American Journal of Epidemiology
ª The Author 2008. Published by the Johns Hopkins Bloomberg School of Public Health.
All rights reserved. For permissions, please e-mail: firstname.lastname@example.org.
Vol. 168, No. 8
Advance Access publication September 8, 2008
Risk of Pancreatitis According to Alcohol Drinking Habits: A Population-based
Louise Kristiansen, Morten Grønbæk, Ulrik Becker, and Janne Schurmann Tolstrup
Received for publication May 6, 2008; accepted for publication June 25, 2008.
The association between alcohol intake and pancreatitis has been examined previously in case-control studies,
mostly consisting of men. The significance of beverage type and drinking pattern is unknown. The objective of this
study was to assess the association between amount, type, and frequency of alcohol intake and risk of pancreatitis.
For this purpose, the authors used data on 17,905 men and women who participated in the Copenhagen City Heart
Study in 1976–1978, 1981–1983, 1991–1994, and 2001–2003 in Copenhagen, Denmark. Alcohol intake and
covariates were assessed by questionnaire. Information on pancreatitis was obtained from national registers.
A high alcohol intake was associated with a higher risk of pancreatitis. Hazard ratios associated with drinking
1–6, 7–13, 14–20, 21–34, 35–48, and >48 drinks/week were 1.1 (95% confidence interval (CI): 0.8, 1.6), 1.2 (95%
CI: 0.8, 1.8), 1.3 (95% CI: 0.8, 2.1), 1.3 (95% CI: 0.7, 2.2), 2.6 (95% CI: 1.4, 4.8), and 3.0 (95% CI: 1.6, 5.7),
respectively, compared with 0 drinks/week (Ptrend< 0.001). Associations were similar for men and women.
Drinking frequency did not seem to be independently associated with pancreatitis.
alcohol drinking; alcoholic beverages; alcohol-induced disorders; pancreatitis; prospective studies
Abbreviations: CI, confidence interval; ICD, International Classification of Diseases.
As early as 1878, alcohol was proposed as a risk factor for
pancreatitis (1), and it is now considered well known that
alcohol increases the risk of pancreatitis. However, epide-
miologic studies on the quantitativeaspect of the association
between alcohol intake and pancreatitis are sparse. To date,
only 4 case-control studies (1–4) and 1 ecologic study (5) on
this subject have been published. In all these studies, an
association was found between alcohol intake and risk of
pancreatitis in men; however, only one of the case-control
studies included women, where surprisingly, no increased
risk of pancreatitis according to alcohol was observed (2).
No studies have assessed the risk of pancreatitis associated
with other dimensions of alcohol intake, such as beverage
type and drinking frequency. It is also uncertain whether
a threshold exists, that is, a level of alcohol intake under
which the risk of pancreatitis is not increased.
In addition to alcohol, gallstone disease is thought to be
an important risk factor for pancreatitis (6) and, because
studies indicate that a moderate intake of alcohol could pro-
tect against gallstone disease (7, 8), the association between
alcohol and risk of pancreatitis might be tempered by gall-
In this study, we examined the association between alco-
hol intake and risk of pancreatitis in a large prospective
cohortconsisting ofmenand womenfromthegeneral Danish
population. Furthermore, we aimed at addressing whether
there may be specific effects of beverage type or drinking
MATERIALS AND METHODS
The data used in this study came from the 4 examinations
of the Copenhagen City Heart Study, performed in 1976–
1978, 1981–1983, 1991–1994, and 2001–2003. Enrolment
and examination procedures have been described in more
Correspondence to J. Tolstrup, National Institute of Public Health, Oster Farimagsgade 5a, DK-1399 Copenhagen K, Denmark (e-mail:
932Am J Epidemiol 2008;168:932–937
by guest on June 5, 2013
detail elsewhere (9, 10). Subjects with pancreatitis before
baseline or missing information on alcohol intake were ex-
cluded, which left 17,905 participants eligible for further
analysis. The study was approved by the ethical committees
for the Copenhagen area (approval reference number: KF
Beer, wine, and spirits were categorized as 0, 1–6, 7–13,
and ?14 drinks/week and total alcohol intake as 0, 1–6,
7–13, 14–20, 21–34, 35–48, and >48 drinks per week. Be-
cause drinking frequency was assessed separately for beer,
wine, and spirits, we could not determine overall drinking
frequency with certainty. We performed exploratory anal-
yses defining drinking frequency from a combination of in-
formation on beverage type and beverage type-specific
drinking frequency, categorizing drinking frequency as rare
drinkers, monthly drinkers, weekly drinkers, almost daily
drinkers, and daily drinkers.
Smoking (never smoker, former smoker, and smoker of
1–14 g, 15–24 g, and >24 g of tobacco/day), body mass
index (<20, 20–24, and ?25 kg/m2), physical activity (sed-
entary, light, moderate, and heavy), school education (<8,
8–11, and >11 years of education, corresponding to lower
primary school, higher primary school, and secondary
school), and income (low, middle, and high income, corre-
sponding to <$30,000, $30,000–$80,000, and >$80,000/
year in Denmark in 1991–1994) were considered to be po-
tential confounders. Information on gallstone disease was
obtained from the Danish Hospital Discharge Register,
which contains data on all hospital admissions in Denmark.
Information on acute and chronic pancreatitis was ob-
tained from the Danish Hospital Discharge Register and
the Danish Register of Causes of Death. For acute pancre-
atitis, the relevant International Classification of Diseases
(ICD), Eighth Revision, codes were 577.00–577.04, 577.08,
and 577.09, and the ICD, Tenth Revision, code was K85.9.
For chronic pancreatitis, the relevant ICD, Eighth Revision,
codes were 577.19 and 577.90–577.92, and the ICD, Tenth
Revision, codes were K86.0–K86.3, K86.8, and K86.9.
Participants accrued person-time from the time of their
first participation in the Copenhagen City Heart Study until
the time of pancreatitis diagnosis, date of death, emigration,
or end of follow-up (July 9, 2007), whichever occurred first.
We had follow-up information on 100% of the study partic-
ipants. Data were analyzed by means of the Cox propor-
tional hazards regression model, with delayed entry
implemented by using SAS/STAT, version 9.1, software
(SAS Institute, Inc., Cary, North Carolina). Age (in days)
was used as the underlying time axis.
Analyses were performed by using updated information on
alcohol and covariates. Tests for linear trend were performed
by treating the median value within categories of alcohol in-
take as a continuous variable and adding this to the model.
Using fractional polynomials, we performed analyses to
study the shape of the risk curve to see if there was any
Baseline Characteristics by AlcoholConsumption in 18,035Danish Men and WomenParticipatingin the Copenhagen City Heart Study,
% of total intake
Mean Age, years
% <8 years
02,95355 (33, 73)541262212649
1–64,244 32581049 (25, 68)571638211427
7–131,55434 541348 (26, 70)591827231525
14–20488 34551150 (31, 69)6718312420 23
>2033433472050 (33, 68)751133242526
094854 (32, 75)62225526 2540
1–61,85653 33 1350 (25, 71)62 20 41 2318 26
7–131,989 56 301550 (24, 70) 65 2039 2516 21
14–20 1,38362 251351 (26, 70)672237251520
21–341,11963241351 (30, 69)75 1743221919
35–485666619 1551 (33, 69) 78 16 4926 23 19
>48 47175111450 (24, 67)841150263227
Pancreatitis and Drinking Habits 933
Am J Epidemiol 2008;168:932–937
by guest on June 5, 2013
indication of a threshold effect (11). The model with the best
fit was a model including alcohol as linear and square terms.
To examine the possibility that latent baseline symptoms
of pancreatitis might have reduced the amount of alcohol
consumed, thereby biasing results, we carried out analyses
in which the first 2 or 4 years of observation time were
excluded, and information was updated with a delay of 2
or 4 years, respectively.
The test for interaction between alcohol intake and smok-
ing was performed by a nested log likelihood test, compar-
ing a model containing the variables as single terms with
a model also including the interaction terms. For this pur-
pose, alcohol was categorized as <7, 7–20, and >20 drinks/
smokers, and current smokers.
Table 1 shows the characteristics of the 9,573 women and
8,332 men participating in this study, categorized by amount
of alcohol intake (assessed on the time of their first partic-
ipation in the Copenhagen City Heart Study). Among both
women and men, those in the higher alcohol categories were
more likely to smoke, whereas individuals in the lowest
alcohol category tended to be older, to have fewer years of
education, and to have lower income levels.
Amount of alcohol intake and risk of pancreatitis
The mean follow-up time in this study was 20.1 years
(range, 0–31 years). At the end of follow-up, 235 partici-
pants (113 women and 122 men) had developedpancreatitis,
and there were 171 cases of acute and 97 cases of chronic
An increasing risk of pancreatitis according to alcohol
intake was observed in both men and women, although this
was statistically insignificant in women (data not shown).
We performed further analyses including men and women in
the same model (Pinteraction between sex and alcohol in
a nested log likelihood test ¼ 0.83). A high alcohol intake
was associated with an increased risk of both acute and
chronic pancreatitis (Table 2). The hazard ratio for acute
and chronic pancreatitis combined (total pancreatitis) in-
creased by 1.13 for every additional drink/day (95% confi-
dence interval (CI): 1.06, 1.21). In multivariate-adjusted
models, smoking was responsible for most of the effect of
adjustment. The fully adjusted risk of pancreatitis in women
compared with men was 0.9 (95% CI: 0.7, 1.2). Including
variables for personal income and physical activity in the
adjusted model had little effect on risk estimates.
Separating never drinkers from the nondrinkers was not
possible in this study, and it is hence possible that the cat-
egory of nondrinkers contains participants with a previously
high intake, resulting in a falsely high incidence rate of
pancreatitis in this category. To address this issue, analyses
were repeated, separating nondrinkers from consistent non-
drinkers, that is, participants who participated in at least 2
examinations and reported no alcohol intake at every exami-
nation. During follow-up, 23 cases occurred in the category of
consistent nondrinkers. The hazard ratios for total pancreatitis
Risk of Acute Pancreatitis, Chronic Pancreatitis, and Total Pancreatitis According to Updated Consumption of Alcohol Intake, Copenhagen, Denmark, 1976–2007
aAdjusted for age and sex.
bAdjusted for age, sex, smoking, education, and body mass index.
934Kristiansen et al.
Am J Epidemiol 2008;168:932–937
by guest on June 5, 2013
using consistent nondrinkers as the reference category were
similar to results from the main analysis (data not shown).
Excluding the first 2 or 4 years of observation time re-
vealed results similar to those from the main analyses (data
Modeling the association between alcohol and risk of
total pancreatitis for men and women separately and to-
gether, respectively, using fractional polynomials did not
indicate that there was a threshold in the risk of pancreatitis
according to alcohol (data not shown). The curves flattened
out at high intakes, but the test for linearity did not provide
evidence for departure from linearity (P ¼ 0.14 for women
and men combined).
Compared with that among never smokers, the hazard
ratio for pancreatitis was 1.6 (95% CI: 1.0, 2.5) among past
smokers and 1.5 (95% CI: 0.9, 2.3), 2.3 (95% CI: 1.5, 3.6),
and 3.1 (95% CI: 1.8, 5.3) among current smokers of 1–14,
15–24, and >24 g of tobacco/day, respectively. We found no
evidence of interaction between alcohol intake and smoking
(P ¼ 0.57).
Gallstone disease, alcohol, and pancreatitis
variable,the hazardratiosof pancreatitisaccordingtoalcohol
were slightly increased: For example, the risk associated with
drinking >48 drinks/week increased from 3.0 (95% CI: 1.6,
5.7) to 3.6 (95% CI: 1.9, 6.9) when including gallstones in
the model. In addition, the hazard ratio for continuous alco-
hol intake increased from 1.13 (95% CI: 1.06, 1.21) to 1.15
(95% CI: 1.08, 1.23) per drink/day. The hazard ratio of
pancreatitis according to gallstone disease was 11 (95%
CI: 7.5, 15).
Analyses of beverage type and drinking frequency
We explored the relation between type of alcoholic bev-
erage and risk of pancreatitis (Table 3). The adjusted hazard
ratio for drinking more than 14 beers/week was 2.0 (95%
CI: 1.3, 3.1). No association was observed regarding wine
and spirits. However, there were only a few participants in
the highest categories of both wine and spirits.
We found no evidence that drinking frequency was asso-
ciated with risk of pancreatitis. In a comparison with never
drinkers, the adjusted hazard ratio was 0.8 (95% CI: 0.5,
1.3), 1.0 (95% CI: 0.7, 1.5), 1.1 (95% CI: 0.6, 1.8), and
1.3 (95% CI: 0.9, 1.9) for monthly, weekly, almost daily,
and daily alcohol drinking. With further adjustment for the
amount of alcohol intake (as a continuous variable), the
hazard ratio for daily drinking attenuated to 1.0 (95% CI:
0.6, 1.5). In this model, amount of alcohol intake remained
significantly associated with increased risk of pancreatitis
(P < 0.001), indicating that amount of alcohol is more
important than drinking frequency for the risk associated
with alcohol drinking.
In this prospective cohort study, we found that a high
alcohol intake was associated with increased risk of pancre-
atitis. Drinking frequency appeared to be unassociated with
the risk of pancreatitis. Further, our results indicate that
gallstones slightly temper the association between alcohol
Limitations include a lack of information on potential
confounders, such as diet and coffee (4, 12). Furthermore,
the diagnosis of pancreatitis has not been validated. In the
Danish Hospital Discharge Register, only hospital admis-
sions and not contacts with general practitioners are regis-
tered; hence, it is likely that we do not have information on
all cases occurring during the study period. Moreover, mis-
classification between the diagnoses of acute and chronic
pancreatitis has most likely occurred, as the symptoms and
diagnostic criteria are overlapping and the 2 diseases can
coexist (13). Such misclassification would result in similar
observed associations between alcohol and risk of acute and
chronic pancreatitis. Therefore, for these reasons and be-
cause of the limited number of cases, we recommend that
readers do not place too much emphasis on this result. How-
ever, results for the joint outcome of pancreatitis are valid.
We found an increased risk of pancreatitis when drinking
>14 drinks of beer/week, whereas no association was ob-
served for wine or spirits. For wine, this result could hypo-
thetically be due to a protective effect of the antioxidants
Consumption of Individual Alcoholic Beverages, Copenhagen,
Risk of Total Pancreatitis According to Updated
0 871.0Referent 1.0Referent
1–6 741.30.9, 220.127.116.11, 1.8
7–13 251.30.8, 18.104.22.168, 2.0
?14492.51.6, 3.8 2.01.3, 3.1
0117 1.0Referent 1.0Referent
1–6 991.00.7, 1.41.1 0.8, 1.5
7–13110.7 0.4, 22.214.171.124, 1.5
?1480.80.4, 126.96.36.199, 1.8
0 1451.0Referent1.0 Referent
1–6 660.80.6, 1.1 0.70.5, 1.0
7–13 16 1.10.6, 1.81.00.6, 1.7
?148 1.00.5, 2.00.9 0.4, 1.8
aAdjusted for age, sex, and other beverage types.
bAdjusted for age, sex, smoking, education, body mass index, and
other beverage types.
Pancreatitis and Drinking Habits935
Am J Epidemiol 2008;168:932–937
by guest on June 5, 2013
found in wine. However, the validity of these results is se-
verely limited by the small number of cases in the high cat-
egories of wine and spirits and by the fact that wine drinking
is associated with a generally healthier lifestyle (14).
Results on drinking frequency are limited by the fact that
our measure of this drinking variable may be too crude to
pick up small effects. Unfortunately, information on binge
drinking (i.e., drinking a minimum number of drinks per
occasion) was not available, and we cannot comment on
this aspect with the present data.
Our data did not suggest a clear threshold effect of alcohol
intake on the risk of pancreatitis. However, this might be due
to misclassification. In the case of underreporting, the risk in
the reference category would increase, and this could lead to
a true threshold amount being either lowered or blurred.
Strengths include the large study size, the prospective
design, and the fact that we had complete follow-up infor-
mation on all 17,905 participants. In addition, to the best of
our knowledge, this is the first study to investigate the as-
sociation between alcohol intake and pancreatitis prospec-
tively, even though it is considered well known that alcohol
is a strong risk factor for pancreatitis.
Our findings are consistent with previous results from
case-control studies (1–4), although the literature on alcohol
and pancreatitis is sparse, and only one previous case-control
study has included women (2). In that study, the risk of
pancreatitis was not associated with alcohol among women;
however, results were limited by the few female cases.
Different mechanisms have been suggested to explain the
toxic effect of alcohol. The pancreas can degrade alcohol by
both oxidative and nonoxidative metabolism— mechanisms
involving the syntheses of acetaldehyde and fatty-acid eth-
anol esters, respectively (15). The latter (i.e., fatty-acid eth-
anol esters) have been demonstrated to cause pancreatic
edema, intracellular trypsin activation, and the induction
of proinflammatory transcription factors in animal studies
(16–18). In addition, the toxic effect of alcohol may be due
to the induction of oxidative stress, that is, the imbalance
between formation and neutralization of reactive oxygen
species (19–21). On the other hand, there is some evidence
that a moderate alcohol intake protects against gallstone
disease (7, 8), which would lower the risk of pancreatitis.
This is in accordance with our finding that the risk of pan-
creatitis according to alcohol was increased when including
gallstones in the model.
In summary, we found that a high alcohol intake was
associated with an increased risk of pancreatitis in both
men and women, whereas drinking frequency did not seem
to be a risk factor for pancreatitis when the amount of alco-
hol intake was taken into consideration.
Author affiliations: Centre for Alcohol Research, National
Institute of Public Health, University of Southern Denmark,
Janne Schurmann Tolstrup); and Alcohol Unit, Hvidovre Hos-
pital, Copenhagen, Denmark (Ulrik Becker).
The study is supported by grants from the Danish
National Board of Health and the Danish Medical Research
Conflict of interest: none declared.
1. Durbec JP, Sarles H. Multicenter survey of the etiology of
pancreatic diseases. Relationship between the relative risk of
developing chronic pancreatitis and alcohol, protein, and lipid
consumption. Digestion. 1978;18(5–6):337–350.
2. Yen S, Hsieh CC, MacMahon B. Consumption of alcohol and
tobacco and other risk factors forpancreatitis. Am J Epidemiol.
3. Talamini G, Bassi C, Falconi M, et al. Alcohol and smoking as
risk factors in chronic pancreatitis and pancreatic cancer. Dig
Dis Sci. 1999;44(7):1303–1311.
4. Lin Y, Tamakoshi A, Hayakawa T, et al. Associations of al-
cohol drinking and nutrient intake with chronic pancreatitis:
findings from a case-control study in Japan. Am J Gastroen-
5. Ramstedt M. Alcohol and pancreatitis mortality at the popula-
tion level: experiences from 14 Western countries. Addiction.
6. Steinberg W, Tenner S. Acute pancreatitis. N Engl J Med.
7. Cuevas A, Miquel JF, Reyes MS, et al. Diet as a risk factor for
cholesterol gallstone disease. J Am Coll Nutr. 2004;23(3):
8. Kono S, Eguchi H, Honjo S, et al. Cigarette smoking, alcohol
use, and gallstone risk in Japanese men. Digestion. 2002;
9. The Copenhagen City Heart Study. Osterbroundersøgelsen.
A book of tables with data from the first examination (1976–78)
and a five year follow-up (1981–83). The Copenhagen City
Heart Study Group. Scand J Soc Med. 1989;170(suppl 41):
10. Schnohr P, Jensen G, Lange P, et al. The Copenhagen City
Heart Study. Østerbroundersøgelsen: tables with data from the
third examination 1991–1994. Eur Heart J. 2001;3(suppl H):
11. Royston P, Ambler G, Sauerbrei W. The use of fractional
polynomials to model continuous risk variables in epidemiol-
ogy. Int J Epidemiol. 1999;28(5):964–974.
12. Morton C, Klatsky AL, Udaltsova N. Smoking, coffee, and
pancreatitis. Am J Gastroenterol. 2004;99(4):731–738.
13. Kloppel G, Maillet B. Development of chronic pancreatitis
from acute pancreatitis: a pathogenetic concept. (In German).
Zentralbl Chir. 1995;120(4):274–277.
14. Tjonneland AM, Gronbaek MN, Stripp C, et al. The connec-
tion between food and alcohol intake habits among 48.763
Danish men and women. A cross-sectional study in the project
‘‘Food, cancer and health.’’ (In Danish). Ugeskr Laeger. 1999;
15. Lange LG. Nonoxidative ethanol metabolism: formation of
fatty acid ethyl esters by cholesterol esterase. Proc Natl Acad
Sci U S A. 1982;79(13):3954–3957.
16. Gukovskaya AS, Mouria M, Gukovsky I, et al. Ethanol me-
tabolism and transcription factor activation in pancreatic acinar
cells in rats. Gastroenterology. 2002;122(1):106–118.
17. Werner J, Saghir M, Warshaw AL, et al. Alcoholic pancreatitis
in rats: injury from nonoxidative metabolites of ethanol. Am J
Physiol Gastrointest Liver Physiol. 2002;283(1):G65–G73.
936Kristiansen et al.
Am J Epidemiol 2008;168:932–937
by guest on June 5, 2013