Critique 242: The relation of light alcohol consumption to the risk of common cancers – 1 October 2020
Caprio GG, Picascia D, Dallio M, Vitiello PP, Giunta EF, De Falco V, et al. Light Alcohol Drinking and the Risk of Cancer Development: A Controversial Relationship. Reviews on Recent Clinical Trials 2020;15:164-177
Background: In accordance with the scientific literature heavy alcohol consumption (>50g per day) represents a risk factor for several diseases development, including cancer. However, the oncogenic role of light alcohol drinking (<12.5g per day) is still unknown.
Objective: To assess the scientific knowledge about light alcohol consumption and the risk of malignancy onset.
Methods: To collect the scientific evidences regarding this topic the keywords “light alcohol drinking”, “light alcohol consumption” and “cancer”, were used. Papers published during the last 15 years were analyzed, in order to select the most recent evidence. Meta-analyses with well-defined levels of alcohol intake were included in the present review. Other studies that focused on biochemical, molecular and genetic aspects, as well as duplicate articles, were excluded.
Results: Twenty-nine large, meta-analyses were included in this review. Light alcohol drinking was not associated with an increased risk of cancer occurrence, with the exception of breast and prostate cancer and melanoma. Furthermore, a possible protective role of light alcohol consumption on the development of bladder, kidney and ovarian cancer and Non Hodgkin Lymphoma was observed.
Conclusion: Light alcohol drinking was not associated with the development of several malignancies, except for a light increase of melanoma, breast cancer in women and prostate cancer in men.
Forum member Finkel was not very impressed by this study, citing a number of weaknesses: “It seems vague on the details of method. It lacks data that might be of value: numbers and sex of subjects, types of beverage, drinking patterns. By now, shouldn’t we be entitled to this potentially vital data in our literature?” He added: “I appreciate that with light drinking, one shouldn’t expect large differences in RRs, but these are small enough to be suspect. This paper seems to me a start, but it hasn’t the muscle to convince. On the other hand, it is reassuring to those who wish to continue to enrich their meals and lives by including a dram or two. There is nothing in the paper that would give me panic, or even pause, and it should encourage impartial investigators to continue and to refine their efforts.”
Reviewer Ellison agreed that this paper had many problems. “While the title states that it is evaluating ‘light alcohol drinking,’ and their pre-analysis search of the literature focused on that, much of their paper deals with the effects of greater amounts of alcohol. It seems that many more meta-analyses may have qualified for inclusion if they searched on all levels of alcohol consumption. And there are numerous inconsistencies in the paper: e.g., for oesophageal cancer they state that there was no significant correlation between light alcohol and cancer but then report a RR of 0.86, CI of 0.75-0.99), indicating an estimated 14 % reduced risk; for pancreatic cancer, they state ‘no significant increase in cancer risk, in case of less than 3 drinks per day consumption’ then list a RR of 0.92, (CI 0.86-0.97), actually a small but significant decrease.”
Ellison continued: “The presentation of details of the studies included in Table 1 are useful, but it is unfortunate that the type of study being reported and the number of subjects are not given. And there are some errors in the table, such as my colleagues and I being quoted for results on oral cavity and pharynx cancer whereas the results they report are presumably from someone else (perhaps reference 42; our reference, number 31, was for alcohol and breast cancer). Under Oral and Pharyngeal Cancer, the authors quote a meta-analysis by Bagnardi et al as showing increased risk of cancer with a RR of 1.13, then in the next sentence state ‘However, for light alcohol consumption, no increase in cancer onset was found (RR 0.86).’ And under Renal Cell Carcinoma, the authors state that 12 g/day of alcohol was correlated with a 5% reduction in the risk, then give a RR of 0.67, CI 0.62-0.73, compatible with a much greater reduction.
“While brevity in scientific papers is always welcome, the very brief sections on each type of cancer presented in the Results section are sometimes based only on one or two reports, markedly inadequate to develop conclusions that reflect our current knowledge on these topics. This is especially so since much of what we know about alcohol and cancer has been based on results from very large, well-evaluated cohorts followed for many decades; such studies contain detailed data on subjects and their behaviors that cannot generally be included when one has access only to combined data, as in meta-analyses.”
Forum member Parente added: “Meta-analyses are often limited by apples-to-oranges comparisons or false equivalences. These limitations are compounded by the meta-meta-analysis approach taken here. The findings are likely to be lauded by lumpers and slammed by splitters; it appears that most Forum members are splitters, with their views driven by the many studies that have shown different effects from different beverages, for different drinking patterns, and in different cultures.” Reviewer Pajak wrote: “We must state that the authors attempted a very ambitious goal to summarize the existing evidence. They could not go beyond the existing evidence and were exposed to the limitations in the interpretation of the results of the cited studies.”
Reviewer Estruch stated: “In relation to the most important of covariates that really matter for the relation between alcohol consumption and incidence of cancer, we should also note the dietary patterns of subjects. I agree with others that the drinking pattern (mainly type of alcoholic beverage and regular intake or in binges) and the consumption with or without food are very important for the final result of this relationship (protection or induction of cancer). However, the dietary pattern may be even more important. If a woman follows a Western dietary pattern, incidence of breast cancer may increase by 14 %, but if she follows a prudent dietary pattern, incidence of breast cancer decreases by 18% (Xian et al). Therefore, the final result of alcohol (wine) consumption on incidence of breast cancer may be positive (increase) or negative (reduction) depending of the diet followed. The key may be to drink moderately wine or other polyphenols-rich alcoholic beverages under the protective umbrella of a healthy dietary pattern such as a Mediterranean diet.”
Forum member Skovenborg stated: “I agree with the reserved views on the paper expressed by other reviewers. I also miss (1) information on drinking pattern; (2) comments on the probable underreporting of alcohol intake; and (3) comments on the possible confounding factors that are always suspected in observational studies with very small increases in risk. The authors note that ‘the mechanisms underlying alcohol-related carcinogenesis are not completely understood.’ I would rather say that plausible biological mechanisms are absent in the case of light drinking, especially in relation to melanoma and prostate cancer.”
Forum members agreed that it is not possible to use only the data presented in this paper for determining appropriate advice to the public regarding light drinking and the risk of various types of cancer. Forum member Waterhouse added: “This error-ridden paper does not deserve much attention. Those who cite it without calling out the errors will reveal an uncritical approach to science.”
References from Forum critique:
Xiao Y, Xia J, Li L, et al. Associations between dietary patterns and the risk of breast cancer: a systematic review and meta-analysis of observational studies. Breast Cancer Res 2019;21:16.
In an attempt to assess the relation between light alcohol consumption (<12.5g per day) and the risk of a number of types of cancer, the authors carried out a review of meta-analyses published within the past 15 years. Using as keywords “light alcohol drinking,” “light alcohol consumption” and “cancer”, they found 29 meta-analyses on the subject. They report their results as indicating that “Light alcohol drinking was not associated with an increased risk of cancer occurrence, with the exception of breast and prostate cancer and melanoma. Furthermore, a possible protective role of light alcohol consumption on the development of bladder, kidney and ovarian cancer and Non Hodgkin Lymphoma was observed.”
Forum reviewers found a number of weaknesses in the paper. While their title focuses on “light alcohol drinking,” most of their reports of results and their discussion deal with harmful effects of heavier drinking; however, by not including papers on ‘heavy drinking’ in their original search, they may have left out many relevant papers. Further, the very brief reviews of each common cancer that they comment upon are sometimes based on only one or two meta-analyses, leaving out a huge amount of key data, especially that from single, well-done, long-term cohort studies with important information on the relation of alcohol to cancer; potential confounding by many factors could not be taken into consideration.
In the analyses in the present paper, the reported average level of alcohol consumed is used, but there is insufficient evaluation of possibly more important indices of alcohol exposure: the pattern of drinking (e.g., type of beverage consumed, in binges or on a regular basis, consumption with or without food, etc.). Further, there is probably inadequate control for important confounders or consideration of the effects of under-reporting, which may have been seen in single studies but not apparent in meta-analyses. In addition, a number of errors are present within the report, such as conflicting statements that may misrepresent the numeric data given. These problems make these results of limited use to professionals in health policy who are developing guidelines for alcohol use in the general public.
Comments on this critique by the International Scientific Forum on Alcohol Research were provided by the following members:
Harvey Finkel, MD, Hematology/Oncology, Retired (Formerly, Clinical Professor of Medicine, Boston University Medical Center, Boston, MA, USA)
Erik Skovenborg, MD, specialized in family medicine, member of the Scandinavian Medical Alcohol Board, Aarhus, Denmark
David van Velden, MD, Dept. of Pathology, Stellenbosch University, Stellenbosch, South Africa
Giovanni de Gaetano, MD, PhD, Department of Epidemiology and Prevention, IRCCS Istituto Neurologico Mediterraneo NEUROMED, Pozzilli, Italy
R. Curtis Ellison, MD, Professor of Medicine, Section of Preventive Medicine & Epidemiology, Boston University School of Medicine, Boston, MA, USA
Ramon Estruch, MD, PhD, Hospital Clinic, IDIBAPS, Associate Professor of Medicine, University of Barcelona, Spain
Andrew L. Waterhouse, PhD, Department of Viticulture and Enology, University of California, Davis, USA
Professor Andrzej Pająk, Epidemiology and Population Studies, Jagiellonian University Medical College, Kraków, Poland
Creina Stockley, PhD, MSc Clinical Pharmacology, MBA; Adjunct Senior Lecturer at the University of Adelaide, Australia
Matilda Parente, MD, consultant in molecular pathology/genetics and emerging technologies, San Diego, CA, USA
Arne Svilaas, MD, PhD, general practice and lipidology, Oslo University Hospital, Oslo, Norway