Abstract
Background
Although previous studies have noted an increased risk of colorectal cancer (CRC) among moderate to heavy alcohol consumers relative to non-drinkers, the relationship between alcohol consumption and CRC survival remains unclear.
Methods
Cases of incident invasive CRC diagnosed between 1997-2007 were identified via population-based cancer registries at four study sites in the Colon Cancer Family Registry. Study participants completed a risk factor questionnaire on pre-diagnostic behaviors, including wine, beer, and liquor consumption, at baseline. Prospective follow-up for survival was conducted for 4966 CRC cases. Using Cox regression, we compared non-drinkers to individuals who consumed, on average, ≥1 serving/day of alcohol in the years preceding CRC diagnosis with respect to overall and disease-specific survival. Separate analyses by beverage type, and stratified by patient and tumor attributes, were also performed. All models were adjusted for age at diagnosis, sex, study site, year of diagnosis, smoking history, body mass index, and education.
Results
Pre-diagnostic beer and liquor consumption were not associated with CRC survival; however, higher levels of wine consumption were modestly associated with better prognosis overall [hazard ratio (HR)CRC-specific=0.70, 95% confidence interval (CI): 0.48-1.03 and HRoverall=0.70, 95% CI: 0.53-0.94). Similar patterns were noted in stratified analyses.
Conclusions
These findings suggest that pre-diagnostic wine consumption is modestly associated with more favorable survival after CRC.
Keywords: colorectal cancer, survival, alcohol consumption, wine, tumor phenotype
Introduction
Colorectal cancer (CRC) remains a leading cause of cancer death in the United States.1 Advances in early detection have contributed to favorable 5-year relative survival (90%) for patients with localized CRC; however, survival with distant-stage disease is only 13%.1 Thus, it is important to identify factors contributing to CRC prognosis.
Alcohol has been classified as a Class I carcinogen, and the role of alcohol consumption as an etiologic risk factor for CRC has been well-characterized.2-5 Previous studies have also noted a greater likelihood of advanced CRC at diagnosis and increased risk of liver metastases among heavy drinkers.6-8 However, evidence for a possible relationship between alcohol and CRC survival has been minimal, with the few existing studies finding suggestive but not statistically significant evidence for an association between overall pre-diagnostic alcohol consumption and CRC prognosis.9-12 Pelser and colleagues recently reported that moderate alcohol intake, in comparison to little or no alcohol consumption, was associated with lower risk of all-cause mortality among CRC patients.12 Other studies have also observed patterns in the relationship between recent pre-diagnostic alcohol consumption and CRC outcomes that differ by beverage type.9-11
We conducted an analysis of data from the Colon Cancer Family Registry (C-CFR) to determine whether survival in individuals diagnosed with CRC varied according to patterns of pre-diagnostic alcohol consumption overall, by beverage type, and across case groups defined by patient and tumor attributes.
Materials and Methods
Study Population
The study sample included women and men diagnosed with incident invasive CRC between January 1997 and June 2007 who participated in the C-CFR. The C-CFR is an international collaborative effort between investigators in Australia, Canada, and the United States, with six contributing study centers. C-CFR recruitment protocols and eligibility criteria have been previously detailed.13 We restricted our analysis to population-based cases, not selected on the basis of family history, recruited through cancer registries at one of four sites (University of Melbourne, Melbourne, Victoria, Australia; Cancer Care Ontario, Toronto, Ontario, Canada; Fred Hutchinson Cancer Research Center, Seattle, Washington, USA; Mayo Clinic, Rochester, Minnesota, USA). All study participants provided informed consent. Institutional review boards from each center approved their respective studies.
Alcohol Consumption Assessment
Study participants completed a standardized risk factor questionnaire regarding pre-diagnostic exposures via telephone interviews or mail at the time of enrollment (http://www.coloncfr.org/questionnaires). Interviews were administered an average of 11.4 months (median: 9.4 months) after diagnosis. Participants recalled their alcohol consumption prior to CRC diagnosis for (as applicable) the intervals between ages 20-30, 30-50, and after age 50. Participants were queried as to whether they had consumed ≥1 serving/week of alcohol for a period of ≥6 months, and were separately asked if they had consumed ≥1 serving/week of beer or hard cider, wine or fortified wines, or hard liquor for ≥6 months. A serving was defined as a 12-ounce can/bottle of beer/cider, a 4-ounce glass of wine / 1-ounce glass of fortified wine, or a 1-ounce shot of hard liquor. For each beverage category, participants reported their average servings/week and the number of years within the age interval during which they drank that beverage. To reflect recent pre-diagnostic consumption, we restricted our analysis to intake during the age interval including the age at diagnosis (referred to here as “pre-diagnostic alcohol consumption”).14 We further categorized alcohol intake as “average alcohol use overall” and by alcoholic beverage type.
Tumor Characteristics
Cancers arising from the cecum through the splenic flexure (ICD-O-3 codes C180, C182-C185) were grouped together as right-sided colon cancers, while neoplasms in the descending or sigmoid colon (C186-C187) were defined as left-sided colon cancers. Tumors originating in the rectosigmoid junction (C199) or rectum (C209) were classified as rectal cancers.15
We assessed mismatch repair status (MMR) status using one of two methods.13 The first method was genetic analysis for microsatellite instability (MSI) based on a 10-marker panel using tumor DNA and DNA from normal surrounding tissue.10 We categorized tumors as MMR deficient (i.e., high MSI) if instability was observed for ≥30% of markers and as MMR proficient (i.e., microsatellite stable) if instability was seen in <30% of markers. For a subset of cases, we evaluated MMR status using immunohistochemistry of four markers.16 Cases with tumors where all markers exhibited positive staining were categorized as proficient (pMMR); tumors with ≥1 negative marker were classified as deficient (dMMR). We assayed tumors for the presence of the V600E BRAF mutation using a fluorescent allele-specific PCR.17 CIMP testing was based on a validated, five-gene panel quantitative DNA methylation assay, and was completed for cases, with complete MMR information, diagnosed before July 2002. Cases were classified as CIMP-high if the percentage of methylated reference (PMR) ratio was ≥10 for ≥3 of 5 markers in a tumor and as non-CIMP if the PMR ratio was ≥10 for <3 markers.18
Vital Status and Cause of Death
We ascertained vital status, date and cause of death via linkage to population-based registries, contact with relatives, and collection of death certificates. For a small number of cases, we determined vital status through other means (e.g., obituaries).19 We categorized a death as CRC-specific when the underlying cause of death was classified as C18.0-C20.0 or C26.0 (ICD-10). Vital status information was available through December 2013. Cases who were alive through the most recent evaluation of vital status were censored at that date.
Statistical Analysis
We employed Cox proportional hazards regression to assess the relationship of pre-diagnostic alcohol consumption with overall and CRC-specific survival. Our time scale was defined as days since CRC diagnosis, with left censoring to account for the time-lag between diagnosis and C-CFR enrollment. Separate analyses were conducted for each outcome. In analyses of CRC-specific survival, we censored participants who died from causes other than CRC. Proportional hazards assumptions were supported by testing for a non-zero slope of the scaled Schoenfeld residuals on ranked failure times.
We conducted separate analyses for alcohol overall and by beverage type. We used non-drinkers as the reference category, including those who consumed an average of <1 serving of alcohol/week. We also conducted analyses stratified by sex, age at diagnosis, smoking history, tumor site, BRAF-mutation status, MMR status, and CIMP status.
We adjusted all regression models for age at diagnosis, year of diagnosis, sex, study site, and several self-reported pre-diagnostic exposures ascertained in the baseline interview: smoking history, body mass index, educational attainment, and use of non-steroidal anti-inflammatory drugs (NSAIDs). All covariates were categorized as shown in Table 1. We conducted all analyses in STATA SE 14 (College Station, Texas).
Table 1. Study population characteristics.
| Pre-diagnostic alcohol consumption: | ||||
|---|---|---|---|---|
|
|
||||
| Non-drinker or <1 drink / week N (%) |
≥1 drink / week but ≤1 drink / day N (%) |
>1 drink / day N (%) |
Chi-square P-value | |
| Age at diagnosis: | <0.01 | |||
| <40 | 233 (10) | 166 (12) | 119 (10) | |
| 40-49 | 759 (32) | 518 (37) | 438 (36) | |
| 50-59 | 545 (23) | 316 (23) | 282 (23) | |
| 60-69 | 530 (22) | 276 (20) | 251 (21) | |
| ≥70 | 298 (13) | 123 (9) | 112 (9) | |
| Median age (SD) | 53 (11.7) | 50 (11.4) | 51 (11.2) | |
| Study center: | <0.01 | |||
| Ontario, Canada | 603 (26) | 477 (34) | 415 (35) | |
| Melbourne, Australia | 251 (11) | 296 (21) | 251 (21) | |
| Minnesota, USA | 273 (12) | 140 (10) | 113 (9) | |
| Seattle-Puget Sound, USA | 1238 (52) | 486 (35) | 423 (35) | |
| Sex: | <0.01 | |||
| Male | 896 (38) | 748 (53) | 938 (78) | |
| Female | 1469 (62) | 651 (47) | 264 (22) | |
| Cigarette smoking history: | <0.01 | |||
| Never smoker | 1229 (52) | 587 (42) | 321 (27) | |
| Former smoker | 769 (33) | 588 (42) | 545 (46) | |
| Current smoker | 357 (15) | 216 (15) | 331 (28) | |
| Missing | 10 | 8 | 5 | |
| Body mass index (kg/m2): | <0.01 | |||
| <25.0 | 786 (34) | 556 (40) | 381 (32) | |
| 25.0-29.9 | 826 (35) | 557 (39) | 535 (45) | |
| ≥30.0 | 718 (31) | 284 (20) | 275 (23) | |
| Missing | 35 | 12 | 11 | |
| Pre-diagnostic NSAID use: | 0.59 | |||
| No | 1837 (80) | 1092 (80) | 925 (79) | |
| Yes | 461 (20) | 268 (20) | 250 (21) | |
| Missing | 64 | 38 | 23 | |
| Educational attainment: | <0.01 | |||
| High school graduate or less | 900 (38) | 471 (34) | 475 (40) | |
| Some college / vocational school | 784 (33) | 439 (32) | 404 (34) | |
| College graduate | 664 (28) | 480 (35) | 318 (27) | |
| Missing | 17 | 9 | 5 | |
| Tumor site: | <0.01 | |||
| Right-sided | 878 (37) | 484 (35) | 357 (30) | |
| Left-sided | 703 (30) | 371 (27) | 355 (30) | |
| Rectal | 765 (33) | 524 (38) | 475 (40) | |
| Missing | 19 | 20 | 15 | |
| Mismatch repair status: | <0.01 | |||
| Proficient | 1562 (84) | 997 (88) | 872 (90) | |
| Deficient | 307 (16) | 139 (12) | 101 (10) | |
| Unknown | 496 | 263 | 229 | |
| BRAF-mutation status: | 0.01 | |||
| Wildtype | 1324 (87) | 861 (90) | 706 (90) | |
| Mutated | 206 (13) | 97 (10) | 77 (10) | |
| Missing | 835 | 441 | 419 | |
| CIMP status: | <0.01 | |||
| CIMP-high | 207 (17) | 75 (10) | 53 (9) | |
| Non-CIMP | 1009 (83) | 612 (90) | 525 (91) | |
| Missing | 1149 | 712 | 624 | |
Results
Compared to individuals who consumed <1 serving of alcohol/week, individuals in the upper alcohol consumption category (>1 serving/day) were more likely to be male (78% vs. 38%), to have a history of smoking (74% vs. 48%), and to have a primary tumor in the rectum (40% vs. 33%), and were less likely to be obese (23% vs. 31%), or to have a primary tumor that was MMR deficient (12% vs. 16%) or CIMP-high (10% vs. 17%) (Table 1). The distribution of participant characteristics was similar among persons in the middle versus upper consumption categories, except for sex (53% vs. 78% male) and smoking status (15% vs. 28% current smokers).
After adjustment for patient and tumor characteristics, there was no evidence of a relationship between overall pre-diagnostic alcohol consumption and survival (Table 2). With respect to specific alcohol types, evidence of a modest but not statistically significant poorer overall survival was noted in those who consumed >1 serving of liquor/day versus <1 serving/week (HRoverall: 1.12, 95% CI: 0.92-1.36). Conversely, there was evidence of an inverse association of wine consumption with survival. This relationship was primarily evident in the upper category of wine consumption (HRCRC-specific: 0.70, 95% CI: 0.48-1.03 and HRoverall: 0.70, 95% CI: 0.53-0.94).
Table 2. Pre-diagnostic alcohol consumption and survival after colorectal cancer diagnosis.
| Overall Survival | CRC-Specific Survival | |||||||
|---|---|---|---|---|---|---|---|---|
|
|
||||||||
| Total N | # deaths | HR (95% CI)* | P | # deaths | HR (95% CI)* | P | ||
| Alcohol type (not mutually exclusive among drinkers) | Non-drinker / <1 drink/week | 2245 | 869 | 1.0 (ref) | 519 | 1.0 (ref) | ||
| Beer / hard cider | 1360 | 476 | 0.93 (0.80-1.08) | 0.36 | 312 | 1.01 (0.84-1.22) | 0.96 | |
| Wine / fortified wines | 1267 | 418 | 0.81 (0.65-1.02) | 0.07 | 264 | 0.90 (0.68-1.2) | 0.39 | |
| Liquor | 1138 | 453 | 0.89 (0.74-1.09) | 0.30 | 249 | 0.94 (0.73-1.21) | 0.66 | |
| Drinker – type unspecified | 16 | 11 | 6 | |||||
| Beer servings: | No beer / <1 serving/week | 3070 | 1181 | 1.0 (ref) | 691 | 1.0 (ref) | ||
| ≤1 serving/day (≥1 serving/week) | 874 | 279 | 0.89 (0.76-1.05) | 0.17 | 185 | 1.01 (0.83-1.23) | 0.93 | |
| >1 serving/day | 486 | 197 | 1.01 (0.83-1.21) | 0.89 | 127 | 1.17 (0.92-1.48) | 0.20 | |
| Wine servings: | No wine / <1 serving/week | 3272 | 1298 | 1.0 (ref) | 772 | 1.0 (ref) | ||
| ≤1 serving/day (≥1 serving/week) | 1038 | 348 | 0.96 (0.83-1.12) | 0.62 | 222 | 0.95 (0.79-1.15) | 0.62 | |
| >1 serving/day | 229 | 70 | 0.70 (0.53-0.94) | 0.02 | 42 | 0.70 (0.48-1.03) | 0.07 | |
| Liquor servings: | No liquor / <1 serving/week | 3242 | 1197 | 1.0 (ref) | 744 | 1.0 (ref) | ||
| ≤1 serving/day (≥1 serving/week) | 838 | 310 | 0.94 (0.81-1.09) | 0.40 | 170 | 0.88 (0.72-1.08) | 0.23 | |
| >1 serving/day | 281 | 137 | 1.12(0.92-1.36) | 0.25 | 76 | 1.09 (0.82-1.44) | 0.46 | |
Adjusted for age at diagnosis, year of diagnosis, sex, smoking history, body mass index, education, study site
Counts exclude those with missing data on adjustment variables (N=130)
When analyses focused on pre-diagnostic wine consumption, we observed generally similar findings within case subgroups (Tables 3-4). This association, strongest in the upper wine consumption category, was only statistically significant in men (HRCRC-specific: 0.57, 95% CI: 0.34-0.97 and HRoverall: 0.61, 95% CI: 0.42-0.90) and in former smokers (HRCRC-specific: 0.59, 95% CI: 0.36-0.98 and HRoverall: 0.62, 95% CI: 0.42-0.91). There was some difference in the association depending on CRC site, with the strongest inverse relationship noted among cases with left-sided colon tumors (HRCRC-specific : 0.44, 95% CI: 0.20-0.963 and HRoverall: 0.59, 95% CI: 0.33-1.02). Conversely, we noted that current smokers who were modest wine drinkers displayed a poorer prognosis than individuals who consumed <1 serving/week (HRCRC-specific: 1.80, 95% CI: 1.21-2.67 and HRoverall: 1.41, 95% CI: 1.00-1.98).
Table 3. Pre-diagnostic wine consumption and overall survival after colorectal cancer diagnosis by patient and tumor attributes.
| No wine (<1 serving/week)† |
≤1 serving/day of wine (≥1 serving/week)† |
>1 servings/day of wine† | ||||||
|---|---|---|---|---|---|---|---|---|
|
|
||||||||
| N (deaths) | HR (95% CI)* | N (deaths) | HR (95% CI)* | P | N (deaths) | HR (95% CI)* | P | |
| Sex | ||||||||
| Male | 1674 (725) | 1.0 (ref) | 554 (196) | 0.87 (0.71-1.05) | 0.15 | 138 (46) | 0.61 (0.42-0.90) | 0.01 |
| Female | 1598 (573) | 1.0 (ref) | 484 (152) | 1.07 (0.85-1.33) | 0.57 | 91 (24) | 0.81 (0.52-1.24) | 0.32 |
| Age at diagnosis | ||||||||
| ≤50 | 1562 (506) | 1.0 (ref) | 509 (149) | 0.95 (0.77-1.18) | 0.66 | 89 (20) | 0.63 (0.38-1.03) | 0.07 |
| >50 | 1710 (792) | 1.0 (ref) | 529 (199) | 0.97 (0.80-1.18) | 0.77 | 140 (50) | 0.73 (0.51-1.03) | 0.07 |
| Smoking history | ||||||||
| Never smoker | 1449 (472) | 1.0 (ref) | 453 (139) | 1.03 (0.82-1.29) | 0.73 | 65 (14) | 0.73 (0.39-1.37) | 0.27 |
| Former smoker | 1.0 (ref) | 459 (151) | 0.77 (0.62-0.97) | 114 (38) | 0.62 (0.42-0.91) | 0.02 | ||
| Current smoker | 1185 (527) 638 (299) | 1.0 (ref) | 126 (58) | 1.41 (1.00-1.98) | 0.02 0.05 | 50 (18) | 0.77 (0.44-1.37) | 0.42 |
| Tumor site | ||||||||
| Right-sided | 1133 (485) | 1.0 (ref) | 342 (122) | 1.01 (0.79-1.28) | 0.96 | 65 (19) | 0.66 (0.39-1.13) | 0.13 |
| Left-sided | 970 (350) | 1.0 (ref) | 272 (83) | 1.14 (0.85-1.52) | 0.39 | 67 (18) | 0.59 (0.33-1.02) | 0.06 |
| Rectal | 1139 (447) | 1.0 (ref) | 410 (137) | 0.91 (0.72-1.17) | 0.47 | 97 (33) | 0.81 (0.54-1.21) | 0.30 |
| MMR status | 2210 | 1.0 (ref) | ||||||
| Proficient | (889) 377 | 775 (269) | 0.97 (0.82-1.14) | 0.69 | 172 (58) | 0.76 (0.55-1.05) | 0.10 | |
| Deficient | (113) | 1.0 (ref) | 96 (15) | 0.54 (0.29-1.02) | 0.06 | 22 (4) | -- | |
| BRAF-mutation status | ||||||||
| Wildtype | 1852 | 1.0 (ref) | 696 (226) | 1.03 (0.85-1.23) | 0.78 | 162 (56) | 0.82 (0.59-1.14) | 0.25 |
| Mutated | (716)251 (107) | 1.0 (ref) | 73 (24) | 1.24 (0.70-2.20) | 0.47 | 19 (3) | -- | |
| CIMP status | ||||||||
| Non-CIMP | 1371 | 1.0 (ref) | 484 (172) | 0.91 (0.73-1.13) | 0.39 | 116 (39) | 0.70 (0.47-1.02) | 0.07 |
| CIMP-high | (581)246 (114) | 1.0 (ref) | 54 (19) | 1.05 (0.58-1.90) | 0.52 | 11 (3) | -- | |
Unadjusted 5-year survival %: <1 serving/week (74.6%), ≤1 serving/day (≥1 serving/week) (77.7%), >1 serving/day (82.1%)
Adjusted for age at diagnosis, year of diagnosis, sex, smoking history, body mass index, education, study site
-- Result based on <5 deaths in one comparison group
Table 4. Pre-diagnostic wine consumption and disease-specific survival after colorectal cancer diagnosis by patient and tumor attributes.
| No wine (<1 serving/week)† |
≤1 serving/day of wine (≥1 serving/week)† |
>1 servings/day of wine† | ||||||
|---|---|---|---|---|---|---|---|---|
|
|
||||||||
| N (deaths) | HR (95% CI)* | N (deaths) | HR (95% CI)* | P | N (deaths) | HR (95% CI)* | P | |
| Sex | ||||||||
| Male | 1622 (423) | 1.0 (ref) | 527 (121) | 0.85 (0.66-1.10) | 0.22 | 131 (24) | 0.57 (0.34-0.97) | 0.04 |
| Female | 1552 (349) | 1.0 (ref) | 468 (101) | 1.10 (0.83-1.46) | 0.50 | 88 (18) | 0.92 (0.56-1.52) | 0.76 |
| Age at diagnosis | ||||||||
| ≤50 | 1499 (363) | 1.0 (ref) | 485 (109) | 1.01 (0.78-1.30) | 0.94 | 82 (13) | 0.63 (0.34-1.18) | 0.15 |
| >50 | 1675 (409) | 1.0 (ref) | 510 (113) | 0.93 (0.71-1.22) | 0.59 | 137 (29) | 0.74 (0.46-1.18) | 0.20 |
| Smoking history | ||||||||
| Never smoker | 1408 (302) | 1.0 (ref) | 432 (92) | 1.01 (0.76-1.34) | 0.88 | 64 (9) | 0.82 (0.39-1.73) | 0.59 |
| Former smoker | 1159 (300) | 1.0 (ref) | 441 (86) | 0.70 (0.51-0.96) | 0.01 | 108 (21) | 0.59 (0.36-0.98) | 0.04 |
| Current smoker | 607 (170) | 1.0 (ref) | 122 (44) | 1.80 (1.21-2.67) | 0.004 | 47 (12) | 0.82 (0.40-1.68) | 0.64 |
| Tumor site | ||||||||
| Right-sided | 1100 (260) | 1.0 (ref) | 330 (86) | 1.31 (0.98-1.76) | 0.07 | 63 (14) | 0.92 (0.48-1.75) | 0.79 |
| Left-sided | 943 (201) | 1.0 (ref) | 256 (42) | 0.76 (0.49-1.20) | 0.24 | 62 (9) | 0.44 (0.20-0.96) | 0.04 |
| Rectal | 1103 (302) | 1.0 (ref) | 397 (91) | 0.89 (0.66-1.20) | 0.45 | 94(19) | 0.72 (0.42-1.22) | 0.22 |
| MMR status | ||||||||
| Proficient | 2154 (558) | 1.0 (ref) | 740 (172) | 0.91 (0.73-1.13) | 0.39 | 164 (35) | 0.76 (0.51-1.15) | 0.19 |
| Deficient | 374 (44) | 1.0 (ref) | 96 (7) | 0.56 (0.24-1.33) | 0.19 | 21 (2) | -- | |
| BRAF-mutation status | ||||||||
| Wildtype | 1816 (435) | 1.0 (ref) | 666 (138) | 0.90 (0.71-1.14) | 0.39 | 156 (35) | 0.80 (0.53-1.23) | 0.32 |
| Mutated | 243 (44) | 1.0 (ref) | 72 (20) | 1.76 (0.93-3.33) | 0.08 | 19 (3) | -- | |
| CIMP status | 1.0 (ref) | 464 (106) | 0.82 (0.62-1.09) | 0.18 | 115 (25) | 0.71 (0.44-1.16) | 0.18 | |
| Non-CIMP | 1346 (356) | |||||||
| CIMP-high | 242 (51) | 1.0 (ref) | 54 (13) | 0.96 (0.46-1.99) | 0.91 | 11 (3) | -- | |
Unadjusted 5-year survival %: <1 serving/week (80.2%), ≤1 serving/day (≥1 serving/week) (81.4%), >1 serving/day (84.7%)
Adjusted for age at diagnosis, year of diagnosis, sex, smoking history, body mass index, education, study site
-- Result based on <5 deaths in one comparison group
Discussion
In this prospective analysis of invasive CRC, there was no evidence for an association between overall pre-diagnostic alcohol consumption and disease survival. When considering different types of beverage types, however, our findings suggested slightly better survival with higher levels of pre-diagnostic wine consumption, especially in men and former smokers.
Consistent with the results presented here, the majority of previous studies have found no evidence for a relationship between pre-diagnostic alcohol consumption overall and CRC outcomes. In the largest evaluation of this association to-date, Pelser et al reported similar findings to ours in that they found a suggestively, but not significantly lower risk of all-cause mortality when comparing moderate drinkers to non-drinkers (RRcolon cancer: 0.86, 95% CI: 0.73-1.01 and RRrectal cancer: 0.84, 95% CI: 0.62-1.14).12 Among colon cancer cases specifically, this observed relationship was likely primarily driven by the association between pre-diagnostic alcohol intake and lower cardiovascular disease (CVD) mortality (RRcolon cancer: 0.47, 95% CI: 0.31-0.73). In a previous analysis of a subset of the data included in the present analysis, Phipps et al similarly found no association between overall alcohol consumption and CRC outcomes;10 although that study also considered separate associations with pre-diagnostic wine, beer, and liquor consumption, results were inconclusive due to small numbers.
Of the three prior studies to perform separate evaluations by beverage type, two observed statistically significant findings indicating more favorable prognosis in those with higher pre-diagnostic wine intake.9-11 Zell and colleagues reported a significant association between overall survival and regular (≥1-3 glasses/month) versus infrequent wine intake (<1 glass/month) among CRC patients with a family history of CRC (HR: 0.50, 95% CI: 0.25-0.99).11 Additionally, a recent study by Phipps and colleagues found significantly better overall survival among stage III colon cancer cases reporting ≥30 servings of wine/month in the year prior to cancer diagnosis (HR: 0.51, 95% CI: 0.30-0.85, versus never consumers),9 and found no evidence for variations in the effect of pre-diagnostic alcohol use on CRC outcomes according to evaluated tumor characteristics.9 This result, and our similar findings, could be impacted by low numbers, especially within wine-only analyses, but may also suggest that the impact of alcohol consumption on disease progression or development does not depend on known molecular pathways of disease process.
Consistent with recent meta-analyses indicating a positive association between alcohol consumption and CRC risk,3,4 in vitro evidence suggests that high alcohol consumption, perhaps via its mechanism as a local solvent, plays a role in early tumorigenesis in the colon and rectum.20 Alcohol damages the epithelial cells of the intestinal tract through direct and indirect means, assisting the absorption of acetaldehyde, a recognized carcinogen, and facilitating the production of species that proliferate DNA damage, oxidative stress, and lipid peroxidation in the intestinal lumen.20 Increased alcohol intake can also lead to a suppression of immune surveillance, which could impact both CRC risk and progression.20 Our results indicated no association between overall pre-diagnostic alcohol consumption and survival, but also hinted at a potentially less favorable prognosis given higher daily liquor intake.
Conversely, recent evidence suggests that moderate wine consumption may elicit benefits not only for CVD survival, but also for CRC prognosis.21 With respect to cardiovascular outcomes, moderate wine consumption has been linked to lower levels of inflammatory markers.22 Red wine also contains resveratrol and polyphenols – elements that appear to exhibit anti-coagulatory,23 anti-oxidant,24 and anti-platelet traits.21 However, whether and how these properties contribute to CRC survival is unclear.11,25 Studies using mice found that resveratrol triggers apoptosis in cancer cell lines,26 downregulated intestinal genes involved in cell proliferation or cell cycle progression,27 interrupted growth of transplanted human primary gastric cancer cells,28 and intensified the anti-tumor effect of 5-fluorouracil.29 Thus, it is plausible that our findings reflect some CRC-specific effects of wine.
Interpretation of these findings is subject to some limitations. Our measure of alcohol consumption required self-reported recall for a time period prior to CRC diagnosis. By relying on this variable, we assumed that patient recall was accurate. Furthermore, because our measure of alcohol intake could reflect an averaged consumption level over an extended time (e.g., ages 30-50 for someone diagnosed at age 50), there is potential for misclassification of the relevant exposure. Additionally, despite the large size of our overall study, wine-specific analyses stratified by tumor markers lacked the statistical precision necessary to conclude differences between MMR, CIMP, and BRAF-mutation strata. We also had limited information on stage at diagnosis in our sample; however, when we performed a sensitivity analysis stratified by tumor stage among those with these data, our findings and interpretations did not change. Finally, our findings may be partially attributable to lifestyle differences between individuals who regularly consume wine and non-consumers. Past evidence indicates that wine consumption is positively associated with healthier diet,30 physical activity,30 educational attainment,31 and income,31 all of which may also be associated with CRC survival. While we did adjust for several patient characteristics, residual confounding is possible.
This study is among the first to focus on the relationship between pre-diagnostic alcohol consumption, overall and by type, and CRC outcomes using population-based data. In line with past findings, we observed that overall pre-diagnostic alcohol use was not associated with overall or disease-specific survival, but also that higher levels of pre-diagnostic wine consumption were associated with a more favorable prognosis. Because the mechanistic action of wine on CRC prognosis is not understood, further study is merited to better elucidate biological pathways and explore possible alternative explanations.
Acknowledgments
Funding: This work was supported by grant UM1 CA167551 from the National Cancer Institute (NCI), National Institutes of Health (NIH), and through cooperative agreements with members of the Colon Cancer Family Registry and Principal Investigators. Centers contributing to this analysis include the Australasian Colorectal Cancer Family Registry (U01/U24CA097735), Mayo Clinic Cooperative Family Registry for Colon Cancer Studies (U01/U24CA074800), Ontario Familial Colorectal Cancer Registry (U01/U24CA074783), and Seattle Colorectal Cancer Family Registry (U01/U24CA074794). This work was also supported by NCI/NIH grant K07CA172298 (A.I.P.) and National Institute of Environmental Health Sciences (NIEHS)/NIH grant T32ES015459 (J.R.R.). The content of this manuscript does not necessarily reflect the views or policies of the NIH or any of the collaborating centers in the Colon Cancer Family Registry, nor does mention of trade names, commercial products, or organizations imply endorsement by the US Government or the CFR.
Footnotes
Conflict of Interest Disclosures: None to report.
Author Contributions: Amanda I. Phipps: Conceptualization, methodology, formal analysis, data curation, writing (original draft, review, editing), visualization, supervision, project administration, funding acquisition. Jamaica R. Robinson: Methodology, writing (original draft, review, editing), visualization. Peter T. Campbell: Methodology, writing (reviewing, editing). Aung Ko Win: Methodology, resources, data curation, writing (reviewing, editing). Jane Figueiredo: Methodology, writing (review and editing). Noralane M. Lindor: Methodology, investigation, resources, data curation, writing (reviewing, editing), funding acquisition. Polly A. Newcomb: Conceptualization, methodology, investigation, resources, data curation, writing (original draft,reviewing, editing), visualization, supervision, funding acquisition.
References
- 1.American Cancer Society. Colorectal Cancer: Facts and Figures 2014-2016. Atlanta: American Cancer Society; 2014. [Google Scholar]
- 2.International Agency for Research on Cancer. Alcohol consumption and ethyl carbamate. Lyon, France: IARC; 2010. [Google Scholar]
- 3.Feng YL, Shu L, Zheng PF, et al. Dietary patterns and colorectal cancer risk: a meta-analysis. Eur J Cancer Prev. 2016 doi: 10.1097/CEJ.0000000000000245. [DOI] [PubMed] [Google Scholar]
- 4.Fedirko V, Tramacere I, Bagnardi V, et al. Alcohol drinking and colorectal cancer risk: an overall and dose-response meta-analysis of published studies. Ann Oncol. 2011;22(9):1958–1972. doi: 10.1093/annonc/mdq653. [DOI] [PubMed] [Google Scholar]
- 5.Magalhaes B, Peleteiro B, Lunet N. Dietary patterns and colorectal cancer: systematic review and meta-analysis. Eur J Cancer Prev. 2012;21(1):15–23. doi: 10.1097/CEJ.0b013e3283472241. [DOI] [PubMed] [Google Scholar]
- 6.Maeda M, Nagawa H, Maeda T, Koike H, Kasai H. Alcohol consumption enhances liver metastasis in colorectal carcinoma patients. Cancer. 1998;83(8):1483–1488. doi: 10.1002/(sici)1097-0142(19981015)83:8<1483::aid-cncr2>3.0.co;2-z. [DOI] [PubMed] [Google Scholar]
- 7.Anton-Culver H. Smoking and other risk factors associated with the stage and age of diagnosis of colon and rectum cancers. Cancer Detect Prev. 1991;15(5):345–350. [PubMed] [Google Scholar]
- 8.Patel SA, Chen YH, Hornick JL, et al. Early-stage rectal cancer: clinical and pathologic prognostic markers of time to local recurrence and overall survival after resection. Dis Colon Rectum. 2014;57(4):449–459. doi: 10.1097/DCR.0b013e3182a70709. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Phipps AI, Shi Q, Limburg PJ, et al. Alcohol consumption and colon cancer prognosis among participants in North Central Cancer Treatment Group phase III trial N0147. Int J Cancer. 2016 doi: 10.1002/ijc.30135. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Phipps AI, Baron J, Newcomb PA. Prediagnostic smoking history, alcohol consumption, and colorectal cancer survival: the Seattle Colon Cancer Family Registry. Cancer. 2011;117(21):4948–4957. doi: 10.1002/cncr.26114. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Zell JA, McEligot AJ, Ziogas A, Holcombe RF, Anton-Culver H. Differential effects of wine consumption on colorectal cancer outcomes based on family history of the disease. Nutr Cancer. 2007;59(1):36–45. doi: 10.1080/01635580701413926. [DOI] [PubMed] [Google Scholar]
- 12.Pelser C, Arem H, Pfeiffer RM, et al. Prediagnostic lifestyle factors and survival after colon and rectal cancer diagnosis in the National Institutes of Health (NIH)-AARP Diet and Health Study. Cancer. 2014;120(10):1540–1547. doi: 10.1002/cncr.28573. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Newcomb PA, Baron J, Cotterchio M, et al. Colon Cancer Family Registry: an international resource for studies of the genetic epidemiology of colon cancer. Cancer Epidemiol Biomarkers Prev. 2007;16(11):2331–2343. doi: 10.1158/1055-9965.EPI-07-0648. [DOI] [PubMed] [Google Scholar]
- 14.Longnecker MP, Newcomb PA, Mittendorf R, et al. The reliability of self-reported alcohol consumption in the remote past. Epidemiology. 1992;3(6):535–539. doi: 10.1097/00001648-199211000-00014. [DOI] [PubMed] [Google Scholar]
- 15.Fritz A, Percy C, Jack A. International Classification of Diseases for Oncology (ICD-O) Geneva: World Health Organization; 2000. [Google Scholar]
- 16.Cicek MS, Lindor NM, Gallinger S, et al. Quality assessment and correlation of microsatellite instability and immunohistochemical markers among population- and clinic-based colorectal tumors results from the Colon Cancer Family Registry. J Mol Diagn. 2011;13(3):271–281. doi: 10.1016/j.jmoldx.2010.12.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Buchanan DD, Win AK, Walsh MD, et al. Family history of colorectal cancer in BRAF p.V600E-mutated colorectal cancer cases. Cancer Epidemiol Biomarkers Prev. 2013;22(5):917–926. doi: 10.1158/1055-9965.EPI-12-1211. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Weisenberger DJ, D Siegmund K, Campan M, et al. CpG island methylator phenotype underlies sporadic microsatellite instability and is tightly associated with BRAF mutation in colorectal cancer. Nat Genet. 2006;38(7):787–793. doi: 10.1038/ng1834. [DOI] [PubMed] [Google Scholar]
- 19.Campbell PT, Newton CC, Newcomb PA, et al. Association between body mass index and mortality for colorectal cancer survivors: overall and by tumor molecular phenotype. Cancer Epidemiol Biomarkers Prev. 2015;24(8):1229–1238. doi: 10.1158/1055-9965.EPI-15-0094. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Klarich DS, Brasser SM, Hong MY. Moderate Alcohol Consumption and Colorectal Cancer Risk. Alcohol Clin Exp Res. 2015;39(8):1280–1291. doi: 10.1111/acer.12778. [DOI] [PubMed] [Google Scholar]
- 21.Renaud S, de Lorgeril M. Wine, alcohol, platelets, and the French paradox for coronary heart disease. Lancet. 1992;339(8808):1523–1526. doi: 10.1016/0140-6736(92)91277-f. [DOI] [PubMed] [Google Scholar]
- 22.Imhof A, Woodward M, Doering A, et al. Overall alcohol intake, beer, wine, and systemic markers of inflammation in western Europe: results from three MONICA samples (Augsburg, Glasgow, Lille) Eur Heart J. 2004;25(23):2092–2100. doi: 10.1016/j.ehj.2004.09.032. [DOI] [PubMed] [Google Scholar]
- 23.Mukamal KJ, Jensen MK, Gronbaek M, et al. Drinking frequency, mediating biomarkers, and risk of myocardial infarction in women and men. Circulation. 2005;112(10):1406–1413. doi: 10.1161/CIRCULATIONAHA.105.537704. [DOI] [PubMed] [Google Scholar]
- 24.Hung LM, Chen JK, Huang SS, Lee RS, Su MJ. Cardioprotective effect of resveratrol, a natural antioxidant derived from grapes. Cardiovasc Res. 2000;47(3):549–555. doi: 10.1016/s0008-6363(00)00102-4. [DOI] [PubMed] [Google Scholar]
- 25.Kwan ML, Kushi LH, Weltzien E, et al. Alcohol consumption and breast cancer recurrence and survival among women with early-stage breast cancer: the life after cancer epidemiology study. J Clin Oncol. 2010;28(29):4410–4416. doi: 10.1200/JCO.2010.29.2730. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Chen Y, Tseng SH, Lai HS, Chen WJ. Resveratrol-induced cellular apoptosis and cell cycle arrest in neuroblastoma cells and antitumor effects on neuroblastoma in mice. Surgery. 2004;136(1):57–66. doi: 10.1016/j.surg.2004.01.017. [DOI] [PubMed] [Google Scholar]
- 27.Schneider Y, Duranton B, Gosse F, Schleiffer R, Seiler N, Raul F. Resveratrol inhibits intestinal tumorigenesis and modulates host-defense-related gene expression in an animal model of human familial adenomatous polyposis. Nutr Cancer. 2001;39(1):102–107. doi: 10.1207/S15327914nc391_14. [DOI] [PubMed] [Google Scholar]
- 28.Zhou HB, Chen JJ, Wang WX, Cai JT, Du Q. Anticancer activity of resveratrol on implanted human primary gastric carcinoma cells in nude mice. World J Gastroenterol. 2005;11(2):280–284. doi: 10.3748/wjg.v11.i2.280. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Wu SL, Sun ZJ, Yu L, Meng KW, Qin XL, Pan CE. Effect of resveratrol and in combination with 5-FU on murine liver cancer. World J Gastroenterol. 2004;10(20):3048–3052. doi: 10.3748/wjg.v10.i20.3048. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Barefoot JC, Gronbaek M, Feaganes JR, McPherson RS, Williams RB, Siegler IC. Alcoholic beverage preference, diet, and health habits in the UNC Alumni Heart Study. Am J Clin Nutr. 2002;76(2):466–472. doi: 10.1093/ajcn/76.2.466. [DOI] [PubMed] [Google Scholar]
- 31.Mullie P, Guelinckx I, Clarys P, Degrave E, Hulens M, Vansant G. Cultural, socioeconomic and nutritional determinants of functional food consumption patterns. Eur J Clin Nutr. 2009;63(11):1290–1296. doi: 10.1038/ejcn.2009.89. [DOI] [PubMed] [Google Scholar]