Critique 162: What is the association between alcohol consumption and cancer? 13 April 2015

This critique relates to the following three recent publications relating alcohol to cancer: 

1.  Klatsky AL, Li Y, Tran HN, Baer D, Udaltsova N, Armstrong MA, Friedman GD. Alcohol Intake, Beverage Choice, and Cancer: A Cohort Study in a Large Kaiser Permanente Population. Perm J 2015;19:March 1, 2015. http//dx.doi.org/10.7812/TPP/14-189

Authors’ Abstract

Context: Heavy intake of alcoholic beverages is associated with an increased risk of developing several types of cancers at specific body sites. However, evidence is conflicting regarding alcohol-associated cancers in other sites of the body as well as the role played by choice of wine, liquor, or beer.

Objective: To study incident cancer risk from 1978 to 1985 and through follow-up in 2012 relative to light-to-moderate and heavy drinking and to the choice of alcoholic beverage in a cohort of 124,193 persons.

Design: Cohort.

Main Outcome Measures: 1) Cox proportional hazards models controlled for 7 covariates to analyze alcohol-associated risk of any cancer and multiple specific types. 2) Similar analyses in strata of drinkers with or without a preponderant choice of wine, liquor, or beer and with or without inferred likelihood of underreporting.

Results: With lifelong abstainers as referent, heavy drinking (≥ 3 drinks per day) was associated with increased risk of 5 cancer types: upper airway/digestive tract, lung, female breast, colorectal, and melanoma, with light-to-moderate drinking related to all but lung cancer. No significantly increased risk was seen for 12 other cancer sites: stomach, pancreas, liver, brain, thyroid, kidney, bladder, prostate, ovary, uterine body, cervix, and hematologic system. For all cancers combined there was a progressive relationship with all levels of alcohol drinking. These associations were largely independent of smoking, but among light-to-moderate drinkers there was evidence of confounding by inferred underreporting. Beverage choice played no major independent role.

Conclusion: Heavy alcohol drinking is related to increased risk of some cancer types but not others. Because of probable confounding, the role of light-to-moderate drinking remains unclear.

2.  Bagnardi V, Rota M, Botteri E, et al. Alcohol consumption and site-specific cancer risk: a comprehensive dose–response meta-analysis. British Journal of Cancer 2015;112:580–593. doi: 10.1038/bjc.2014.579  

Authors’ Abstract

Background:  Alcohol is a risk factor for cancer of the oral cavity, pharynx, oesophagus, colorectum, liver, larynx and female breast, whereas its impact on other cancers remains controversial.

Methods:  We investigated the effect of alcohol on 23 cancer types through a meta-analytic approach. We used dose-response meta-regression models and investigated potential sources of heterogeneity.

Results:  A total of 572 studies, including 486 538 cancer cases, were identified. Relative risks (RRs) for heavy drinkers compared with nondrinkers and occasional drinkers were 5.13 for oral and pharyngeal cancer, 4.95 for oesophageal squamous cell carcinoma, 1.44 for colorectal, 2.65 for laryngeal and 1.61 for breast cancer; for those neoplasms there was a clear dose–risk relationship.  Heavy drinkers also had a significantly higher risk of cancer of the stomach (RR 1.21), liver (2.07), gallbladder (2.64), pancreas (1.19) and lung (1.15).  There was indication of a positive association between alcohol consumption and risk of melanoma and prostate cancer.  Alcohol consumption and risk of Hodgkin’s and Non-Hodgkin’s lymphomas were inversely associated.

Conclusions: Alcohol increases risk of cancer of oral cavity and pharynx, oesophagus, colorectum, liver, larynx and female breast.  There is accumulating evidence that alcohol drinking is associated with some other cancers such as pancreas and prostate cancer and melanoma.

3.  Wienecke A, Barnes B, Neuhauser H, Kraywinkel K. Incident cancers attributable to alcohol consumption in Germany, 2010.  Cancer Causes Control 2015;[Epub ahead of print].

Authors’ Abstract

Purpose:  Germany lacks an up-to-date assessment of the cancer burden attributable to alcohol. Therefore, cancer incidence attributable to this exposure was estimated for colorectal, liver, breast, and upper aerodigestive tract (UADT) cancer. Additionally, the impact of alcohol on UADT cancer was analyzed by smoking status, to account for synergistic interactions between these two risk factors.

Methods: Alcohol consumption and smoking prevalence from a nationwide survey in Germany 2008-2011 were combined with relative risks of incident cancer from meta-analyses to obtain population attributable risks (PARs), indicating the proportion of cancers that could be avoided by eliminating a risk factor. Each PAR was multiplied with the respective cancer incidence for 2010 to calculate the absolute number of attributable cases.

Results: In Germany, for the year 2010, approximately 13,000 incident cancer cases could be attributed to alcohol consumption (3 % of total cases). PAR was highest for esophageal cancer (men: 47.6 % and women: 35.8 %) and lowest for colorectal cancer in men (9.7 %) and breast cancer in women (6.6 %). Among women, moderate consumption levels account for the greatest PAR overall, whereas heavy drinking contributes considerably to overall PAR among men. Additionally, moderate-to-heavy drinking among smokers substantially contributes to the overall PAR of UADT cancers compared to drinking among non-smokers.

Conclusion: In Germany, a substantial proportion of cases of common cancers can be attributed to alcohol consumption, even when consumed at moderate levels. Alcohol consumption with concurrent tobacco smoking is especially important for cancers of the UADT. These findings strengthen the rationale for prevention measures that address exposure at all levels.

Forum Comments

There is no question that heavy alcohol consumption is associated with most upper aero-digestive cancers, including those of the mouth, pharynx, esophagus, etc.  In fact, as these are so often seen in heavy drinkers, they are often referred to as “alcohol-related cancers.”  For these cancers, coexisting heavy cigarette smoking has been shown to markedly increase the risk of cancer.

One feature of the “alcohol-related cancers” is that alcohol comes in direct contact with these tissues.  Other cancers appear to be related to alcohol consumption more indirectly, perhaps related to blood alcohol levels.  The level of association with alcohol for these cancers is usually less than for the alcohol-related cancers, but an association is often discovered through observational epidemiologic studies or animal experiments.

Comments on paper by Klatsky et al:  The paper by Klatsky et al, based on the very large Kaiser-Permanente Study, had 17.8 years of follow up among white, black, Asian, and Latino subjects.  The investigators used persistent abstainers as the referent group, with categories of an average of < 1 drink/day as light drinkers, 1 to 2 drinks per day as moderate drinkers, and 3 or more drinks/day as heavy drinkers.  The investigators had data allowing them to adjust for age, race, education, BMI, marital status, and smoking, and had the ability to estimate which of the moderate drinkers were under-reporting their intake.  No significant differences were noted between abstainers and drinkers for liver cancer, although there was a tendency toward increased risk for moderate and heavy drinkers.  The strongest alcohol-cancer associations were found for melanoma and breast cancer in women, with small but significant increases in risk of the latter even among light drinkers.  No significant differences according to type of alcoholic beverage were found.   This finding could not be understood by Lanzmann-Petithory, who noted that many studies show a decrease in cancer risk to be associated with increased fruit intake, and “Wine is the combination of alcohol and polyphenol enriched fruit extract.  We could guess that at moderate intake (still to be specified), fruit effect wins over alcohol effect.”

The study found that under-reporting of alcohol consumption strongly affected the risk of cancer. Among subjects reporting 1 to 2 drinks/day, those whose data within the Kaiser-Permanente system suggested that they were under-reporting their alcohol intake (findings indicating greater intake on other occasions, evidence of alcohol misuse, alcohol-related liver disease, etc.), the HR versus abstainers for any cancer was 1.4 (95% CI 1.3-1.7).  In contrast, for those reporting the same amount but having no evidence within their medical records suggesting greater intake or misuse (thus, subjects deemed unlikely to be under-reporters), the HR for any cancer was 1.1 (95% CI 0.9 – 1.2).  The study concludes that consuming an average of 3 or more drinks/day increases the risk of cancers of the upper airway/digestive tract, lung, female breast, colorectal, and melanoma; this study did not find an increase for cancers of the stomach, pancreas, liver, or prostate.

Overall, the authors state: “At present, a possible increased cancer risk at moderate intake should enter into individual estimation of the overall risk-benefit equation for alcohol drinking, especially for young persons.  For most persons older than age 50 years, the overall benefits of lighter drinking, especially the reduced risk of atherothrombotic disease, outweigh possible cancer risk.”

Comments on paper by Bagnardi et al:  The paper by Bagnardi et al was a meta-analysis based primarily on cancer incidence from a total of 572 studies, of which 163 were cohort studies and 409 were case-control studies (in general, the latter tend to have higher estimates of alcohol effect than prospective cohort studies, such as that of Klatsky et al).  As with any meta-analysis, the authors used confounder-adjusted estimates when provided by the authors of individual studies, but were otherwise forced to use “unadjusted RRs from the raw data presented in the paper.”

The investigators of this meta-analysis defined ≤ 12.5 g/day of alcohol (about one typical drink) as light, > 12.5 to ≤ 50 g/day (1 to 4 or 5 typical drinks) as moderate, and > 50 g/day as heavy.  They report increased risks for cancers of the upper airway/digestive tract, lung, female breast, colorectal, and melanoma for heavy drinkers.  In addition, significant increases were reported for heavy drinking for cancers of the stomach, liver, gallbladder, pancreas, and lung, although the risk ratios for most of these were lower.

Reviewer Ellison noted that the results among subjects assigned to the “light” drinking category (<12.5 g/day) in this study can be related to those in the Klatsky lightest drinking category, and the results were similar.  “However, the inclusion of drinkers averaging up to 50 grams of alcohol per day (up to 4 or 5 typical drinks) in the “moderate” category makes it impossible to compare the results of this meta-analysis with those of Klatsky et al and much other previous research.  Current US Dietary Guidelines (2010) define moderate as no more than 1 drink/day for women and no more than 2 drinks/day for men.”

Reviewer de Gaetano was also worried about the wide range of alcohol intake defined as “moderate.”  He stated: “In our meta-analysis (Di Castelnuovo et al) we saw already at 40 g daily that the significant beneficial effect of alcohol against total mortality had completely disappeared . . . and between 40 and 50 g/day there was a clear increase in mortality, in respect to abstainers.  Thus, in Bagnardi’s meta-analysis, moderate was not moderate indeed!”

As for potential problems associated with combining data from cohort and case-control studies, the authors remark that “heterogeneity across studies was high for some types of cancer.”  They state that therefore, “some of the estimates should be interpreted with caution.”  Further, they state that because of lack of data, they could not judge the effects of different drinking patterns, different types of beverage, or potential under-reporting of alcohol intake among subjects.

Comments on paper by Wienecke et al:  The paper by Wienecke et al based an estimate of alcohol consumption from a nation-wide survey in Germany in 2008-2011.  Alcohol data from a food-frequency questionnaire in the survey recorded consumption during the 4-week period before the survey.  Results were stratified into categories of “moderate” (< 3 drinks/day) and “heavy” (at least 3 drinks/day), with 3 drinks considered to be approximately 30 g of alcohol.  Cancer incidence in Germany was obtained from estimates by the German Center for Cancer Registry Data at the Robert Koch Institute.

The authors used relative risks for cancer derived from previous meta-analyses to test for association between alcohol and cancer.  The meta-analyses used to estimate risk ratios were published between 1999 and 2014 by Hashibe et al, Castellsagué et al, Fedirko et al, Turati et al, and Ridolfo et al.  Only two of the meta-analyses also considered the effects of tobacco use.  Of the 149 separate studies included in the present analysis, 104 are based on case-control studies and 45 on cohort studies.  Their focus was on the population-attributable cancer burden in Germany from alcohol compared with zero alcohol intake.  They evaluated the effects of “moderate” intake, defined as < 3 drinks/day, and “heavy” drinking (at least 3 drinks/day).   They found that there was an apparent increase in the population attributable risk of cancer for men for upper aero-digestive tract (UADT) and of the colon-rectum, but decreases from alcohol for liver cancer.  For women, there was a decrease in colon-rectal cancer and liver cancer from alcohol, but increases for breast and UADT cancers.

The authors comment on the importance of considering coexisting tobacco use when judging the effects of alcohol on UADT cancers.  They conclude that “Our analysis indicates that approximately 3% of all incident cancer cases among German adults ages ≥ 35 years could be attributable to alcohol consumption in 2010.”  The highest proportion of alcohol-attributable cases was estimated for esophageal cancer, but the absolute numbers of attributable cases were for breast cancer and colorectal cancer, as these are much more common diseases in the population.  In their discussion, the authors point out: “The high burden of alcohol-associated diseases other than cancer should be considered as well, since cancer prevention is not the only benefit to be expected when eliminating or reducing alcohol consumption.”  They they go on to mention fetal alcohol syndrome and liver cirrhosis but, interestingly, do not mention the demonstrated much more important effects of moderate drinking on reducing cardiovascular disease.

Determining thresholds for effects of alcohol on risk of cancer; the net health effects of alcohol consumption:  All three studies describe increased risk for many cancers among heavy drinkers, especially UADT cancers.  However, light-to-moderate alcohol intake is by far the most common pattern of drinking in the population, and has been related in most studies to a lower risk of cardiovascular diseases and other diseases of ageing and a decrease in the risk of total mortality.  Hence, Forum members think that the most important questions we need to answer are the net effects on health for various drinking levels.  We do not need further data to know that heavy drinking leads to certain cancers, and many other adverse health and societal problems.  However, given that the majority of people in industrialized nations consume alcohol, what we need are reliable threshold levels of drinking that increase the risk of cancer and other diseases; even with the three large studies reviewed here, there are still limitations in our ability to determine this for some cancers.  Reviewer Ellison considered it especially important to evaluate and comment upon “the net effects of alcohol consumption at different levels of consumption.  Only when these are considered can reasonable drinking guidelines be presented for individual patients and the public.” 

The study by Klatsky et al includes confounder-adjusted data from a well-monitored cohort, and suggests that even consumption of up to 1 drink/day may be associated with small but significant increases in the risk of some cancers (UADT, colo-rectal, breast, and especially melanoma).  Under-reporting of alcohol intake is said to affect the risk, but separate results for subjects not considered to be under-reporting their intake are not given for specific cancers in this paper.  However, it is stated that for subjects reporting an average of 1 to 2 drinks/day, the HR for any cancer decreases from 1.4 (CI 1.3 – 1.7) for likely under-reporters to 1.1 (CI 0.9 – 1.2) for those deemed unlikely to be under-reporting their intake.

Individual study results versus results from meta-analysis:  There are strengths and weaknesses in reaching conclusions from meta-analyses.  The number of cases is generally much larger than the numbers from individual studies so that risk ratios of lesser magnitude can often be found to be statistically significant; this allows cancers with more modest increases from alcohol to be detected.  However, the categories in the multiple studies included in a meta-analysis often differ, so compromises must be made in the analysis.  The degree of control of potential confounding factors also varies among studies included in a meta-analysis.

Large, well-done prospective studies, especially those with repeated assessments of alcohol intake, offer more precise estimates of alcohol intake, and can account for changes in alcohol intake over time.  However, the number of cases may be inadequate to judge more subtle increases in risk.  For studies (such as the Kaiser-Permanente Study) with considerable health data over many decades, better adjustment for potential confounders is possible.  Also, as has been demonstrated from the Kaiser-Permanente Study, determining which subjects may be under-reporting their typical alcohol consumption can be estimated, which can be important when seeking to judge the effects of light-to-moderate drinking, especially when seeking thresholds for adverse effects of alcohol.

Reviewer Finkel added: “There are differences between cohort papers and meta-analyses.  These differences, while usually small and perhaps the expected biological variation, may be expressions of meaningful information.  Shouldn’t we treasure and analyze them?  Otherwise, we risk seeing and talking about the same issues over and over again: statistical and epidemiological nuances, confounding factors, choice of beverage, accuracy of self-reporting, adequacy of folate intake, nutrition, endocrine balance, etc.  I suppose that’s our job, but sometimes it feels repetitive.”  Finkel also stated: “I strongly endorse other members’ emphasis on the net effects of alcohol consumption, too often neglected by investigators with tunnel vision.”

Forum member Thelle had some interesting observations:  “The issues of interest among all of these papers are the etiological role of alcohol in carcinogenesis (about which these observational studies only can provide associations for further research) and the public health message regarding thresholds.  The paper by Klatsky et al has an interesting remark concerning the public health issue and the potential net benefit for those aged 50 and over.  This is a conclusion that may be reached by a large longitudinal study based upon one single population, but less so by meta-analysis.”  Thelle added: “Unfortunately, the meta-analysis assessed both case-control and longitudinal studies.  Further, I am slightly surprised that little has been said about interaction of smoking and alcohol in these papers.”

General comments on the topic of alcohol and cancer:  Forum member Thelle provided an overview of the topic.  “First a few general comments on cancer epidemiology.  Cancer of the stomach has been declining the last fifty years, probably because of improved hygiene and fewer infected by helicobacter pylori plus a reduced intake of salt.  Cancer of the colon has increased during the same period of time. Breast cancer is difficult to assess due to screening procedures likely to affect incidence figures (and those women who are first screened — the better off– also consume more alcohol). The strongest association between alcohol and cancer is seen for esophagus cancer and upper airway cancers.  Liver cancer, or rather hepatocellular carcinoma, is also among the more strongly associated cancers in some studies.  Both cancers can be considered rare in the Nordic countries, and it is interesting to note that the incidence of these cancers has not changed much during a period when alcohol consumption has increased considerably.

“But even for these cancer types the risk ratios are relatively low, and the possibility of residual confounding must be taken seriously.  I think we must consider some general problems when relating alcohol to risk of cancer:

1.  The heterogeneity of cancer. It is not one particular disease; the causes are different and they should not be lumped together.

2.  Exposure time.  The decline in risk estimates after 20 years; what does this mean? Are other factors taking over?

3.  Are there cohort effects in meta- analyses which will affect the average effects?

4.  Does the effect of alcohol at meals differ from alcohol independent of food intake?

5.  The relative risk estimates are rather low, but still sufficient to say that they satisfy the Hill criteria for strong associations when discussing etiology (or is Hill old-fashioned today with causal algebra?).

6.  Acetaldehyde is a possible culprit, but the half-time is short. Is this a plausible mechanism?

“And I have some particular issues with regard to the three papers under review:

1.  Klatsky et al. Why are the ex-drinkers having increased risk for UADT and liver cancer, but less so for melanoma and colon-rectal cancer?

2.  The Klatsky paper showed higher risk for smokers who consumed alcohol. An interaction effect is suggested. but not commented upon.

3.  The association with melanoma; to what extent is UV radiation a confounder here? Sunny holidays and drinking?

4.  Wienecke et al used risk estimates from meta-analyses and applied them to German survey data to calculate how many cases could be attributed to alcohol. They end up with the same fraction of total cancer attributed to alcohol, about 3%, but I am not sure whether that is a circular argument.

“My conclusion is that the area is in shades of grey.  A high alcohol consumption is likely to contribute to some cancers, but at lower levels, the case remains unsolved.”

Reviewer Ursini commented: “I fully agree with the comments by Thelle.  As a basic scientist, I feel uncomfortable with the present expansion of computational evaluation of a risk, which for non-specialists becomes a ‘cause.’   Are the Hill criteria completely abandoned, and what about the biological rationale that must be in agreement with the epidemiological results ?  We teach to students that basic science drives and inspires clinical studies and epidemiology, and that meta-analyses can only provide the final confirmatory evidence.  Apparently, the need for getting results and publishing quickly is not dimmed by the indispensable requirement of scientific thoroughness.  Finally, I must also admit my surprise in reading the name of the journal where the paper by Wienecke has been published: Cancer Causes Control.  Isn’t this promising too much at the present level of knowledge?  Also considering that a recent paper in Science identifies the real cause of cancer is ‘bad luck.’”

Forum member Finkel stated: “Based on my years practicing and teaching, I still view some upper aerodigestive (especially in concert with tobacco) and hepatocellular cancers as clearly tied to alcohol, to state it generally, with, no doubt, nuances to be unraveled.  I remain confused about breast cancer and, perhaps, colorectal cancer in this regard.  I would return a guilty verdict versus heavy drinking, but am uneasy when it takes intricate statistics to implicate light or moderate drinking as either promoters of or protectors from various cancers.  The definitions of light-moderate-heavy sometimes vary awkwardly.  I also feel the need for clarification of beverage types and drinking pattern — and so many confounders.”

Reviewer Mattivi commented: “There is a clear consensus that a high alcohol consumption is likely to contribute to several cancers, and these three papers further support the IARC classification of ethanol in alcoholic beverages as carcinogenic to humans.  On the other hand, there is a strong need to improve our understanding of the biological rationale and to further clarify the dose-effect, especially at a lower level of consumption, and also to distinguish among different alcoholic beverages.”

Dr. Samir Zakhari, former director of the Division of Metabolism and Health Effects at the National Institute on Alcohol Abuse and Alcoholism (NIAAA), has noted when evaluating the relation of alcohol to breast cancer, “Numerous risk factors are involved in breast carcinogenesis; some are genetic and beyond the control of a woman; others are influenced by lifestyle factors  The current state-of-knowledge about alcohol and breast cancer association is ambiguous and confusing to both a woman and her physician” (Zakhair & Hoek).

Forum member Skovenborg provided his impressions of the overall data relating alcohol to cancer, reaching the following conclusions: “The association between heavy drinking and certain cancer types is strong and convincing and a trustworthy biological pathway has been described for most of these cancer types.

“However, there is a deplorable lack of a uniform definition of light, moderate, and heavy drinking.  A recent example, that has had a large coverage by the media, is from a World Cancer Research Fund (WCRF) report regarding the associations between food, nutrition and physical activity and the risk of liver cancer, which stated that ‘Three drinks a day can cause liver cancer’ (wcrf.org).  A closer look at their published data indicate that the slope of the non-linear dose-response figure for ethanol intake and liver cancer is fairly flat below 45 grams of alcohol a day.  According to the WRCF, that is up to ‘three drinks a day’ – however 45 grams of alcohol amounts to 5.25 British drinks, 4.5 French drinks and 3.75 Danish drinks, which is heavy drinking by almost any definition and not moderate drinking as the WCRF headline implies.

“Further, a serious flaw in most reports on alcohol and cancer is the lack of information on confounding, drinking patterns, beverage choice, as well as the effects of underreporting of alcohol intake.  Also, a damaging flaw in the present reports is the omission of a discussion of total mortality vs cancer mortality.  Such omissions support sometimes misguided information being widely circulated to the public.”

References from Forum review

Di Castelnuovo A, Costanzo S, Bagnardi V, Donati MB, Iacoviello L, de Gaetano G.  Alcohol dosing and total mortality in men and women: an updated meta-analysis of 34 prospective studies.  Arch Intern Med. 2006;166:2437-2445.

Hashibe M, Brennan P, Chuang S et al.  Interaction between tobacco and alcohol use and the risk of head and neck cancer: pooled analysis in the international head and neck cancer epidemiology consortium. Cancer Epidemiol Biomark Prev 2009;18:541–550.

Castellsague´ X, Munoz N, De Stefani E, et al.  Independent and joint effects of tobacco smoking and alcohol drinking on the risk of esophageal cancer in men and women. Int J Cancer 1999;82:657–664.

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;22:1958–1972.

Turati F, Galeone C, Rota M, et alcohol and liver cancer: a systematic review and meta-analysis of prospective studies. Ann Oncol 2014;25:1526–1535.

Ridolfo B, Stevenson C.  The quantification of drug-caused mortality and morbidity in Australia, 1998. In: Australian Institute of Health and Welfare (ed); Drug statistics series. Australian Institute of Health and Welfare, 2001. Canberra.

Dietary Guidelines for Americans 2010, U.S. Department of Agriculture, U.S. Department of Health and Human Services, www.dietaryguidelines.gov.

WCRF report.  Available at wcrf.org/sites/default/files/Liver-Cancer-2015-Report.pdf.

Zakahari S, Hoek JB.  Alcohol and Breast Cancer: Reconciling Epidemiological and Molecular Data.  In, V. Vasiliou et al. (eds.), Biological Basis of Alcohol-Induced Cancer, Advances in Experimental Medicine and Biology 815. Springer International Publishing, Switzerland; DOI 10.1007/978-3-319-09614-8_2

Forum Summary

Three major papers on the association of alcohol consumption and cancer have recently been published.  Forum members considered that all were well done, and presented valuable new information on the topic.  While the key findings in each study are similar, each brings specific information on how alcohol relates to the risk of developing cancer.  In this critique, we present the authors’ abstracts of each paper, our comments on the paper, and then an overall discussion of the topic.

Problems can occur when combining case-control studies with prospective cohort studies, as was done in one of the papers reviewed; risk estimates are usually higher in the former type of study and control for potential confounding may be less complete.  Further, it is especially important to consider the interaction between alcohol consumption and smoking for upper aero-digestive tract cancers, as was not always done.  When able to be estimated, under-reporting of alcohol intake among those claiming to be moderate drinkers can help define the relation between alcohol and disease.  And none of the studies covered in this critique emphasized the net effects of alcohol consumption: because of protective effects against cardiovascular disease, moderate drinking is almost always associated with a lower risk of total mortality than that which occurs among abstainers.

In summary, the Forum considers that the three papers reviewed provide important data on one of the few lifestyle factors that have been shown to relate to the risk of cancer.  And cumulative research data clearly show that heavy alcohol intake increases the risk of upper aero-digestive tract cancers and some other cancers.  Further, in these and many previous reports, even light alcohol consumption was associated with an increase in risk of breast cancer in women.  Unfortunately, for most cancers, the threshold level of drinking associated with an increase in risk is not clearly defined.

Forum members agree with the conclusions presented by Klatsky et al in his paper for providing advice regarding alcohol consumption: “At present, a possible increased cancer risk at moderate intake should enter into individual estimation of the overall risk-benefit equation for alcohol drinking, especially for young persons.  For most persons older than age 50 years, the overall benefits of lighter drinking, especially the reduced risk of atherothrombotic disease, outweigh possible cancer risk.”

                                     *                *                      *

Comments on this critique by the International Scientific Forum on Alcohol Research were provided by the following members:

Giovanni de Gaetano, MD, PhD, Department of Epidemiology and Prevention, IRCCS Istituto Neurologico Mediterraneo NEUROMED, Pozzilli, Italy

R. Curtis Ellison, MD. Section of Preventive Medicine & Epidemiology, Department of Medicine, Boston University School of Medicine, Boston, MA, USA

Harvey Finkel, MD, Hematology/Oncology, Boston University Medical Center, Boston, MA, USA

Ulrich Keil, MD, PhD, Professor Emeritus, Institute of Epidemiology & Social Medicine, University of Muenster, Germany

Dominique Lanzmann-Petithory,MD, PhD, Nutrition/Cardiology, Praticien Hospitalier Hôpital Emile Roux, Paris, France

Fulvio Mattivi, PhD, Head of the Department of Food Quality and Nutrition, Research and Innovation Centre, Fondazione Edmund Mach, in San Michele all’Adige, Italy

Erik Skovenborg, MD, specialized in family medicine, member of the Scandinavian Medical Alcohol Board, Aarhus, Denmark

Creina Stockley, PhD, MSc Clinical Pharmacology, MBA; Health and Regulatory Information Manager, Australian Wine Research Institute, Glen Osmond, South Australia, 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

Dag S. Thelle, MD, PhD, Senior Professor of Cardiovascular Epidemiology and Prevention, University of Gothenburg, Sweden; Senior Professor of Quantitative Medicine at the University of Oslo, Norway

Fulvio Ursini, MD, Dept. of Biological Chemistry, Universityof Padova, Padova, Italy

Andrew L. Waterhouse, PhD, Department of Viticulture and Enology, University of California, Davis, USA

Note on potential conflict of interest:  An author of one of these papers, Arthur Klatsky, is a member of the International Scientific Forum on Alcohol Research.   Dr. Klatsky had no input into the preparation of this critique.