
)
91.
Hansel
B
Thomas
F
Pannier
B
, et al.  . 
Relationship between alcohol intake, health and social status and cardiovascular risk factors in the urban Paris-Ile-De-France Cohort: is the cardioprotective action of alcohol a myth?
Eur J Clin Nutr
 , 
2010
, vol. 
64
 
6
(pg. 
561
-
568
)
92.
Paavola
M
Vartiainen
E
Haukkala
A
Smoking, alcohol use, and physical activity: a 13-year longitudinal study ranging from adolescence into adulthood
J Adolesc Health
 , 
2004
, vol. 
35
 
3
(pg. 
238
-
244
)