Rovira P, Rehm J. Estimation of cancers caused by light to moderate alcohol consumption in the European Union. European Journal of Public Health 2021; pre-publication. doi:10.1093/eurpub/ckaa236.
Background: Research has identified alcohol to be an important risk factor for several types of cancers. This study estimates the number of incident cancers attributable to alcohol consumption in the European Union (EU) in 2017, with a special focus on those caused by light to moderate drinking levels.
Methods: The attributable-fraction methodology is used to estimate the number of new cancer cases in the year 2017 in the EU caused by alcohol use, and further examines those due to light to moderate drinking levels, defined here as alcohol consumption of <20 g of pure alcohol per day.
Results: Light to moderate drinking levels of alcohol caused almost 23 000 new cancer cases in the EU in 2017, and accounted for 13.3% of all alcohol-attributable cancers, and 2.3% of all cases of the seven alcohol-related cancer types. Almost half of these ( ̴ 11 000 cases) were female breast cancers. Also, more than a third of the cancer cases due to light to moderate drinking resulted from a light drinking level of <1 standard drink per day (total: 37%; women: 40%; men: 32%).
Conclusions: Alcohol use, including light to moderate drinking, continues to cause considerable cancer burden, and efforts should be made to reduce this burden. In addition to the alcohol control policies suggested by the World Health Organization, public information campaigns and the placement of warning labels on alcohol containers advising of the cancer risk associated with alcohol use should be initiated to increase knowledge about the alcohol-cancer link.
Forum member Waterhouse, like several other Forum members, had difficulty trying to follow the complex definitions and categorization of subjects in this paper, stating: “It seems to be organized to obfuscate.” Other reviewers were not sure how the authors were categorizing “alcohol-attributable” cases of cancer. Further, many also had concerns about the combining of data from so many diverse groups into one estimate: to whom do the results really apply?
Reviewer Ellison also had some concerns about this study: “I note that the authors include daily consumption of up to 20 grams of alcohol, or up to about 2 typical drinks, as being ‘light-moderate’ for both men and women. For women, this is twice what the US Guidelines say, which is no more than 1 drink/day. Also, the majority of the alcohol-attributed cases were from the relation of alcohol to the risk of breast cancer, where the influence of alcohol is slight but the number of women with the disease is great. The authors report in the Results that ‘the majority of all cancer cases due to light to moderate drinking was in the breast cancer category, 68.7% of all alcohol attributable cancer for both sexes.’ If adjustments had been made for numerous factors that modify the alcohol-breast cancer relation (as discussed in detail below by Skovenborg) the impact would probably have been considerably smaller.
Forum member Stockley wrote: “The only exposure variable considered in the study was the average consumption of alcohol, in grams per day of ethanol, with no data on the type of alcoholic beverage, the pattern of alcohol consumption, or many unmeasured cultural differences among the countries that were included in the study’s analyses.
“In addition, the relationship between alcohol and each cancer is different and is influenced by other factors such as diet and exercise, genetic predisposition and tobacco smoking. The role of different drinking patterns may also modify the effect of the total amount of alcohol consumed and, therefore, risk (Bagnardi, et al). For example, it has also been suggested that the risk of developing a cancer of the aero-digestive tract is less if you consume alcohol with food (Dal Maso, et al), while risk is increased among current tobacco smokers (Allen, et al). Further, only approximately 10 to 15% of alcohol dependent drinkers develop cirrhosis of the liver and, of those, only 10% develop liver cancer (Hall). Similarly, not all heavy drinkers develop cancer and some light-to-moderate drinkers develop cancer; this suggests that an individual’s genetic predisposition influences their risk of developing cancer. This has also been suggested for colorectal cancers, such that there is a stronger relationship between alcohol consumption and colorectal cancers among individuals with a family history of colorectal cancer (Cao, et al).
Confounders and effect modifiers of the relation of alcohol to cancer: Reviewer Skovenborg provided an extensive review of scientific data that relate to bias, confounding, and effect- modifying factors that need to be taken into consideration when relating light to moderate alcohol consumption to the risk of cancer. He also provides a number of references as examples of studies upon which such conclusions are based.
“Body mass index: Shin, et al noted in their paper on the association of alcohol and breast cancer risk: ‘Overall, we found no statistically significant association between alcohol intake and breast cancer risk after adjustment for confounding, with an estimated relative risk (RR) of 1.01 (95 % CI: 0.98–1.04) for an increment in alcohol consumption of 5 g/day. A statistically significant 3% elevated breast cancer risk associated with higher alcohol consumption was found only among women with BMI ≤25 (RR 1.03, 95 % CI 1.0–1.05 per 5 g/day increase)’.
“Folate intake: Analysis of 39 studies on dietary folate intake and 12 studies on plasma folate level by Ren, et al found folate intake inversely correlated with the breast cancer risk when the highest and lowest categories (OR = 0.85, 95% CI = 0.79–0.92) were compared, and the dose-response result showed that folate intake had an inverse linear correlation with the breast cancer risk. A higher folate intake correlated with a lower breast cancer risk in premenopausal women (OR = 0.80, 95% CI = 0.66–0.97), but not in postmenopausal women (OR = 0.94, 95% CI = 0.83–1.06). Plasma folate levels were not correlated with the breast cancer risk (OR = 0.98, 95% CI = 0.82–1.17).
Diet: Buga, et al found that a higher intake of total meat, or red or processed meats, or foods with a high glycemic index, or eggs seemed to be associated with a higher risk of breast cancer. Vegetables seemed to have an inverse association with cancer risk. A study of association between healthy lifestyle score and breast cancer found that individuals with the highest Healthy Eating Index (HEI) score were 46% less likely to have breast cancer than those with the lowest score (OR: 0.54; 95% CI: 0.35, 0.82) (Ghosn, et al). A randomized trial of a long-term dietary intervention on breast cancer incidence found a beneficial effect of a Mediterranean diet supplemented with extra-virgin olive oil in the primary prevention of breast cancer (Toledo, et al).
Beverage choice: Overall the epidemiologic evidence suggest that it is the alcohol content – not the type of beverage – that drives breast cancer risk. In a review of large observational cases, Teissedre, et al found that alcohol consumption or alcohol misuse related to cancers of the upper aerodigestive tract, liver, colorectum, breast, pancreatic, and prostate; they found no evidence of ovarian, gastric, head and neck, and lung cancer being linked to the moderate consumption of alcoholic beverages. While total alcohol seems to be related to breast cancer, they found insufficient data to affirm that moderate red wine consumption during meals in the frame of a healthy lifestyle is associated with an increased risk of breast cancer.
Drinking pattern: In a study by Mørch, et al, the relative risk of breast cancer was 2.30 [CI: 1.56–3.39] for alcohol intake of 22–27 drinks per week, compared to 1–3 drinks per week. Binge drinking of 4–5 drinks on a weekday increased risk by 55%, compared with consumption of one drink. Women in the binge drinking group in the SUN Project showed a higher risk of breast cancer (HR = 1.76; 95% CI: 1.03–2.99) compared to women in the non-binge drinking category, even after controlling for total alcohol consumption (Sánchez-Bayona, et al).
Physical activity: From a study by Feng, et al, among subjects with a physical activity score (PA) that was ≤ 7.5 MET-hour/week, there was a significant association between alcohol consumption and risk of alcohol-related cancer mortality: exdrinkers (HR =1.53, 95% CI= (1.11, 2.12)), drinkers at hazardous levels (HR=1.47, 95% CI= (1.07, 2.02)) and harmful levels (HR=1.64, 95% CI= (1.07, 2.52)) had significantly higher mortality risks than never-drinkers. However, the increased risks from alcohol were eliminated among individuals with a PA > 7.5 MET-hour/week individuals.
Smoking: Analysis of data from the Nurses’ Health Study and the Health Professionals Follow-up Study found that light to moderate drinking was associated with minimally increased risk of overall cancer. They reported that for men who had never smoked, risk of alcohol related cancers was not appreciably increased for light and moderate drinking (up to 29.9 g alcohol/day). For women who had never smoked, risk of breast cancer increased within the range of up to 5 – 14.9 g alcohol a day (Cao, et al). In the Million Women Study of cancers of the upper aerodigestive tract, the alcohol-associated risk was confined to current smokers, with little or no effect of alcohol among never and past smokers (Allen, et al).
Underreporting of alcohol intake: The apparent increased risk of cancer among light–moderate drinkers may be substantially due to underreporting of intake, as described by Klatsky, et al. In their very large Kaiser-Permanente study, among subjects reporting 1 or 2 drinks/day in one survey, they found a group of subjects who had other data in their massive datasets indicating excessive drinking or an alcohol use disorder (they were subsequently categorized as ‘probable under-reporters of alcohol’ on the current assessment). Others of the subjects reporting 1 to 2 drinks/day whose extensive data had no evidence of problems with alcohol were categorized as ‘probable not under-reporters.’ When relating their alcohol to risk of cancer of any type, only the under-reporters showed a positive relation (HR = 1.33 (1.21-1.45); for the group categorized as ‘not under-reporters,’ the hazard ratio for any cancer, versus abstainers, was 0.98 (0.87-1.09).
Cancer screening: In a report by Mu & Mukamal, moderate alcohol consumers were 84.7% more likely to report mammography screening than non-consumers. Among binge consumers the weighted prevalence of screening was lower than in non-binge consumers (for binge vs non-binge moderate consumers, 80.5 vs 85.5%). Given the likelihood of over-diagnosis, the increased risk of incident breast cancer in moderate alcohol consumers may be overestimated by more intensive screening.
The conclusion of Skovenborg’s assessment, shared by most other Forum members, is that there are certainly too many potential confounders and modifiers to use alcohol-attributable data from diverse populations to try and prove causality for any relation. Further, while many studies have shown a small increase in incidence of breast cancer but not mortality from breast cancer to be associated with moderate alcohol intake (discussed below), we have located 34 separate papers published since 1983 reporting no statistically significant relation between alcohol consumption and breast cancer in women. Two of the most recent large studies were the Prospective cohort Nurses’ Health Study II (93,835 US women), which reported that alcohol consumption was not associated with breast cancer risk overall for intake of ≥10 g/day vs. nondrinking (HR = 1.07 [0.94–1.22]) (Kim, et al). However, a positive association between alcohol consumption and breast cancer was found among women with a family history and folate intake < 400 µg/day (HR = 1.82 [1.06–3.12]) Also, Li, et al reported in 2020 that for both BRCA1 and BRCA2 mutation carriers, alcohol consumption was not associated with breast cancer risk; the study involved a retrospective analysis of 5,707 BRCA1 mutation carriers and 3,525 BRCA2 mutation carriers and a prospective analysis of 2,276 BRCA1 mutation carriers and 1,610 BRCA2 mutation carriers. There may well be a small effect of moderate drinking on breast cancer, but that effect seems to be grossly exaggerated in the analyses in the present paper, as that association provides the vast majority of cases used to estimate ‘alcohol-attributable cancer cases’ for both men and women.
Breast-cancer specific mortality related to alcohol consumption: As for mortality among women with breast cancer, Zeinomar, et al reported the following: “Compared to non-drinkers, we found that both low and moderate to heavy levels of alcohol intake were not [significantly] associated with greater overall mortality (≤3 drinks/week: HRO: 0.66, 95% CI: 0.38–1.14) or breast cancer–specific mortality (≤ 3 drinks/week: HRBC:0.62, 95% CI: 0.19–2.03). For subjects reporting >3 drinks/week, HR for total mortality was 1.16 (CI: 0.85–1.58) and for breast-cancer specific mortality it was 0.96 (CI: 0.49–1.89).”
Similarly, among more than 4,500 women with invasive breast cancer followed by Ma, et al, 1,055 women died, 824 from breast cancer, during an 8.6 year follow up. These authors report: “Average weekly alcohol consumption from age 15 years until breast cancer diagnosis was inversely associated with breast cancer-specific mortality (Ptrend = 0.01). Compared to non-drinkers, women in the highest average weekly alcohol consumption category (≥7 drinks/week) had 25% lower risk of breast cancer-specific mortality (HR = 0.75, 95% CI = 0.56–1.00).” They conclude: “Overall, we found no evidence that alcohol consumption before breast cancer diagnosis increases subsequent risk of death from breast cancer.”
While many studies have found no significant effects of alcohol on cancer mortality, some actually suggest a lower risk of death from cancer to be associated with moderate drinking. For example, Reding, et al evaluated drinking and mortality among 1,286 young women with breast cancer and reported that “Women who consume alcohol before a diagnosis of breast cancer have improved survival, which does not appear to be attributable to differences in stage, screening, or treatment.” Further, a paper from the Nurses’ Health Study of 22,890 women with incident invasive breast cancer reported 7,780 total deaths, including 3,484 resulting from breast cancer, during 11.3 years of follow up (Newcomb, et al). These investigators reported that pre-diagnostic moderate alcohol consumption “. . . was modestly associated with disease-specific survival (compared with nondrinkers, HR = 0.93 [95% CI, 0.85 to 1.02], 0.85 [95% CI, 0.75 to 0.95], 0.88 [95% CI, 0.75 to 1.02], and 0.89 [95% CI, 0.77 to 1.04] for two or more, three to six, seven to nine, and ≥ 10 drinks/wk, respectively).” These authors concluded: “Women consuming moderate levels of alcohol, either before or after diagnosis, experienced better cardiovascular and overall survival than nondrinkers.” There was no effect on mortality for alcohol consumption after the diagnosis of invasive breast cancer was made.
A study from the Women’s Health Initiative included data from 7,835 women diagnosed with breast cancer. The study concluded: “Women who were consuming alcohol prior to their breast cancer diagnosis had a non-statistically significant 24% (95% CI, 0.56–1.04) reduced risk of breast cancer specific-mortality and a 26% significant reduced risk (95% CI, 0.61–0.89) of all-cause mortality (Lowry, et al).
In contrast to the studies described above, Jain, et al evaluated alcohol intake among 223 Canadian women with death from breast cancer and reported “The hazard ratios for the risk of death from breast cancer increased with intakes of total alcohol of 10–20 g/day (1.039, 1.009–1.071) and > 20 g/day (1.063, 1.029–1.098).” They concluded: “Although our results were statistically significant, the magnitude of the change in risk was small.” Vreiling, et al, in a cohort study of 2,522 post-menopausal women in Germany with breast cancer, reported an increased risk of breast cancer-specific mortality for women consuming ≥12 vs. <0.5 g/day, but concluded: “Our findings show that consumption of alcohol before diagnosis is non-linearly associated with increased breast cancer-specific mortality but may be associated with decreased risk of mortality due to other causes.”
A key factor in judging causality of any outcome, including cancer, is reproducibility. It will be important that future studies designed to evaluate light to moderate drinking for effects on cancer and mortality provide assessments of alcohol intake that include adjustments for important confounders and modifiers of alcohol’s effect. Forum member Mattivi pointed out some important observations on this topic published by the team of John Ioannidis in their review on “research reproducibility.” They include the following: “Multiplicity, combined with incomplete reporting, might be the single largest contributor to the phenomenon of non-reproducibility, or falsity, of published claims. Scientific fields that routinely work with multiple hypotheses without correcting for or reporting the occurrence of multiplicity run a higher risk of non-reproducibility of results or inferences (Goodman, et al).”
Overall, we interpret current scientific data to show that moderate alcohol intake may slightly increase the diagnosis of breast cancer, but has little (or no) effect on mortality from the disease. In fact, in the larger studies such as the Nurses’ Health Study and the Women’ Health Initiative, there is even a tendency for lower risk of breast cancer-specific mortality, and essentially all studies show reductions in all-cause mortality from moderate drinking, mainly through beneficial effects on cardiovascular disease. We look forward to future studies of moderate alcohol consumption and cancer risk, evaluating the amount and pattern of drinking with adjustments for multiple confounders and modifiers of effect.
Using population–attributable risk calculations to determine impact of the exposure. Forum member Ellison noted: “In general, I am often concerned about papers that use the attributable-fraction values produced by the WHO (based on the combination of data from many divergent countries and cultures) for estimating new cases of cancer; such methodology lumps everyone together into a single estimate, and does not take under-reporting of intake by heavy drinkers into account. (And it has been shown that the vast majority of truly alcohol-attributable cases of cancer, such as the upper aero-digestive ones, relate primarily to heavy consumption).”
Forum member Zhang stated: “We realize that population attributable risk % is not a causal effect measure: it is a measure of impact of exposure on a population. It varies by relative risk (a causal measure) and prevalence of the exposure. In general, assessing population attributable risk percent (PAR%) of a specific risk factor on a disease is a challenging task. First, to estimate PAR%, we assume that the risk factor, here alcohol consumption, is causally related to the risk of cancer. I am not sure this issue has been settled yet. Second, obtaining a precise causal estimate (i.e., relative risk, RR) of alcohol (here light-to moderate alcohol) on the risk for each cancer is a pre-requisite of calculating PAR%. I do not know how authors obtained this effect estimate (RR) for each of the different countries.
“I noted, as did others, the very striking similarities between the percentage of alcohol attributable cancer cases (for all types of cancer) reported for consumers reporting < 10 g/day and for those reporting 10-20 g/day. Although I admit that it could happen because PAR% is a function of both RR and prevalence of the exposure of interest, such similarity in PAR% of all of the types of cancer among subjects reporting <10g/day of alcohol intake and those reporting 10-20g/day intake seems unusual.” Added reviewer Djoussé: “I agree that that PAR% is conditioned on correct assumptions. As such, there is no way to make a strong inference since many assumptions must be made and could be off; they cannot easily be verified, and this includes determining causality.”
Thus, it appears that the authors of this paper have essentially constructed average relationships across a range of countries, with the weight for each country being effectively the sample sizes in the studies for that country. So what population would this average value relate to? Actually, the answer is a non-existent hypothetical population; different countries have different risk relationships, so what is the relevance of this average value to each country in the EU or to populations around the world?
Focusing only on the effects of alcohol on cancer: Many Forum members have noted the increasing focus of certain scientists on only the relation of alcohol consumption to cancer, tending to ignore the much more substantiated protective effects of moderate drinking against coronary heart disease and total mortality. Reviewer de Gaetano wrote: “Alcohol consumption has been inconsistently associated with cancer risk in observational studies. Mendelian randomization (MR) has been more recently used in support of a causal relationship between alcohol consumption and several site-specific cancers. However, a recent study by Larsson, et al found, by MR, no evidence supporting a relationship between alcohol consumption and overall cancer, or the risks of 19 different site-specific cancers. Another recent study (Zhu, et al) did not find, by MR, 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. I wonder why data showing a protective effect of alcohol are not accepted by many as deriving from observational studies, while the negative effects, also based on the same type of studies, are accepted and even presented as showing a causal link. I also wonder why data on all-cause disease/mortality are not presented together with data only on a single disease.”
How large is the true effect of moderate alcohol intake on cancer risk? The authors state that 2.3% of new cases of common cancers (those included in the alcohol-related cancer category) in Europe are “caused by alcohol.” They do not point out that this is a very small percentage of cancers. Then they state separately that alcohol causes 13.3% of “all alcohol-attributable cancers” without clarifying very well what their two separate categories really mean.
When mortality from cancer is evaluated, there are generally very small effects on risk among light or moderate drinkers. For example, Breslow et al, in an analysis based on the National Health Survey in the USA of more than 300,000 subjects with 8,362 cancer deaths, pointed out that, overall, the role of light or moderate total alcohol intake is a relatively small part of cancer deaths related to lifestyle factors. These authors reported that, in comparison with lifetime abstainers, the risk of all-cancer mortality for current light drinkers (<3 drinks/week) at baseline was not increased, with a RR of 0.87 (95% CI 0.80-0.94); for baseline moderate drinkers (> 3 to 7 drinks/week for women and >3 to 14 drinks/week for men) it was 0.97 (CI 0.87-1.06). Among all current alcohol drinkers at baseline, only heavier drinkers had a statistically significant increased risk of subsequent total cancer mortality when compared with lifetime abstainers.
Bergmann, et al evaluated lifetime alcohol use among 111,953 men and 268,442 women from eight countries participating in the European Prospective Investigation into Cancer and Nutrition (EPIC) study. Self-reported alcohol consumption at ages 20, 30, 40 or 50 years and at enrollment were used for the analysis; 26,411 deaths were observed during an average of 12.6 years of follow up. The authors state: “Compared with men and women who used alcohol lightly throughout their lifetime, those who used it in restricted amounts (i.e., below the recommended limits of ≤ 12 g/day for women or ≤ 24 g/day for men throughout their lifetime) had the lowest relative risk of overall death, whereas those who used alcohol heavily, either at enrolment or in the past, had the highest relative risk of death.” The authors point out further that the lower rates among moderate drinkers related primarily to a lower risk of cardiovascular disease, but there were no statistically significant increases in risk among such drinkers for cancer, respiratory causes, digestive cancers, external causes, or other causes.
Overview of the present paper: Forum member Finkel stated: “The arguments presented in this paper are quite familiar: just read any number of the senior author’s previous, and I’d guess, future publications. The cited references are weighted toward his views. Some refer to issues long since laid to rest. As an oncologist, I have continued to share skepticism about ‘alcohol-attributable cancers.’ From the beginning, this was a convenient ill-based dumping ground enabling, especially in the case of light-to-moderate drinking, those with an axe to grind. The text reads more like an editorial, polemics rather than objective data.”
Ellison agreed, stating: “It is interesting that the authors do not provide estimates of total mortality; it would probably suggest that if everyone currently drinking light to moderate amounts of alcohol were to stop drinking completely because of the danger of cancer, population rates of total mortality would increase. Heavy alcohol use is recognized by all as a serious impediment to the health of individuals, and to society at large. However, recommending lower intake of alcohol for everyone does not take into consideration that light, in contrast to heavy, drinkers might be more likely to follow such advice and might decrease their consumption to zero, which would be expected to increase, rather than decrease, total mortality rates and the health of the population.”
Reviewer van Velden wrote: “I just want to add that focusing on only one potential risk factor for disease or mortality is not a good idea. We should be focusing on all components of a responsible lifestyle. This would not only include being moderate in any alcohol consumption, but also weight management, exercise, no smoking and effective stress management. To single out only one component of the exposure or the outcome is misleading.”
In their conclusions, the authors only mention briefly the results of their analyses, then start the usual chant of the need for extensive public health measures (warning labels, public information campaigns, etc.) to spread the word on the dangers of any alcohol consumption. However, the analyses in this paper do not include data that are necessary if one is to set appropriate drinking guidelines (that should be population-specific, age-specific, sex-specific, etc.). The necessary elements for developing guidelines for a population have been well described more than a decade ago by Harding & Stockley.
References from Forum critique
Allen NE, Beral V, Casabonne D, Kan SW, Reeves GK, Brown A, Green J. Moderate alcohol intake and cancer incidence in women. Journal of the National Cancer Institute 2009;101:296-305.
Bagnardi V, Zatonski W, Scotti L, La Vecchia C, Corrao G. Does drinking pattern modify the effect of alcohol on the risk of coronary heart disease? Evidence from a meta-analysis. Journal of epidemiology and community health 2008;62:615-619.
Bergmann MM, Rehm J, Klipstein-Grobusch K, Boeing H, Schutze M, et al. The association of pattern of lifetime alcohol use and cause of death in the European Prospective Investigation into Cancer and Nutrition (EPIC) study. Inter J Epidemiol 2013;42:1772–1790. doi:10.1093/ije/dyt154
Breslow RA, Chen CM, Graubard BI, Mukamal KJ. Prospective Study of Alcohol Consumption Quantity and Frequency and Cancer-Specific Mortality in the US Population. Am J Epidemiol 2011;174:1044–1053.
Buja A, Pierbon M, Lago L, Grotto G, Baldo V. Breast Cancer Primary Prevention and Diet: An Umbrella Review. Int J Environ Res Public Health 2020;17:4731.
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. doi: 10.1136/bmj.h4238
Dal Maso L, La Vecchia C, Polesel J, Talamini R, Levi F, Conti E, Zambon P, Negri E, Franceschi S. Alcohol drinking outside meals and cancers of the upper aero-digestive tract. International journal of cancer 2002;102:435-437.
Feng Y, Powell L, Vassallo AJ, Hamer M, Stamatakis E. Does adequate physical activity attenuate the associations of alcohol and alcohol-related cancer mortality? A pooled study of 54 686 British adults. Int J Cancer 2020;147:2754-2763.
Ghosn B, Benisi-Kohansal S, Ebrahimpour-Koujan S, et al. Association between healthy lifestyle score and breast cancer. Nutr J 2020;19:4. https://doi.org/10.1186/s12937-020-0520-9
Goodman SN, Fanelli D, Ioannidis JPA. Scientific Integrity. What does research reproducibility mean? Sci Transl Med 2016;8:341ps12.
Hall P. Pathological spectrum of alcoholic liver disease. Hall, P. (ed.) Alcoholic liver disease: pathobiology and pathogenesis. 2nd edition ed. 1995. London: Edward Arnold: pp. 41-68.
Harding R, Stockley CS. Communicating through government agencies. Ann Epidemiol 2007;17:S98-S102.
Jain MG, Ferrence RG, Rehm JT, Bondy SJ, Rohan TE, Ashley MJ, Cohen JE, Miller AB. Alcohol and Breast Cancer Mortality in a Cohort Study. Breast Cancer Research and Treatment volume 64, pages 201–209 (2000).
Kim HJ, Jung S, Eliassen AH, Chen WY, Willett WC, Cho E. Alcohol Consumption and Breast Cancer Risk in Younger Women According to Family History of Breast Cancer and Folate Intake. Am J Epidemiol 2017;186:524-531. https://doi.org/10.1093/aje/kwx137
Klatsky AL, Udaltsova N, Li Y, Baer D,Tran HN, Friedman GD. Moderate alcohol intake and cancer: the role of underreporting. Cancer Causes Control 2014;25:693-699.
Larsson SC, Carter P, Kar S, Vithayathil M, Mason AM, Michaëlsson K, Burgess S. Smoking, alcohol consumption, and cancer: A mendelian randomisation study in UK Biobank and international genetic consortia participants. PLoS Med 2020;23:17:e1003178. doi: 10.1371/journal.pmed.1003178.
Li H, Terry MG, Antoniou AC, et al. Alcohol Consumption, Cigarette Smoking, and Risk of Breast Cancer for BRCA1 and BRCA2 Mutation Carriers: Results from The BRCA1 and BRCA2 Cohort Consortium. Cancer Epidemiol Biomarkers Prev 2020;29:368-378.
Lowry SJ, Kapphahn K, Chlebowski R, Li CI. Alcohol Use and Breast Cancer Survival among Participants in the Women’s Health Initiative. Cancer Epidemiol Biomarkers Prev 2016;25;1268–1273.
Ma H, Malone KE, McDonald JA, et al. Pre-diagnosis alcohol consumption and mortality risk among black women and white women with invasive breast cancer. BMC Cancer 2019;19:800. https://doi.org/10.1186/s12885-019-5991-8.
Mørch LS, Johansen D, Thygesen LC, Tjønneland A, Løkkegaard E, Stahlberg C, Grønbaek M. Alcohol drinking, consumption patterns and breast cancer among Danish nurses: a cohort study. European Journal of Public Health 2007;17:624-629.
Mu L, Mukamal KJ. Alcohol consumption and rates of screening: Is cancer risk overestimated? Cancer causes Control 2016;27:281-289.
Newcomb PA, Kampman E, Trentham-Dietz A, Egan KM, Titus LJ, Baron JA, Hampton JM, Passarelli MN, Willett WC. Alcohol Consumption Before and After Breast Cancer Diagnosis: Associations With Survival From Breast Cancer, Cardiovascular Disease, and Other Causes. Journal of Clinical Oncology 2013;31:1939-1946.
Reding KW, Daling JR, Doody DR, O’Brien CA, Porter PL, Malone KE. Effect of Prediagnostic Alcohol Consumption on Survival after Breast Cancer in Young Women. Cancer Epidemiol Biomarkers Prev 2008;17:1988-1996). DOI: 10.1158/1055-9965.EPI-07-2897.
Ren Z, Xu P, Zhang D, Liu K, Song D, Zheng Y, et al. Association of folate intake and plasma folate level with the risk of breast cancer: a dose-response meta-analysis of observational studies. Aging 2020;12:21355-21375.
Sánchez-Bayona R, et al. Binge Drinking and Risk of Breast Cancer: Results from the SUN Project. Nutrients 2020;12:731.
Shin A, Sandin S, Lof M, Margolis KL, Kim K, Couto E, et al. Alcohol consumption, body mass index and breast cancer risk by hormone receptor status: Women’ Lifestyle and Health Study. BMC Cancer 2015;15:881.
Teissedre P-L, Rasines-Perea Z, Ruf J-C, Stockley C, Antoce AO, Romano R, Fradera U, Kosti RI. Effects of alcohol consumption in general, and wine in particular, on the risk of cancer development: a review. OENO One 2020;4:813-832.
Toledo E, Salas-Salvadó J, Donat-Vargas C, Buil-Cosiales P, Estruch R, et al. Mediterranean Diet and invasive breast cancer among women at high cardiovascular risk in the PREDIMED Trial: A Randomized Clinical Trial. JAMA Intern Med 2015;175:1752-1760.
Vrieling A, Buck K, Heinz J, Obi N, Benner A, Flesch-Janys D, Chang-Claude J. Pre-diagnostic alcohol consumption and postmenopausal breast cancer survival: a prospective patient cohort study. Breast Cancer Research and Treatment volume 136, pages195–207 (2012).
Zeinomar N, Thai A, Cloud AJ, McDonald JA, Liao Y, Terry MB. Alcohol consumption and breast cancer-specific and all-cause mortality in women diagnosed with breast cancer at the New York site of the Breast Cancer Family Registry. PLoS One 2017;12: e0189118. doi: 10.1371/journal.pone.0189118.
Zhu J, Jiang X, Niu Z. Alcohol consumption and risk of breast and ovarian cancer: A Mendelian randomization study. Cancer Genet 2020;245:35-41. doi:10.1016/j.cancergen.2020.06.001.
The present paper attempts to use alcohol-attributable data, based on combining average alcohol intake from diverse European populations, to judge the effects of moderate alcohol consumption on the risk of cancer for all populations. It is based on earlier calculations of risk, extensively covered in previous publications by the senior author. The investigators conclude that the risk of a number of types of cancer are increased even among subjects reporting no more than 20 grams of alcohol per day (up to about 2 typical drinks), an amount they state to be within recommended levels of consumption (but exceed current US Drinking Guidelines for women of no more than 1 drink/day). The largest contribution to their calculated effects of alcohol was from breast cancer in women (which made up “68.7% of all alcohol attributable cancer for both sexes”). Breast cancer is a condition for which many factors other than just average alcohol intake are important: factors such as pattern of drinking, type of beverage, folate intake, smoking, obesity, physical activity, etc., all of which tend to confound the association.
Forum members had real concerns about their methodology of using an estimate of “alcohol-attributable cancer” based on reported average intake from 28 separate countries, including the large countries in western Europe (e.g., Germany, France, UK, Italy, Spain, Sweden), East-European countries (such as Poland and Hungary), and many smaller countries (e.g., Malta, Cyprus, Luxembourg, Latvia). Given the marked variation in culture and drinking patterns among these countries, it was considered that there are a large number of confounders and modifiers of alcohol effects that vary by country, and make it impossible to simply take the average alcohol intake from such diverse cultures to estimate causality of alcohol and cancer for everyone. By lumping data from all of these countries, the analyses end up with values that relate to a non-existent hypothetical population; different countries have different risk relationships, so the estimates of effect described in the paper apply to no single country.
Further, the investigators do not report rates of other outcomes, especially total mortality, associated with drinking. Taking only one outcome condition, without at least noting that moderate drinking has almost uniformly been found to decrease all-cause mortality, limits the usefulness of their results for setting alcohol guidelines. Heavy alcohol use is recognized by all as a serious impediment to the health of individuals, and to society at large. However, recommending lowering the intake of alcohol for everyone does not take into consideration that light drinkers might be more likely than heavy drinkers to follow such advice and might decrease their consumption to zero. Because of moderate drinking’s almost uniformly demonstrated beneficial effect on cardiovascular disease, if many light drinkers stopped drinking completely it would be expected to increase, rather than decrease, total mortality rates and the health of the population.
In their conclusions, the authors only mention the results of their actual analyses briefly, then start the usual chant of the need for extensive public health measures (warning labels, public information campaigns, increasing alcohol taxes, etc.) which tend to emphasize the dangers of any alcohol consumption. However, the analyses in this paper do not provide the data that are needed if one is to set appropriate drinking guidelines, which should be age-specific, sex-specific, and population-specific. Such factors would include not only the alcohol-disease relation but the particular characteristics of the population, the objective of guidelines (to influence and change behavior among target populations), the behavior that is thought to be in need of change, the culture and mindset of the target population, and the kind of message that is likely to be effective. Forum members realize that there are tensions between advice intended only to reduce the prevalence of misuse versus advice that takes into account the well demonstrated beneficial health effects of light-to-moderate consumption. When setting guidelines, it is important to take into account the prevailing drinking culture of a country or population, because only in that way is it possible to produce a public health message that is likely to be respected and regarded.
Comments on this critique by the International Scientific Forum on Alcohol Research were provided by the following Forum members:
Erik Skovenborg, MD, specialized in family medicine, member of the Scandinavian Medical Alcohol Board, Aarhus, Denmark
Creina Stockley, PhD, MSc Clinical Pharmacology, MBA; Principal, Stockley Health and Regulatory Solutions; Adjunct Senior Lecturer, The University of Adelaide, Adelaide, Australia
Arne Svilaas, MD, PhD, general practice and lipidology, Oslo University Hospital, Oslo, Norway
Pierre-Louis Teissedre, PhD, Faculty of Oenology–ISVV, University Victor Segalen Bordeaux 2, Bordeaux, France
David Van Velden, MD, Dept. of Pathology, Stellenbosch University, Stellenbosch, South Africa
Andrew L. Waterhouse, PhD, Department of Viticulture and Enology, University of California, Davis, USA
Yuqing Zhang, MD, DSc, Clinical Epidemiology; Massachusetts General Hospital, Harvard Medical School, Boston, MA, USA.
Giovanni de Gaetano, MD, PhD, Department of Epidemiology and Prevention, IRCCS Istituto Neurologico Mediterraneo NEUROMED, Pozzilli, Italy
Luc Djoussé, MD, DSc, Dept. of Medicine, Division of Aging, Brigham & Women’s Hospital and Harvard Medical School, Boston, MA, USA
Harvey Finkel, MD, Hematology/Oncology, Retired (Formerly, Clinical Professor of Medicine, Boston University Medical Center, Boston, MA, USA)
Fulvio Mattivi, MSc, CAFE – Center Agriculture Food Environment, University of Trento, via E. Mach 1, San Michele all’Adige, Italy
R. Curtis Ellison, MD, Professor of Medicine, Section of Preventive Medicine & Epidemiology, Boston University School of Medicine, Boston, MA, USA