Critique 092: The association of alcohol and tobacco with age at diagnosis among subjects with pancreatic cancer — 2 October 2012

Anderson MA; Zolotarevsky E; Cooper KL; Sherman S; Shats O; Whitcomb DC; Lynch HT; et al.  Alcohol and tobacco lower the age of presentation in sporadic pancreatic cancer in a dose-dependent manner: A multicenter study.  Am J Gastroenterol 2012; advance online publication, doi: 10.1038/ajg.2012.288

Authors’ Abstract

OBJECTIVES: The objective of this study was to examine the association between tobacco and alcohol dose and type and the age of onset of pancreatic adenocarcinoma (PancCa).

METHODS: Prospective data from the Pancreatic Cancer Collaborative Registry were used to examine the association between age of onset and variables of interest including: gender, race, birth country, educational status, family history of PancCa, diabetes status, and tobacco and alcohol use.  Statistical analysis included logistic and linear regression, Cox proportional hazard regression, and time-to-event analysis.

RESULTS: The median age to diagnosis for PancCa was 66.3 years [95 % confidence intervals (CIs), 64.5 – 68.0].  Males were more likely than females to be smokers (77 % vs. 69 % , P = 0.0002) and heavy alcohol and beer consumers (19 % vs. 6 % , 34 % vs. 19 % , P < 0.0001).  In univariate analysis for effects on PancCa presentation age, the following were significant: gender, alcohol and tobacco use (amount, status and type), family history of PancCa, and body mass index.  Both alcohol and tobacco had dose-dependent effects.  In multivariate analysis, alcohol status and dose were independently associated with increased risk for earlier PancCa onset with greatest risk occurring in heavy drinkers (HR 1.62, 95 % CI 1.04 – 2.54).  Smoking status had the highest risk for earlier onset pancreatic cancer with a HR of 2.69 (95 % CI, 1.97 – 3.68) for active smokers and independent effects for dose (P = 0.019).  The deleterious effects for alcohol and tobacco appear to resolve after 10 years of abstinence.

CONCLUSIONS: Alcohol and tobacco use are associated with a dose-related increased risk for earlier age of onset of PancCa.  Although beer drinkers develop pancreatic cancer at an earlier age than nondrinkers, alcohol type did not have a significant effect after controlling for alcohol dose.

Forum Comments

Background:  Pancreatic adenocarcinoma (PancCa) is a deadly disease, with essentially 100% mortality.  Screening for the early detection of such cancer has not been shown to be feasible, and is currently not advised for asymptomatic people.1  Except for a genetic link for a small percentage of patients who have familial disease, the causes of PancCa are not known.  Among environmental factors that have been implicated in some studies are smoking, alcohol use, high-fat diet, excessive intake of carbonated soft drinks, obesity, chronic pancreatitis, and diabetes.

Previous epidemiologic studies have had conflicting results, with some showing an increase in risk among consumers of alcohol, especially among heavy drinkers,2,3 but most showing no significant effect of moderate drinking.2,4  A monograph from IARC in 2009 concluded that there was insufficient evidence to support a role of alcohol in pancreatic cancer development.5

The present study does not evaluate the etiology of PancCa, as it deals only with cases of the disease and has no normal controls.  Instead, it reports the age of onset of PancCa according to exposures to tobacco and alcohol.  It concludes that smokers and drinkers, especially heavy drinkers, have the onset of disease at an earlier age than do non-smokers and non-drinkers.  Subjects who had stopped smoking or alcohol consumption for more than 10 years had the same age at onset of their cancers as lifetime non-smokers and non-drinkers.

Specific comments on the paper:  Overall, 32% of subjects were never drinkers; 26% of subjects were never smokers.  The authors used the following cutpoints for alcohol consumption categories: “mild” = ≤ 13 grams of alcohol/day (about 1 typical US drink); “moderate (>14 – 39 g/day, up to about 2 or 3 drinks); “heavy (> 39 g of alcohol/day, more than about 3 drinks/day by US standards or 4 drinks/day by UK standards).  Unfortunately, the authors were not able to evaluate the effects according to the pattern of drinking.

The analyses appear to be appropriate.  Most of the results presented are based on univariate associations, not adjusted for confounding.  In multivariate analysis there was a tendency for the age at diagnosis of PancCa to be earlier for obese subjects (HR=1.42; CI 1.06, 1.89) and for current smokers and ex-smokers who had stopped less than 10 years earlier.  In comparison with never smokers, past smokers within < 10 years of stopping had an estimated 64% increase in risk of earlier diagnosis of cancer while the calculated hazard ratio was more than doubled (HR=2.69; CI=1.97, 3.68) for active smokers.  Interestingly, subjects who had stopped smoking for more than 10 years had the same risk as lifetime non-smokers.

In terms of alcohol consumption, in comparison with never drinkers the estimated hazard ratios (HR) for an earlier diagnosis of cancer were 0.95 (CI=0.54, 1.68) for ex-drinkers who had quit drinking more than 10 years earlier, 1.27 (CI =0.66, 2.43) for ex-drinkers who had stopped drinking fewer than 10 years earlier, and 1.62 (CI=0.96, 2.81) for subjects reporting that they were consuming alcohol at entry into the study (when their cancer was diagnosed).  In multivariate analysis, the authors state that the only statistically significant finding was comparing heavy versus mild or moderate drinking, where the HR=1.62, CI=1.04, 2.54.

For beverage-specific analyses, with adjustment for total alcohol consumed, there were no significant effects (HRs were 1.06 for beer, 0.88 for wine, 0.90 for spirits).

Potential bias in assigning a specific time for diagnosis of pancreatic cancer:  As pointed out by several Forum members, the clinical onset of pancreatic cancer is a very soft event in the natural history of the disease, localized or metastatic.  Thus, the point in time when the diagnosis is made in a patient who has developed PancCa could well be affected by socioeconomic, educational, and other lifestyle factors.  For example, more affluent persons (who, especially in the southern and mid-western American sites involved in this study, are more likely to consume alcohol) may seek investigation of mild symptoms much earlier than would be sought by poor people with limited access to medical care.  In addition, subjects who are heavy smokers or heavy drinkers are much more likely (than non-smokers and non-heavy drinkers) to be admitted to hospital for other diseases, and thus would be more likely to have medical investigations (x-rays, CT scans, MRI exams, etc.) carried out that might lead to the diagnosis of a previously undetected PancCa.  Such scenarios could bias the estimates of the age of onset of cancer according to alcohol consumption reported in this paper. 

Other limitations of the study:  Forum reviewers agreed with the many limitations of the study that were acknowledged by the authors.  These include not having normal controls (that prevents evaluation of alcohol and smoking as causative factors of these cancers), and their inability to evaluate other potential factors such as coffee exposure, exposure to second-hand smoke, specific origin of tumor in the pancreas, or a history of chronic pancreatitis. 

Forum reviewers were especially concerned that the authors did not have the ability to judge the effects of the pattern of drinking.  The calculated weekly average of “drinks per day” includes subjects primarily binge drinking on the week-end as well as subjects with regular moderate intakes.  Typically, even when the average weekly intake of alcohol may be the same, blood alcohol levels are much higher following binge drinking than with regular moderate drinking.  Because of this and other limitations, one Forum reviewer concluded that “This study is likely to have only a small impact on clinical gastroenterology.”


1.  Calonge N.  Screening for Pancreatic Cancer. AHRQ Pub No 05-0554-A [serial on the Internet] 2004.

2.  Gupta S, Wang F, Holly EA, et al.  Risk of pancreatic cancer by alcohol dose, duration, and pattern of consumption, including binge drinking: a population-based study.  Cancer Causes Control 2010;21:1047–1059.

3.  Michaud DS, Vrieling A, Jiao L, et al.  Alcohol intake and pancreatic cancer: a pooled analysis from the pancreatic cancer cohort consortium (PanScan).Cancer Causes Control 2010;21:1213–1225.

4.  Lucenteforte E, La Vecchia C, Silverman D, et al.  Alcohol consumption and pancreatic cancer: a pooled analysis in the International Pancreatic Cancer Case-Control Consortium (PanC4). Ann Oncol 2012;23:374 – 382 .

5.  Secretan B, Straif K, Baan R, et al.  A review of human carcinogens — Part E: tobacco, areca nut, alcohol, coal smoke, and salted fish.  Lancet Oncol 2009;10:1033–1034.

Forum Summary

This analysis from a group of distinguished scientists supports previous research showing that smoking is associated with an earlier onset of pancreatic adenocarcinoma.  Other research has shown that smoking may also be a causative factor in the development of this type of cancer.  For alcohol, this study shows that heavy drinking also appears to be associated with earlier diagnosis; previous research from some epidemiologic studies has suggested further that heavy intake of alcohol may be associated with the development of pancreatic cancer. 

Forum reviewers were concerned, however, about potential bias in regards to the time of diagnosis of such cancer, and about many other limitations of the study.  For example, the inability to separate drinkers by the pattern of drinking (binge drinking versus regular, moderate intake), by the socioeconomic status of subjects, and by a lack of information on chronic pancreatitis, a known risk factor, weaken the implications of this paper.  Forum members do not think that the results of this study will have a large impact on clinical practice, or on measures for the prevention of pancreatic cancer. 

                                                                   *                      *                      *

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

Francesco Orlandi, MD, Dept. of Gastroenterology, Università degli Studi di Ancona. Italy

Erik Skovenborg, MD, Scandinavian Medical Alcohol Board, Practitioner, Aarhus, Denmark

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

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

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

Yuqing Zhang, MD, DSc, Epidemiology, Boston University School of Medicine, Boston, MA, USA

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