Critique 184: A major new meta-analysis on alcohol consumption and the risk of pancreatic cancer — 4 April 2016

Wang Y-T, Gou Y-W, Jin WW, Xiao M, Fang H-Y.  Association between alcohol intake and the risk of pancreatic cancer: a dose–response meta-analysis of cohort studies.  BMC Cancer 2016;16:212.  DOI 10.1186/s12885-016-2241-1

Authors’ Abstract

Background: Studies examining the association between alcohol intake and the risk of pancreatic cancer have given inconsistent results. The purpose of this study was to summarize and examine the evidence regarding the association between alcohol intake and pancreatic cancer risk based on results from prospective cohort studies.

Methods: We searched electronic databases consisting of PubMed, Ovid, Embase, and the Cochrane Library identifying studies published up to Aug 2015.  Only prospective studies that reported effect estimates with 95 % confidence intervals (CIs) for the risk of pancreatic cancer, examining different alcohol intake categories compared with a low alcohol intake category were included.  Results of individual studies were pooled using a random-effects model.

Results: We included 19 prospective studies (21 cohorts) reporting data from 4,211,129 individuals. Low-to-moderate alcohol intake had little or no effect on the risk of pancreatic cancer.  High alcohol intake was associated with an increased risk of pancreatic cancer (risk ratio [RR], 1.15; 95 % CI: 1.06–1.25). Pooled analysis also showed that high liquor intake was associated with an increased risk of pancreatic cancer (RR, 1.43; 95 % CI: 1.17–1.74).  Subgroup analyses suggested that high alcohol intake was associated with an increased risk of pancreatic cancer in North America, when the duration of follow-up was greater than 10 years, in studies scored as high quality, and in studies with adjustments for smoking status, body mass index, diabetes mellitus, and energy intake.

Conclusions: Low-to-moderate alcohol intake was not significantly associated with the risk of pancreatic cancer, whereas high alcohol intake was associated with an increased risk of pancreatic cancer.  Furthermore, liquor intake in particular was associated with an increased risk of pancreatic cancer.

Forum Comments

While the majority of prospective cohort studies have not suggested that light-to-moderate intake of alcohol increases the risk of pancreatic cancer, some groups continue to insist that any amount of alcohol increases the risk of many types of cancer.  And, indeed, in some meta-analyses heavy alcohol intake has been shown to increase the risk of pancreatic cancer, along with other risk factors such as smoking, diabetes, and obesity.

The present meta-analysis provides an up-to-date appraisal of the association by reporting data from more than four million subjects in prospective cohort studies, among whom 11,846 incident cases of pancreatic cancer were diagnosed.  The authors used the lowest intake group (non-drinkers or occasional drinkers) as the referent group, and defined “light” consumption as up to 12 grams/day (essentially one typical drink); 12-24 g/day as “moderate”; and ≥24 g/day as “heavy” drinking.

Forum members considered this to be an excellent paper, using appropriate methods for a meta-analysis on alcohol and pancreatic cancer; it was limited to data from prospective cohort studies.  The key results of the study were that (1) neither light drinkers (RR = 0.97) nor moderate drinkers (RR = 0.98) showed an increase in risk of pancreatic cancer when compared with the referent group; (2) overall, subjects classified as heavy drinkers had a slight increase in risk (RR= 1.15, 95% CI 1.06 – 1.25); (3) in beverage-specific analyses, there was no significant increase in risk even for heavy drinkers of beer (RR = 1.08, CI 0.90 – 1.30) or wine (RR = 1.09, CI 0.79 -1.49), but there was a significant increase for heavy drinkers of liquor (RR = 1.43, CI 1.17 – 1.74).

A large group of other sub-analyses carried out by the authors (of groups defined by gender and duration of follow up, as well as those adjusted for smoking, BMI, diabetes, energy intake, and study quality) tended to strongly support the overall findings.  The conclusions of the authors accurately reflect their results, indicating that high alcohol intake, especially liquor intake, might play a role in the risk of pancreatic cancer, but no increase is associated with light or moderate intake of alcohol.

Specific comments on paper by Forum members:  Forum members agreed that this was a very-large and well-done analysis with a number of strengths: it was based only on prospective cohort studies (most classified as quite “high quality” studies), and all of the studies adjusted at least for smoking and most for BMI (smoking and obesity are two of the known risk factors for pancreatic cancer).

Reviewer Ellison mentioned a number of weaknesses of the study, most of which had been acknowledged by the authors.  These include the fact that the meta-analysis was based on pooled data from the studies (as data on individual subjects were not available), there was a mixture of lifetime abstainers and ex-drinkers included in the referent group, and the same cut-points for category of alcohol intake were used for both men and women.  He further noted that the amount of alcohol consumed by subjects in the “heavy” drinking category was not given, so it is not possible to judge how many of such subjects in this group were regular drinkers (perhaps with only slight increases over “moderate” limits) and how many may have been heavy binge drinkers or had a diagnosed alcohol use disorder.

Forum member Estruch added: “I only would like to add that when we analyze pancreatic disease, the pattern of drinking may be more important than the amount of alcohol intake.  References to the pattern of drinking and/or binge drinking are missing in the paper and in my opinion should be included as a limitation of the study.”   And reviewer de Gaetano stated: “In particular I support Estruch’s comments on the pattern of drinking.  Further, I was intrigued by the possible J-shaped dose-response curve, at least in women.  I wonder whether comparison with control groups including only life-time abstainers would not have evidentiated a more clear,  though minimal, protective effect of low-moderate doses of wine or beer?”

Forum member Lanzmann-Petithory commented: “In an analysis of 35,292 men from the COLOR cohort in the Eastern part of France, among whom 183 deaths from pancreatic cancer occurred,  the only significant factors we found for pancreatic cancer death were smoking > 20 cigarettes / day and age (Lanzmann-Petithory et al).  There were non-significant trends for increased BMI and for being a high school graduate.  Concerning alcohol, in the COLOR cohort, compared with abstainers, at 10 g/day the RR = 0.86, CI 0.42-1.74; at 40 g/day the RR = 1.03, CI 0.49-2.19; for 80 g/day, the RR = 1.17, CI 0.54-2.54; and for > 100 g/day the RR = 1.69, CI 0.68-4.17.  Thus, the only trends that suggested an increase in cancer risk was associated with quite large amounts of alcohol.  Further, there was no relation between the consumption of wine and cancer in the COLOR study, as was shown similarly in the present paper.  Overall, scientific data strongly indicate that moderate drinking is not a risk factor for pancreatic cancer.”  Stated reviewer Van Velden: “This new study confirms our previous beliefs, even what I was taught in medical school in 1963!”

Reviewer Thelle stated: “Even if pancreatic cancer has been shown to be associated with obesity, diabetes and smoking, it still remains a causal enigma.  The present paper is appreciated as it contributes to the clarification of the role of alcohol.  Still, as is usually the case, more targeted research is needed to confirm the particular effect of liquor and assess the cumulative effect of life-time exposure to alcohol.”  Added Forum member Skovenborg: “I agree with the comments of others, especially those of Thelle regarding the particular effect of liquor.  In my view, this association could generate hypotheses regarding the importance of drinking patterns and possible confounding by other lifestyle factors.”  Reviewer Finkel commented: “I was wondering whether pancreatitis might be either an intermediate step in the genesis of (excessive) alcohol-inspired pancreatic cancer, or at least a marker of individuals who were at greater risk.”

Forum members note that in the present paper the significant increase in risk was only among men, with no significant effect of alcohol found among women.  Overall, the study suggests that heavy consumption of liquor may be the primary reason for a positive association of “alcohol” with pancreatic cancer.  And heavy liquor consumption is known to increase the risk of chronic pancreatitis, one potential mechanism of effect on the risk of cancer (Dufour & Adamson).  The Forum also notes that the study shows no significant association with cancer risk for any level of consumption of beer or wine, which could relate to the lower concentration of alcohol per volume of these beverages as compared with liquor (Devos-Comby & Lange; Heinen et al; Zheng et al).  It could also relate to non-alcoholic substances (such as polyphenols, present in wine and beer) or even to different drinking practices among subjects consuming different beverages.

While the authors of this paper conclude that “Our study suggests that high alcohol intake, especially liquor intake, might play an important role in the risk of pancreatic cancer,” the Forum notes that the overall increase in risk associated with heavy drinking was not large, a 15% increase (RR = 1.15, 95% CI 1.06 – 1.25).  However, given that pancreatic cancer is so deadly, usually being incurable by the time it is diagnosed, any lifestyle factor associated with the disease should be recognized.

References from Forum review

Devos-Comby L, Lange JE. “My drink is larger than yours”? A literature review of self-defined drink sizes and standard drinks. Curr Drug Abuse Rev 2008;1:162–176.

Dufour MC, Adamson MD. The epidemiology of alcohol-induced pancreatitis. Pancreas 2003;27:286–290.

Heinen MM, Verhage BAJ, Ambergen TAJ, et al. Alcohol consumption and risk of pancreatic cancer in the Netherlands Cohort Study. Am J Epidemiol 2009;169:1233–1242.

Lanzmann-Petithory D, Renaud S, et al.  Unpublished data.  Data presented at meeting of National Research Agency, Paris, France.

Zheng W, McLaughlin JK, Gridley G, et al. A cohort study of smoking, alcohol consumption, and dietary factors for pancreatic cancer (United States). Cancer Causes Control 1993;4:477–482.

Forum Summary

The present meta-analysis was based on data from more than four million subjects in prospective cohort studies, among whom 11,846 incident cases of pancreatic cancer were diagnosed.  With the lowest intake group (non-drinkers or occasional drinkers) as the referent group, the authors defined “light” consumption as up to 12 grams/day (essentially one typical drink); 12-24 g/day as “moderate”; and ≥24 g/day as “heavy” drinking.  The key results of the study were that, overall, neither light drinkers (RR = 0.97) nor moderate drinkers (RR = 0.98) showed an increase in risk of pancreatic cancer, while subjects classified as heavy drinkers had a slight increase (RR= 1.15, 95% CI 1.06 – 1.25).  The increase in risk was due to heavy drinkers of liquor, as there was no significant increase in risk even for heavy drinkers of beer (RR = 1.08, CI 0.90 – 1.30) or wine (RR = 1.09, CI 0.79 -1.49).

To summarize, Forum members considered this to be an excellent paper on the association between alcohol consumption and pancreatic cancer.  The authors used appropriate methods and limited subjects to those in prospective cohort studies, which would tend to limit bias. The paper shows that the significant increase in risk occurred only among men, with no significant effect of alcohol being found among women.  Among the weaknesses of the study were that there was a mixture of lifetime abstainers and ex-drinkers included in the referent group, and the same cut-points for category of alcohol intake was used for both men and women, whereas drinking guidelines are generally lower for women than for men.  Further, data on the pattern of drinking (regular versus binge) were not available.  The study showed no significant association with cancer risk for any level of consumption of beer or wine, which could relate to their lower concentration of alcohol per volume of the beverage, to non-alcoholic substances (such as polyphenols, present in wine and beer), or even to different drinking practices among subjects consuming different beverages.  Thus, Forum members agree with the conclusions of the authors that heavy alcohol consumption, especially of liquor, increases the risk of pancreatic cancer, but the intake of beer or wine may not be associated with an increased risk.

Reference:  Wang Y-T, Gou Y-W, Jin WW, Xiao M, Fang H-Y.  Association between alcohol intake and the risk of pancreatic cancer: a dose–response meta-analysis of cohort studies.  BMC Cancer 2016;16:212.  DOI 10.1186/s12885-016-2241-1

*                        *                      *                      *                      *                      *

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

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

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

R. Curtis Ellison, MD, Professor of Medicine & Public Health, Boston University School of Medicine, Boston, MA, USA

Ramon Estruch, MD, PhD.  Associate Professor of Medicine, University of Barcelona, Spain

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

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

Fulvio Mattivi, MSc, 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

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

Arne Svilaas, MD, PhD, general practice and lipidology, Oslo University Hospital, Oslo, Norway