Zhu J, Jiang X, Niu Z. Alcohol consumption and risk of breast and ovarian cancer: A Mendelian randomization study. Cancer Genetics 2020;245:35–41
Background: Alcohol consumption has been found to increase the risk of breast cancer in observation studies, yet it remains unknown if alcohol is related to other hormone-dependent cancers such as ovarian cancer. No Mendelian randomization (MR) studies have been performed to assess a potential causal relationship between alcohol use and risk of breast and ovarian cancer.
Methods: We aim to determine if alcohol consumption is causally associated with the risk of female hormone-dependent cancers, by using summary level genetic data from the hitherto largest genome- wide association studies (GWAS) conducted on alcohol consumption (N = ~1.5 million individuals), breast (N case = 122,977) and ovarian cancer (N case = 25,509). We examined three different alcohol intake exposures, drinks per week (drinks/week), alcohol use disorder (AUD) and age-adjusted alcohol use disorder identification test (AUDIT-C), to reflect the general and harmful drinking behavior. We constructed updated and stronger instruments using ninety-nine drinks/week-related SNPs, nine AUD-related SNPs and thirteen AUDIT-C-related SNPs and estimated the causal relationship applying several two-sample MR methods.
Results: We did not find any evidence to support for a causal association between alcohol consumption and risk of breast cancer [OR drinks/week = 1.01 (0.85–1.21), P = 0.89; OR AUD = 1.04 (95%CI: 0.89–1.21), P = 0.62; OR AUDIT-C = 1.07 (0.90–1.28), P = 0.44]; neither with its subtypes including ER-positive and ER- negative breast cancer, using any of the three alcohol-related exposures. For ovarian cancer, however, we identified a reduced risk with alcohol consumption, where a borderline significance was found for AUDIT-C but not for drinks/week or AUC [OR drinks/week = 0.83 (0.63–1.10), P = 0.19; OR AUD = 0.92 (0.83–1.01), P = 0.08; OR AUDIT-C = 0.83 (0.71–0.97), P = 0.02]. The effect attenuated to null excluding SNPs associated with potential confounders [OR drinks/week = 0.81 (0.53–1.21), P = 0.31; OR AUD = 0.96 (0.78–1.18), P = 0.68; OR AUDIT-C = 0.89 (0.68–1.16), P = 0.38].
Conclusion: We do not find any compelling evidence in support for a causal relationship between genetically predicted alcohol consumption and risk of breast or ovarian cancer, consistent across three different alcohol-related exposures. Future MR studies validating our findings are needed, when large-scale alcohol consumption GWAS results become available.
Relation of reported alcohol intake to risk of breast cancer: Many observational studies have demonstrated a positive relation of alcohol consumption to the risk of breast cancer, even among women reporting no more than up to 1 drink/day. In 1998, Smith-Warner SA et al reported a linear increase of breast cancer risk: “For alcohol intakes less than 60 g/d (reported by >99% of participants), risk increased linearly with increasing intake; the pooled multivariate relative risk for an increment of 10 g/d of alcohol (about 0.75-1 drink) was 1.09 (95% confidence interval [CI] 1.04-1.13. The multivariate-adjusted relative risk for total alcohol intakes of 30 to less than 60 g/d (about 2-5 drinks) vs nondrinkers was 1.41 (95% CI 1.18-1.69).”
In 2012, Seitz et al meta-analyzed 113 studies with 122,091 cases and identified a modest but significant association between light drinking (≤12.5 g/day ethanol; ≤1 drink/day) and breast cancer (RR = 1.04, 95% CI: 1.02–1.07). While the relation of alcohol to cardiovascular disease demonstrates a “J-shape” that is non-linear, the relation of alcohol to breast cancer may be linear, as Seitz et al conclude that “A significant increase of the order of 4% in the risk of breast cancer is already present at intakes of up to one alcoholic drink/day. Heavy alcohol consumption, defined as three or more drinks/day, is associated with an increased risk by 40-50%.”
Dam et al evaluated the association between changes in alcohol consumption and breast cancer risk in a large prospective cohort of 21,523 postmenopausal women and found that women reporting between 1 and 2 drinks per day had a 13% higher risk of breast cancer when compared with women reporting no more than 1 drink/day. “Analyses modelling five year change in alcohol intake with cubic splines showed that women who increased their alcohol intake over the five year period had a higher risk of breast cancer and a lower risk of coronary heart disease than women with a stable alcohol intake.” For example, they found that women who increased their alcohol intake by seven drinks per week had a hazard ratio of breast cancer of 1.13 (95% confidence interval 1.03 to 1.23) compared to women with stable intake; for coronary heart disease, the hazard ratio was 0.89 (0.81 to 0.97) for those who increased their intake by the equivalent of 1 drink/day in comparison with women reporting stable intakes.
The Nurses’ Health Study was one of the first large epidemiological cohort studies to demonstrate an increase in risk of breast cancer to be associated with alcohol consumption; an updated summary report from that study was published by Chen et al in 2011. Those authors stated: “After controlling for cumulative alcohol intake, binge drinking, but not frequency of drinking, was associated with breast cancer risk. Alcohol intake both earlier and later in adult life was independently associated with risk.” They conclude: “In summary, our study provides a comprehensive assessment of the relationship between alcohol intake and breast cancer risk in terms of timing, frequency, quantity, and types of alcohol in a large prospective cohort with detailed information on breast cancer risk factors. We did find an increased risk at low levels of use, but the risk was quite small. However, an individual will need to weigh the modest risks of light to moderate alcohol use on breast cancer development against the beneficial effects on cardiovascular disease to make the best personal choice regarding alcohol consumption.” Other reports from the Nurses’ Health Study (Cao et al; Mostofsky et al) support their findings of a slight increase in breast cancer risk, but a lower risk of total mortality, for light drinkers when compared with non-drinkers or heavier drinkers. An overview on the topic of alcohol and breast cancer has just been published by Freudenheim.
Alcohol and ovarian cancer: There are fewer studies on ovarian cancer, and the association is generally found to be null. Genkinger et al conducted a pooled analysis of prospective cohort studies (529,638 individuals) and reported a non-significant association between moderate alcohol consumption and ovarian cancer risk. Yan-Hong et al, in a meta-analysis of 13 case-control studies with more than 5,000 cases, reported weak evidence of an association between alcohol consumption and risk of ovarian cancer. They state: “Alcohol consumption had little to no effect on ovarian cancer incidence when compared to non-drinkers (risk ratio [RR], 1.03; 95% CI, 0.96–1.10; P = 0.473). Similarly, low (RR, 0.96; 95% CI, 0.93–1.00; P = 0.059), moderate (RR, 1.08; 95% CI, 0.92–1.27; P = 0.333), and heavy (RR, 0.99; 95% CI, 0.88–1.12; P = 0.904) alcohol consumption was not associated with the risk of ovarian cancer.”
Kelemen et al conducted a pooled analysis using prospective cohorts of 17,155 individuals and reported a null association between > 3 drinks per day compared to none (OR=0.92, 95% CI=0.76-1.10). They state: “Total alcohol intake was not associated with all ovarian cancers. Among beverage types, a statistically non-significant decreased risk was observed among women who consumed >8 oz/d of wine compared to none (OR=0.83, 95% CI=0.68-1.01, P trend=0.08).” The authors conclude: “We found no evidence that recent moderate alcohol drinking is associated with increased risk for overall ovarian cancer.”
Comments of Forum members on the present MR paper: Forum member Erik Skovenborg provided an extensive summary and critique of the present study, stated to be the first MR study conducted to explore a putative causal relationship of alcohol consumption with breast and ovarian cancer. He wrote, “Using summary level genetic data from the hitherto largest genome-wide association studies (GWAS) conducted on alcohol consumption (N = ~1.5 million individuals), with 122,977 cases of breast cancer and 25,509 cases of ovarian cancer, Zhu et al examined three different alcohol intake exposures, drinks per week (drinks/week), alcohol use disorder (AUD) and age-adjusted alcohol use disorder identification test (AUDIT-C), to reflect general and harmful drinking behavior. These authors constructed updated and stronger instruments in a Mendelian randomization study, using as an instrumental variable (IV) of 99 SNPs found to be related to drinks/week, 9 SNPs related to alcohol use disorders (AUD), and 13 SNPs related to AUDIT-C-scores, and estimated the causal relationship applying several two-sample MR methods between genetic factors related to these three indices of alcohol consumption and the risk of breast cancer and ovarian cancer.
“These authors did not find any evidence to support a causal association between alcohol consumption and risk of breast cancer, neither was there a significant association with subtypes of breast cancer (including ER-positive and ER- negative). For ovarian cancer they identified a reduced risk with alcohol consumption, where a borderline significance was found for AUDIT-C [OR for AUDIT-C = 0.83 (0.71–0.97), P = 0.02], but not for drinks/week [OR drinks/week = 0.83 (0.63–1.10), P = 0.19] or AUC [OR for AUD = 0.92 (0.83–1.01), P = 0.08]. The effect attenuated to null excluding SNPs associated with potential confounders. Further, the authors give a precise description of phenotypic variation regarding alcohol intake, predicted by genetic variants. According to the three alcohol GWAS(s), 4.2% of phenotypic variance of drinks/week could be explained by the 99 index SNPs, 5.6% of AUD variance could be explained by the 9 index SNPs, and 6.8% of AUDIT-C variance could be explained by the 13 index SNPs.
“This paper is strengthened by the authors providing a calculation of statistical power, using a method proposed by Brion et al. For drinks/week, the study had 80% power to detect a causal effect of a relative 5.9% increase in breast cancer risk and 11.9% increase in ovarian cancer risk. Corresponding estimates for cancer subtypes (ER +, ER −, and serous) were 7.0%, 10.4% and 13.9% increase, respectively. For AUD, the minimal detectable effects for overall breast and ovarian cancer were 5.1% and 10.2% increase; and corresponding estimates for ER + , ER − and serous subtypes were 6.1%, 9.0% and 12% . For AUDIT-C, the minimal detectable effects for overall breast and ovarian cancer were 4.6% and 9.2% increase, and the relevant estimates for ER + , ER –, and serous subtypes were 5.5%, 8.1% and 10.9% increase, respectively. Further, the authors also report an analysis in which they excluded potential confounders (e.g., BMI, smoking, and education). Results did not alter substantially when they performed these analyses: [OR for drinks/week: 1.06 (0.82–1.38); for AUD: 1.23 (0.81–1.87); for AUDIT-C: 1.28 (0.91–1.81)], again not showing statistically significant results. Further, the paper is strengthened by the use of the same GWAS database (Liu et al) as was used by Larsson et al for their analysis of alcohol consumption and cardiovascular disease.”
Reviewer Skovenborg also noted that “Zhu et al found no evidence in support for an association between genetically predicted alcohol consumption and the risk of either breast or ovarian cancer. However, instead of underlining the positive news for drinking women with low breast cancer risk, the authors chose the following conclusion: ‘Our findings deliver an important public health message on that drinking alcohol does not seem to pose a protective effect on cancer risk.’ However, a protective effect of alcohol on breast cancer was never the issue. Even so, it is better to be told that a glass of wine will not protect you from breast cancer rather than be told that it increases your risk of breast cancer.”
Do self-reports of average alcohol intake reflect actual exposure to alcohol? Forum member Estruch provided an excellent discussion of potential reasons why self-reported average amounts of alcohol, as used in all observational studies, may be inaccurate. He wrote: “Inconsistencies among studies of alcohol and cancer may be due to several methodological issues. One main issue is systematic underreporting of alcohol drinkers, which could result in an increased risk of adverse outcomes from reportedly small amounts of alcohol. Measurement in urine of objective biomarkers of consumption of different alcoholic beverages may help to support data obtained by questionnaires.
“Another important issue is the alcohol drinking pattern. Several epidemiological studies have observed a significant positive relationship between binge drinking and incidence of certain cancers, such as breast cancer. I believe that the average amount of alcohol consumed may be less important in studies in which the main outcome is cardiovascular disease, but the pattern of alcohol consumption may be more important for the relation with cancer. Taking into account that binge drinking in women is associated with an increased incidence of breast cancer (White et al), the measurement of “alcohol intake per week” (used in several epidemiological studies and the present study using MR) did not exclude women with binge drinking, whereas for cancer studies peak blood levels of alcohol may be the important measure. In addition, in the present MR paper, the other measures of alcohol consumption (alcohol use disorder and AUDIC-C) that were used are more related to “heavy drinking” than “moderate drinking”. Thus, the authors did not truly tested the effects of moderate alcohol drinking on the incidence of breast and ovarian cancers.”
Reviewer Skovenborg noted in a paper from Denmark by Mørch et al that week-end drinking had a stronger effect on breast cancer risk than similar amounts when consumed throughout the week, as those authors stated: “For alcohol consumption above the intake most frequently reported (1-3 drinks/day), the risk of breast cancer is increased; a possible threshold in risk estimates was found for consumption above 27 drinks per week. The risk is minor for moderate levels but increases for each additional drink consumed during the week. Weekend consumption and binge drinking imply an additional increase in breast cancer risk.”
Forum member Estruch emphasized the importance of other lifestyle habits: “While genetic factors play a role, lifestyle habits may have even more important effects on the risk of many types of cancer. Consequently, many types of cancer might be prevented, or at least postponed, by favorable environmental conditions and healthy lifestyle. For example, some specific bioactive compounds from foods, such as polyphenols, n-3 fatty acids and monounsaturated fatty acids, exhibit preventive effects against tumors, whereas intake of alcohol or nitrosamines correlated with an increased risk of cancer. However, nowadays, rather than focusing on single nutrients, the assessment of dietary patterns might be a more adequate approach to analyze the connections between nutrition and cancer. The Mediterranean diet has demonstrated a powerful protective effect on both cardiovascular disease and cancer. A meta-analysis of randomized, cohort and case-control studies has concluded that higher adherence to a Mediterranean diet markedly reduces the incidence of breast cancer (Schwingshackl et al). Surprisingly, when effects of the different components of the diet were analyzed, moderate alcohol intake (mainly wine) exhibits a significant inverse association with cancer risk, suggesting a protective effect of moderate alcohol consumption against cancer. Zeng et al have shown that some vitamins (especially folates) are also important.”
Forum member Van Velden points out that “Human nutrition (including moderate wine/alcohol consumption) has been under scrutiny for a long time. To single out one factor could be misleading. Refined carbohydrate intake is much more ‘dangerous’ than moderate alcohol consumption. Rather, stop taking in sugar.” Regarding the risk of cancer, he adds: “We do not have the answers right now, but a responsible lifestyle including exercise, no smoking, weight control, and stress reduction is part of the best intervention to prevent cancer and other chronic and degenerative diseases.” Reviewer Teissedre agrees: “We can also emphasize that some synergy habits, bad or good, can be essential for cancer development. The effect of alcohol on cancer risk depends also of the context (such as alcohol with or without a meal) and nature of the other foods consumed (e.g., processed and red meat, aflatoxins, salt) and tobacco consumption. The average consumption of omega-6/omega-3 in habitual diets should also be considered.”
Reviewer Estruch commented on effects of different alcoholic beverages: “The type of beverage may be important since non-alcoholic compounds of beverages may equipoise the possible carcinogenic effect of alcohol. Polyphenols from alcoholic beverages (wine, beer and some spirits) may exert anti-carcinogenic effects that may influence the final result (induction of carcinogenesis vs protection against it).” In fact, a paper by Bessaoud & Daurés reported on the relation between alcohol intake and breast cancer risk among women in France: “Women drinkers who had an average consumption of less than 1.5 drinks per day had a lower risk (odds ratio [OR] 0.58, 95% confidence interval [CI] 0.34–0.97) when compared with nondrinkers. This protective effect was due substantially to wine consumption since the proportion of regular wine drinkers is predominant in our study population. Furthermore, women who consumed between 10 and 12 g/d of wine had a lower risk (OR 0.51; 95% CI 0.30–0.91) when compared with non–wine drinkers.”
Some limitations when applying the results of this study to women in the US: Forum member Parente warned that she has some reservations about the applicability of the results of this study to women in the US population. “While I applaud and am heartened by this attempt to examine large numbers of women using GWAS and MR, with results that do not favor alcohol consumption-associated increased risk for breast and ovarian cancers, more studies will be needed that further refine these methods in female and multi-ethnic populations to validate these findings and enhance their usefulness to clinical practice and public health.
“Specifically, the present study by Zhu et al analyzed nine and thirteen index SNPs associated with AUD and AUDIT-C, respectively, using data obtained in men. Correlation and statistical validity of the use of these SNPs in a female population is unclear. The 99 index SNPs for ‘drinks per week’, inadequately clarified beyond some discussion of ALDH2*2, were obtained from 1.2 million men and women of European ancestry. The weight and relevance of the data contributed by the male population in those determinations to female cancers are uncertain. Additionally, GWAS outcome data for breast and ovarian cases and controls were obtained from women of European ancestry. The collective relevance of the use of these data to the current US female population is unknown and a study limitation, as noted in part by the authors’ discussion. These uncertainties add to the challenge of interpreting these data in risk discussions with individual adult female patients encountered in current US practice [60% of whom are Caucasian, non-Hispanic, non-Latinx, according to US census data regarding ovarian cancer, a disease with a lifetime risk of 1 in 78 (1.3%), according to US cancer statistics].
“Further, if we consider only the ‘drinks per week’ exposure GWAS data, which were obtained from a mixed male/female European population, this study had ‘80% power to detect a causal effect of a relative 5.9% (i.e., ORs of 1.059 or more) increase in overall breast cancer risk and [an] 11.9% increase in overall ovarian cancer,’ i.e., an OR for ovarian cancer of 1.119 or more. Only looking at drinks/week, the authors found ‘a 17% decreased risk for drinks/week with overall ovarian cancer [OR (95%CI) = 0.83 (0.63–1.10)], although estimate did not reach statistical significance (P=0.19).’ While they noted a greater degree of risk reduction for the serous ovarian cancer subtype with other statistical methods, they also state ‘the effect attenuated to null in the sensitivity analysis.’”
Use of MR in studies of other types of cancer related to alcohol consumption: It is clear that heavy drinking increases the risk of many cancers. For example, Lewis & Smith aggregated summary results from seven studies with a total of 905 cases of esophageal cancer and provided strong evidence that the risk of esophageal cancer was lower among ALDH2 ∗2 ∗2 homozygotes (OR = 0.36; 95%CI = 0.16–0.80) as compared ALDH2 to ∗1 ∗1 homozygotes. Forum member Ellison notes: “The study of Lewis & Smith gives a clear indication that certain ALDH2 SNPs increase the risk of esophageal cancer, especially through their effects on alcohol consumption. The fact that their study showed a large effect on risk may be due to the fact that esophageal cancer is related especially to heavy alcohol intake, while the endocrine cancers studied by Zhu et al are very weakly related to alcohol and presumably related more to other factors.”
Continues Ellison, “Further, Williams has just published a paper on the heritability of alcohol intake in the Framingham Heart Study in which he reports that heritability of alcohol intake is greater for heavy drinkers than for light drinkers. For a disease especially related to very heavy consumption, such as esophageal cancer, a MR is more likely to show an effect from alcohol than a MR will for types of cancer that are not associated only with heavy drinking (such as breast cancer). Williams states: ‘The heritability of alcohol consumption, but not other nutrients, depends upon whether average consumption is high or low relative to the population distribution. Genetic heritability estimated from the offspring-parent regression slope [from two generations studied in the Framingham Study] increased significantly from lowest to highest gram/day of alcohol consumption (0.006 ± 0.001 per percent, P = 1.1 × 10−7). Heritability at the 90th percentile of the sample distribution was 4.5-fold greater than at the 10th percentile.’ Thus, MR studies for diseases associated more with light or moderate alcohol intake may not show genetic effects as much as diseases associated more with heavy drinking. Even though the authors in the present paper included some indices of heavy use, their MR analyses did not show much of an alcohol effect.”
A need to consider data from different types of studies when judging the “true effect” of any exposure: While some authors have considered that results from MR analyses are the only way to obtain a true, unbiased, estimate of risk associated with alcohol consumption, Lewis & Smith provide a very good discussion of appropriate applications of MR to various diseases, and urge us to use caution in their interpretation. In their paper from 2005, when few genetic factors related to alcohol intake were known, they stated: “In using ADLH2*2*2 as a surrogate for measuring alcohol intake, this study shows that alcohol drinking is related to elevated risk of esophageal cancer, thus illustrating the potential of the Mendelian randomization concept. However, this study also shows that ALDH2*1*2 is related to both lower alcohol consumption and elevated risk of esophageal cancer when the amount of alcohol consumed is identical; thus, analyzing the relationship between ALDH2*1*2 and risk of esophageal cancer without considering the amount of alcohol consumption is misleading. The result of ALDH2*1*2 illustrates a potential limitation of a Mendelian randomization approach in that in less well characterized situations, similar gene-environment interactions may be occurring which are not recognized, leading to spurious conclusions being drawn from looking at the main effects.”
Even though we now have identified many more genetic factors that we can include in MR analyses, it is clear that results from many types of studies must be considered when attempting to judge health effects of alcohol. This is especially the case because type of beverage, drinking patterns, smoking and other lifestyle habits, diet, and many other environmental factors relate to the effects of alcohol consumption. Thus, the combination of data from observational studies, clinical trials, animal experiments, and MR analyses will be needed to improve our knowledge on the relation of alcohol intake to health and disease; it remains a continuing challenge.
References from Forum Critique
Bessaoud F, Daurés JP. Patterns of Alcohol (Especially Wine) Consumption and Breast Cancer Risk: A Case-Control Study among a Population in Southern France. Ann Epidemiol 2008;18:467–475.
Brion MJ, Shakhbazov K, Visscher PM. Calculating statistical power in Mendelian randomization studies. Int J Epidemiol 2013;42:1497–1501.
Cao Y, Willett WC, Rimm EB, Stampfer MJ, Giovannucci EL. Light to moderate intake of alcohol, drinking patterns, and risk of cancer: results from two prospective US cohort studies. BMJ 2015;351:h4238 .
Chen WY, Rosner B, Hankinson SE, Colditz GA, Willett WC. Moderate alcohol consumption during adult life, drinking patterns, and breast cancer risk. JAMA 2011;306:1884–1890. doi:10.1001/jama.2011.1590.
Dam MK, Hvidtfeldt UA, Tjonneland A, Overvad K, Gronbaek M, Tolstrup JS. Five year change in alcohol intake and risk of breast cancer and coronary heart disease among postmenopausal women: prospective cohort study. BMJ 2016;353:i2314 .
Freudenheim JL. Alcohol’s Effects on Breast Cancer in Women. Alcohol Res 2020;40:11. https://doi.org/10.35946/arcr.v40.2.11.
Genkinger JM, Hunter DJ, Spiegelman D, et al. Alcohol intake and ovarian cancer risk: a pooled analysis of 10 cohort studies. Br J Cancer 2006;94:757–762 .
Kelemen LE, Bandera EV, Terry KL, et al. Recent alcohol consumption and risk of incident ovarian carcinoma: a pooled analysis of 5,342 cases and 10,358 controls from the Ovarian Cancer Association Consortium. BMC Cancer 2013;13:28.
Larsson SC, Burgess S, Mason AM, Michaelsson K. Alcohol consumption and cardiovascular disease: A Mendelian randomization study. Circ Genom Precis Med 2020;13:e002814. doi:10.1161/CIRGEN.119.002814
Lewis SJ, Smith GD. Alcohol, ALDH2, and esophageal cancer: a meta-analysis which illustrates the potentials and limitations of a Mendelian randomization approach. Cancer Epidemiol Biomarkers Prev 2005;14:1967–1971.
Liu M, Jiang Y, Wedow R, Li Y, Brazel DM, Chen F, Datta G, Davila-Velderrain J, McGuire D, Tian C, et al; 23andMe Research Team; HUNT All-In Psychiatry. Association studies of up to 1.2 million individuals yield new insights into the genetic etiology of tobacco and alcohol use. Nat Genet 2019;51:237–244.
Mørch LS, Johansen; D. Thygesen LC, Tjønneland A, Løkkegaard E, Stahlberg C, Grønbæk M. Alcohol drinking, consumption patterns and breast cancer among Danish nurses: a cohort study. European Journal of Public Health 2007;17:624–629.
Mostofsky E, Mukamal KJ, Giovannucci EL, et al. Key Findings on Alcohol Consumption and a Variety of Health Outcomes From the Nurses’ Health Study. Am J Public Health 2016;106:1586–1591. doi:10.2105/AJPH.2016.303336)
Schwingshackl L, Schwedhelm C, Galbete C, Hoffmann G. Adherence to Mediterranean Diet and Risk of Cancer: An Updated Systematic Review and Meta-Analysis. Nutrients 2017;9:1063; https://doi.org/10.3390/nu9101063
Seitz HK, Pelucchi C, Bagnardi V, LaVecchia C. Epidemiology and pathophysiology of alcohol and breast cancer: Update 2012. Alcohol Alcohol 2012;47:204–212.
Smith-Warner SA, Spiegelman D, Yaun S-S, et al. Alcohol and breast cancer in women: a pooled analysis of cohort studies. JAMA 1998;279:535–540.
US Census data: https://www.census.gov/quickfacts/fact/table/US/LFE046218
White AJ, DeRoo LA, Weinberg CR, Sandler DP. Lifetime Alcohol Intake, Binge Drinking Behaviors, and Breast Cancer Risk. Am J Epidemiol 2017;186:541-549.
Williams PT. Quantile‑Specific Heritability of Intakes of Alcohol but Not Other Macronutrients. Behavior Genetics 2020; pre-publication July 13, 2020. https://doi.org/10.1007/s10519-020-10005-z
Yan-Hong H, Jing L, Hong L, et al. Association between alcohol consumption and the risk of ovarian cancer: a meta-analysis of prospective observational studies. BMC Public Health 2015;15:223. DOI 10.1186/s12889-015-1355-8
Zeng J, Wang K, Ye F, Lei L, Zhou Y. Chen J. Zhao G, Chang H. Folate intake and the risk of breast cancer: an up-to-date meta-analysis of prospective studies. Eur J Clin Nutr 2019;73:1657–1660.
The authors of this ambitious study used a Mendelian Randomization (MR) approach in an attempt to determine if alcohol consumption is causally associated with the risk of female hormone-dependent cancers. They used summary level genetic data from the hitherto largest genome-wide association study (GWAS) conducted on alcohol consumption (N = ~1.5 million individuals), breast cancer (N cases = 122,977) and ovarian cancer (N cases = 25,509). They examined three different alcohol intake exposures: drinks per week, alcohol use disorder (AUD) and age-adjusted alcohol use disorder identification test (AUDIT-C), to reflect general and harmful drinking behavior. For instrumental variables (IVs), they constructed updated and stronger genetic instruments using 99 SNPs related to drinks/week, 9 SNPs related to AUD, and 13 AUDIT-C-related SNPs, and estimated the causal relationship applying several two-sample MR methods.
The key result of these analyses was that the investigators did not find any evidence that alcohol intake increased the risk of breast cancer. For ovarian cancer, an initial inverse association with alcohol intake became null when adjustments were made for confounders. While members of our Forum agreed with the overall conclusions of the authors, who stated that they did not find “any compelling evidence in support for a causal relationship between genetically predicted alcohol consumption and risk of breast or ovarian cancer, consistent across three different alcohol-related exposures,” Forum members identified a number of weaknesses of the analyses. For example, there were very limited data on the pattern of drinking or the type of beverage consumed, and no data on smoking and other lifestyle habits, diet, and many other environmental factors that relate to the effects of alcohol consumption. Further, much of the data utilized to generate the IVs of alcohol exposure were from men, rather than women, although the two conditions being studied are present only in women.
Forum members agreed with the authors about the need to combine data from MR analyses with data from observational and experimental studies before making conclusions about the relation of alcohol consumption to health and disease. Given the rapid expansion of genetic studies, it is likely that MR analyses will be increasingly important in our research efforts, but we agree that the consideration of a combination of data from observational studies, clinical trials, animal experiments, and MR analyses will be needed as we seek to improve our knowledge on the relation of alcohol intake to health and disease; it is, and will remain, a continuing challenge.
* * * * * * *
Contributions to this critique by the International Scientific Forum on Alcohol Research was provided by the following members:
Erik Skovenborg, MD, specialized in epidemiology and family medicine, member of the Scandinavian Medical Alcohol Board, Aarhus, Denmark
Ramon Estruch, MD, PhD, Hospital Clinic, IDIBAPS, Associate Professor of Medicine, University of Barcelona, Spain
R. Curtis Ellison, MD, Professor of Medicine, Section of Preventive Medicine & Epidemiology, Boston University School of Medicine, Boston, MA, USA
David Van Velden, MD, Dept. of Pathology, Stellenbosch University, Stellenbosch, South Africa
Pierre-Louis Teissedre, PhD, Faculty of Oenology–ISVV, University Victor Segalen Bordeaux 2, Bordeaux, France
Creina Stockley, PhD, MSc Clinical Pharmacology, MBA; Adjunct Senior Lecturer at the University of Adelaide, Australia
Harvey Finkel, MD, Hematology/Oncology, Retired (Formerly, Clinical Professor of Medicine, Boston University Medical Center, Boston, MA, USA)
Arne Svilaas, MD, PhD, general practice and lipidology, Oslo University Hospital, Oslo, Norway
Matilda Parente, MD, consultant in molecular pathology/genetics and emerging technologies, San Diego, CA, USA