Statins and Pancreatic Cancer Risk in Patients with ... · INTRODUCTION Chronic pancreatitis is an...
Transcript of Statins and Pancreatic Cancer Risk in Patients with ... · INTRODUCTION Chronic pancreatitis is an...
Statins and Pancreatic Cancer Risk in Patients with Chronic Pancreatitis: A
Danish nationwide population-based cohort study
Jakob Kirkegård1,2, Jennifer L. Lund2,3, Frank Viborg Mortensen1, Deirdre Cronin-Fenton2
1 Department of Surgery, Section for Hepato-Pancreato-Biliary Surgery, Aarhus University
Hospital, Denmark
2 Department of Clinical Epidemiology, Aarhus University Hospital, Denmark
3 Department of Epidemiology, Gillings School of Global Public Health, University of North
Carolina at Chapel Hill, North Carolina, USA
Correspondence: Jakob Kirkegård, Department of Surgery, HPB Research Unit, Aarhus
University Hospital, Palle Juul-Jensens Boulevard 99, DK-8200 Aarhus N, DENMARK
Phone: +45 78 45 00 00 / E-mail: [email protected]
Short tile: Statins, chronic pancreatitis, and pancreatic cancer.
Category: Cancer Epidemiology
Keywords: statins; pancreatitis; pancreatic cancer; epidemiology; risk factors
Novelty and impact: Statins are widely used prescription drugs that may have anti-cancer
properties. In the present nationwide population-based cohort study of 8,311 chronic
pancreatitis patients, exposure to statins was not associated with an increased or decreased
risk of pancreatic cancer. This finding that was robust to several sensitivity analysis and
Conflicts of Interest: Dr. Lund’s spouse is a full-time, paid employee of GlaxoSmithKline. Dr. Kirkegård, Dr. Mortensen, and Dr. Cronin-Fenton have nothing to disclose.
This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process which may lead to differences between this version and the Version of Record. Please cite this article as doi: 10.1002/ijc.32264
This article is protected by copyright. All rights reserved.
underlines that prevention of pancreatic cancer in patients with chronic pancreatitis, a high-
risk group, is still a major challenge.
This article is protected by copyright. All rights reserved.
ABSTRACT
Statins (HMG-CoA reductase inhibitors) have anti-inflammatory and possibly anti-cancer
properties. We hypothesized that statin use is associated with lower risk of pancreatic cancer
in patients with chronic pancreatitis. This nationwide population-based cohort study included
all Danish patients diagnosed with incident chronic pancreatitis from 1 January 1996 to 31
December 2012. We used the Danish National Prescription Registry to ascertain information
on statin prescriptions for members of the study population before and after their pancreatitis
diagnosis. We computed crude incidence rates, incidence rate ratios (IRRs), and adjusted
hazard ratios (HRs) with associated 95% confidence intervals (CIs) for pancreatic cancer,
comparing statin users with non-users. We computed HRs using Cox proportional hazards
regression with statins treated as a time-varying exposure lagged by one year, adjusting for
age, sex, socioeconomic status, and individual comorbidities. The study included 8,311
chronic pancreatitis patients with a median age of 54 years. In total, 153 pancreatic cancers
were observed during 60,365 person-years of follow-up. The unadjusted IRR comparing
statin users with non-users was 1.00 (95% CI: 0.60-1.60). Adjustment for potential
confounders only had a small impact on the estimate (adjusted HR: 0.90; 95% CI: 0.56-1.44).
Our findings suggest that statin use is not associated with pancreatic cancer risk in patients
with chronic pancreatitis.
This article is protected by copyright. All rights reserved.
INTRODUCTION
Chronic pancreatitis is an inflammatory disease characterized by progressive and irreversible
destruction of the exocrine and endocrine pancreas.1 Some patients with chronic pancreatitis
eventually progress to pancreatic cancer,2 which has a poor prognosis.3 While this
progression has not been fully elucidated, inflammatory activation of pancreatic stellate cells
may play an important role.4 Accordingly, inhibition of inflammation within the pancreas may
help to protect against pancreatic cancer.
Statins (3-hydroxy-3-methyl-glutaryl-coenzyme A [HMG-CoA] reductase inhibitors)
are widely used cholesterol-lowering drugs. In addition to their anti-inflammatory properties,
statins may also inhibit carcinogensis.5 However, a previous meta-analysis found no
association between regular statin use and risk of pancreatic cancer.6 The anti-inflammatory
effect of statins is intriguing due to the close relation between inflammation and cancer, both
in the pancreas2, 7 and in general.5 Furthermore, experimental studies suggest that statins
can decrease pancreatic cancer risk by interfering with vital intracellular signaling pathways.8-
23
A recently published Danish case-cohort study of approximately 4,800 chronic
pancreatitis patients suggested that statin use was associated with an 80% reduction in
pancreatic cancer risk.24 Although a compelling finding, the study reported a crude null
finding and the beneficial effect of statins was only evident in a propensity-score matched
analysis of 339 statin users (7% of the total population). Given the rising incidence of chronic
pancreatitis25 and high mortality from pancreatic cancer, this potential association requires
further clarification.
This article is protected by copyright. All rights reserved.
We therefore conducted a population-based cohort study, using nationwide Danish
healthcare registries, to examine the association between statin use and pancreatic cancer
risk in patients with chronic pancreatitis.
MATERIALS AND METHODS
Setting and data sources
This nationwide population-based cohort study used data from the 1996-2015 period,
obtained from the Danish National Patient Registry, the National Prescription Registry, the
Civil Registration System, and the Danish Cancer Registry.26-29 These registries can be
linked on an individual level using the Civil Personal Registration (CPR) number, which is
assigned to every Danish resident at birth or upon immigration.
The Danish National Patient Registry was established in 1977. It contains information
on all inpatient hospitalizations at Danish public hospitals. Outpatient and emergency room
visits have been included since 1995. Patients are registered in the Danish National Patient
Registry with diagnoses coded according to the International Classification of Diseases (ICD)
8th revision (ICD-8) from 1977 through 1993 and ICD 10th revision (ICD-10) thereafter.
The Danish National Prescription Registry has collected detailed information on all
prescriptions filled at Danish community-based pharmacies since 1995.27 This registry
contains Anatomical Therapeutic Chemical (ATC) classification codes for drugs, purchase
dates, and dose information, among other variables.
The Danish Cancer Registry includes information on all cancers diagnosed in
Denmark since 1943, including date of diagnosis, cancer site, tumor characteristics, and
other variables.
This article is protected by copyright. All rights reserved.
The Civil Registration System, which was established in 1968, is an administrative
registry containing data on such variables as CPR-number, birth date, sex, sequential dates
of migration, vital status, and date of death for every resident in Denmark. The Civil
Registration System is updated daily and is virtually complete.
Study population
We identified all patients with an incident diagnosis of chronic pancreatitis recorded in the
Danish National Patient Registry during 1 January 1996 through 31 December 2012 (see
eFigure 1 and eTable 1). We excluded patients with a chronic pancreatitis diagnosis prior to
1 January 1996 to avoid including prevalent cases. We also excluded patients with a
diagnosis of hereditary or autoimmune pancreatitis as hereditary pancreatitis is associated
with a 70-fold relative risk of pancreatic cancer,30 and autoimmune pancreatitis is a systemic
autoimmune immunoglobin G4 disease and not confined to the pancreas.31, 32 Therefore,
including these diseases would limit homogeneity of our population and could bias our
estimates and complicate interpretation of our findings. We also excluded patients aged less
than 20 years at diagnosis, as such young patients often have autoimmune pancreatitis.
Other exclusions were birth or living outside Denmark prior to chronic pancreatitis diagnosis,
and a diagnosis of pancreatic cancer prior to chronic pancreatitis.
Follow-up and outcomes
We defined each patient’s index date as the date of first chronic pancreatitis diagnosis plus
one year. We required at least one full year of follow-up after the index date to reduce the
risk of misclassifying pancreatic cancer as chronic pancreatitis, as it is difficult to distinguish
these two conditions (i.e., a chronic pancreatitis diagnosis occurring less than a year prior to
This article is protected by copyright. All rights reserved.
a pancreatic cancer diagnosis would most likely be attributable to misdiagnosis of pancreatic
cancer).33 We followed each patient from the assigned index date to the first diagnosis of
pancreatic cancer recorded in the Danish Cancer Registry, death, emigration, or 31
December 2015. We considered pancreatic ductal adenocarcinomas and pancreatic cancers
with unknown histology (as the vast majority are expected to be adenocarcinomas34) as the
outcome of interest. Patients diagnosed with pancreatic cancer with a morphology other than
ductal adenocarcinoma were censored on the cancer diagnosis date (see eTable 2). The
follow-up schema is depicted graphically in eFigure 2. Information on tumor stage (recorded
in the Danish Cancer Registry as localized, regional, or metastatic) was available since study
start. Information on the American Joint Committee of Cancer stage, which is based on the
tumor-node-metastasis classification,35 was available since 2004.
Information on drug exposure
From the Danish National Prescription Registry, we identified all prescriptions filled by the
study participants since 1995 for the following medications: statins, other cholesterol-
lowering drugs, proton pump inhibitors (PPIs), non-steroidal anti-inflammatory drugs
(NSAIDs), and pancreatic enzyme supplements (see eTable 3).
Information on comorbidity
For each patient, we retrieved a medical history based on ICD-8 and ICD-10 codes from the
Danish National Patient Registry, restricting to diagnoses in the 10 years prior to chronic
pancreatitis diagnosis. We identified the following selected comorbid conditions: alcohol- and
smoking-related diseases (as proxies for excessive alcohol consumption and tobacco
smoking), diabetes, obesity, cerebrovascular disease, cardiovascular disease, liver disease,
This article is protected by copyright. All rights reserved.
renal disease, hyperlipidemia, previous malignancy, ulcer, and chronic hepatitis B or C (see
eTable 4). We also captured alcohol-related diseases, smoking-related diseases, and
diabetes not registered in the Danish National Patient Registry using ATC codes identified in
the National Prescription Registry. If a patient filled a prescription for a drug used to treat any
of these three conditions, he or she was considered to be living with the disease from the
date the first prescription was filled (see eTable 3).
Covariates
Age, sex, and socioeconomic status
We retrieved information on the patients’ age, sex, and socioeconomic status in the year of
chronic pancreatitis diagnosis from the Civil Registration System. We defined four levels of
socioeconomic status: employed, unemployed, retired, and other/unknown.
Drug exposure
We defined drug exposure as at least two filled prescriptions in the same drug class (e.g., a
patient with one prescription for simvastatin and one prescription for atorvastatin would be
classified as exposed to statins from the date of the second prescription). For each
prescription, we calculated the number of days’ supply as the daily dose multiplied by the
number of pills dispensed. For statins, we used the daily dose recommended by the Danish
Board of Health36 and for the other drugs of interest we used the defined daily dose specified
in the World Health Organization’s classification (for details, see eTable 5).
We considered drug exposure to be time-varying with a one-year lag period, allowing
patients to switch between exposed and unexposed status. Patients were considered
exposed from the prescription start date, plus one year, until one year following the last
This article is protected by copyright. All rights reserved.
prescription date (see eFigure 2) The one-year lag period served two purposes: 1) to allow a
reasonable induction period for a drug effect to occur, and 2) to preclude that an imminent
pancreatic cancer diagnosis might have influenced a physician’s prescribing behavior (i.e.,
reverse causation). For each drug class of interest, we considered the exposure to be
continuous if two prescriptions plus their days’ supply overlapped, allowing a 30-day grace
period for delays in prescription filling. To reduce the impact of outlier prescription durations,
we tabulated the days’ supply for each prescription and excluded prescriptions with a
duration below the 5th percentile. Prescriptions with a duration above the 95th percentile were
truncated at this level.
Comorbidity
In addition to the specific comorbidities, we constructed a composite measure of the overall
comorbidity burden for each patient using the Gagne Comorbidity Index, including
comorbidities diagnosed in the year preceding chronic pancreatitis diagnosis (see eTable
6).37 Based on this index, we defined three levels of comorbidity burden: low (score ≤0),
moderate (score 1-2), and severe (score >2). The Gagne Comorbidity Index combines the
Elixhauser and Charlson Indices and has proved superior to the other indices used alone.37
Statistical analyses
Main analysis
We tabulated descriptive characteristics of the study population according to statin use (ever
versus never use). We defined ever users as patients exposed to statins at any time during
follow-up. Because we allowed exposure to be time-varying, we also present descriptive
characteristics according to person-years accrued in each category. Continuous variables
This article is protected by copyright. All rights reserved.
are presented as medians with interquartile ranges (IQRs). For both statin users and non-
users, we computed the incidence rate of pancreatic cancer as the number of events divided
by the person-years accrued in each exposure category. We calculated the crude incidence
rate ratio (IRR) as the ratio between the incidence rate among statin users and that among
non-users. Using a multivariable Cox proportional hazards regression model, we estimated
the hazard ratio (HR) of pancreatic cancer comparing statin users with non-users. Based on
a directed acyclic graph (eTable 7), we adjusted our multivariable models for age as a
restricted cubic spline with three knots, sex, socioeconomic status, and Gagne Comorbidity
Index score. We included the Gagne Comorbidity Index score as a time-varying covariate,
updated each time the drug exposure changed to mitigate the impact of potential time-
varying confounding of comorbidities diagnosed after chronic pancreatitis. To examine
potential effect measure modification from age, sex, calendar period of chronic pancreatitis
diagnosis, and burden of comorbidity as of the index date, we stratified our models on these
variables. All estimates are presented with associated 95% confidence intervals (CIs).
Statistical analyses were performed using Stata 15 (StataCorp LP, College Station, Texas,
USA).
Sensitivity analyses
We conducted several sensitivity analyses to examine the robustness of our results. First, we
examined the potential effect of reverse causation on our estimates by alterations in the drug
exposure lag period from one year to one month and two years, respectively. Second, we
varied the grace period to 50% of the previous prescription duration to assess the robustness
of the assumed 30-day period used in our primary analysis. In a third analysis, we restricted
to patients with at least five years of prescription history available (patients diagnosed with
This article is protected by copyright. All rights reserved.
chronic pancreatitis on or after 1 January 2001) and excluded patients exposed to statins
during the five years preceding chronic pancreatitis diagnosis to mitigate potential selection
bias attributable to inclusion of prevalent statin users.38, 39
Ethical considerations
This study was approved by the Danish Data Protection Agency (J.no. 1-16-02-139-16),
Statistics Denmark, and the Danish Health Data Authority. According to Danish law, ethics
board approval is not required for registry-based studies.
RESULTS
Descriptive characteristics
We identified 8,311 patients with incident chronic pancreatitis in Denmark during the study
period (Table 1). Median age was 54 years (IQR: 45-64 years), and 5,498 (66.2%) were
men. On their index date, 654 patients (7.9%) were prevalent statin users, and 2,318 patients
(27.9%) filled a statin prescription during follow-up. Median duration of patients’ cumulative
statin prescriptions was 2.8 years (IQR: 1.3-5.3 years). Statin users had a higher Gagne
Comorbidity Index score compared with non-users. Socioeconomic status was similar among
statin users and non-users.
Pancreatic cancer characteristics
We observed 153 pancreatic cancers, 95 (62%) of whom had an assigned tumor stage. Of
these, 13 (13.7%) were early-stage cancer (localized or stage I-II), 28 (29.5%) had spread to
the lymph nodes (regional or stage IIB-III), and 54 (56.8%) were metastatic (stage IV).
This article is protected by copyright. All rights reserved.
Pancreatic cancer risk
The 153 pancreatic cancers were observed during 60,365 person-years of follow-up, yielding
a crude IRR of 1.00 (95% CI: 0.60-1.60). The estimate did not change substantially after
adjustment for potential confounders (adjusted HR: 0.90; 95% CI: 0.56-1.44); Table 2.
Additional adjustment for diabetes decreased the risk estimate but the precision did not
change (adjusted HR: 0.74; 95% CI: 0.46-1.19). We observed a tendency towards a
decreased pancreatic cancer risk with increasing age among statin users but no effect
measure modification by sex, calendar period of chronic pancreatitis diagnosis, or
comorbidity. However, because of the low number of events in each stratum, a meaningful
interpretation of these analyses is difficult. Alterations in the lag period or grace period or
restriction to non-prevalent users had no measurable impact on our estimates (Table 2).
DISCUSSION
In the current population-based cohort study, we found little evidence of an association
between statin use during the course of chronic pancreatitis and the risk of pancreatic
cancer. Our null finding remained robust in several sensitivity analyses.
Chronic pancreatitis patients constitute a group of individuals with a high pancreatic
cancer risk,2 and the burden of chronic pancreatitis is increasing.25 Therefore, measures to
prevent the progression from chronic pancreatitis to pancreatic cancer is important. However,
the association between statin use and pancreatic cancer risk in patients with chronic
pancreatitis has only been examined in one previous study,24 a Danish case-cohort study,
which had an overlapping study population. Although they reported a crude null finding in
This article is protected by copyright. All rights reserved.
4,807 patients with chronic pancreatitis, findings from their propensity-score matched
analysis suggested an 80% relative reduction in pancreatic cancer risk among statin users.
There are multiple differences between our studies. Bang et al. were able to find matches for
only 339 (7%) statin users, restricting inference to a small and likely selected population.
Despite this, substantial imbalances in baseline covariates were still present, increasing the
risk of residual confounding. The propensity score matching led to substantial differences in
the distribution of person-time between statin users and non-users. In the unmatched
analysis, 38% of the person-time was accrued by statin users (14% in our study). After
matching, this number increased to 58%. This imbalance inflates the person-time among
statin users relative to non-users, potentially generating a spuriously strong protective effect
of statins. Last, their estimate was based on only four cancers. By incorporating time-varying
methods and lagging prescription exposure, we reduced bias in our study.
Although our findings agree with a previous meta-analysis that found no association
between statin use and pancreatic cancer risk in healthy individuals,6 they contrast with the
substantial experimental evidence suggesting that statins inhibit pancreatic carcinogenesis.8-
23 An explanation for this discrepancy may be that poor lifestyle choices such as excessive
alcohol consumption, tobacco smoking, or a high level of comorbidity, which is frequently
seen in chronic pancreatitis patients40, 41 may mitigate any potentially beneficial effect of
statins. Furthermore, the potential anti-cancer effect of statins may differ according to statin
solubility (i.e., lipophilic or hydrophilic),42 which our study was large enough to examine. In
addition, statins may elicit their anti-cancer effect by inhibiting the Ras cellular pathway.43
The Ras pathway is activated by a K-ras mutation, which is prevalent in nearly all pancreatic
cancers and occurs at an early stage of malignant transformation.34 However, a substantial
This article is protected by copyright. All rights reserved.
proportion of chronic pancreatitis patients also harbor K-ras mutations,44, 45 precluding an
inhibitory effect of statins on such a mutation.
The present study has strengths. First, we conducted a nationwide, population-based
study with near-complete follow-up, limiting the possibility of selection bias. Our long look-
back period in the Danish National Patient Registry to 1977 decreased the likelihood of
including prevalent chronic pancreatitis patients, further mitigating potential bias. Second,
chronic pancreatitis diagnoses in the Danish National Patient Registry have been shown to
have high validity throughout the study period.46 Third, we had information on all statin
prescriptions filled in Denmark during the study period, as all prescriptions are captured in
the Danish National Prescription Registry and over-the-counter sale of statins is not allowed
in Denmark. The prevalence of statin use (8%) in our study approximates that of the general
population (11%), limiting concern of exposure misclassification.47 Fourth, our design using
time-varying exposures effectively decreased the risk of immortal time bias.48 To reduce a
potential effect of reverse causation, we lagged drug exposure by one year. Although the
choice of a one-year lag period was arbitrary and could have affected our analyses, findings
from our sensitivity analyses indicates that reverse causation is of limited concern.
Some limitations should be noted though. The number of pancreatic cancers in the
cohort was low, reducing precision of our estimates. However, it is reassuring that all
estimates were centered around the null. Our information on statin use was limited to
prescriptions filled rather than actual consumption of the prescribed medication. As
previously noted, compliance with prescription drugs among chronic pancreatitis patients
may be substantially impaired, possibly mitigating any potentially beneficial effect of statins
on pancreatic cancer risk. We also lacked information on prescription duration. Instead, we
used the recommended dose to calculate the number of days’ supply. If statin users in fact
This article is protected by copyright. All rights reserved.
used a different dose than recommended, this would have led to non-differential
misclassification (independent of the outcome), biasing our estimate towards the null.
Furthermore, we had limited information on tobacco smoking and alcohol consumption, both
of which are associated with pancreatic cancer.49, 50 Instead, we assessed exposure to these
substances using proxy diagnoses in the Danish National Patient Registry, which are known
to have limited ability to capture exposure to alcohol and tobacco.51 However, we observed a
high prevalence of these diagnoses in our study, limiting the concern of underreporting.
In conclusion, findings from the present nationwide population-based cohort study
provide evidence against an association of statin use with the risk of pancreatic cancer in
patients with chronic pancreatitis. However, it is not clear from our study if the pancreatic
cancer risk differs according to statin solubility. This should be further investigated in a larger
study population.
This article is protected by copyright. All rights reserved.
REFERENCES
1. Majumder S, Chari ST. Chronic pancreatitis. Lancet 2016;387:1957-1966.
2. Kirkegård J, Mortensen FV, Cronin-Fenton D. Chronic Pancreatitis and Pancreatic
Cancer Risk: A Systematic Review and Meta-analysis. Am J Gastroenterol 2017;112:1366-
1372.
3. Kamisawa T, Wood LD, Itoi T, Takaori K. Pancreatic cancer. Lancet 2016;388:73-
85.
4. Kleeff J, Korc M, Apte M, La Vecchia C, Johnson CD, Biankin AV, Neale RE,
Tempero M, Tuveson DA, Hruban RH, Neoptolemos JP. Pancreatic cancer. Nat Rev Dis
Primers 2016;2:16022.
5. Katz MS. Therapy insight: Potential of statins for cancer chemoprevention and
therapy. Nat Clin Pract Oncol 2005;2:82-89.
6. Cui X, Xie Y, Chen M, Li J, Liao X, Shen J, Shi M, Li W, Zheng H, Jiang B. Statin
use and risk of pancreatic cancer: a meta-analysis. Cancer Causes Control 2012;23:1099-
1111.
7. Momi N, Kaur S, Krishn SR, Batra SK. Discovering the route from inflammation to
pancreatic cancer. Minerva Gastroenterol Dietol 2012;58:283-297.
8. Liu Q, Li Y, Niu Z, Zong Y, Wang M, Yao L, Lu Z, Liao Q, Zhao Y. Atorvastatin
(Lipitor) attenuates the effects of aspirin on pancreatic cancerogenesis and the
chemotherapeutic efficacy of gemcitabine on pancreatic cancer by promoting M2 polarized
tumor associated macrophages. J Exp Clin Cancer Res 2016;35:33.
9. Ding N, Cui XX, Gao Z, Huang H, Wei X, Du Z, Lin Y, Shih WJ, Rabson AB,
Conney AH, Hu C, Zheng X. A triple combination of atorvastatin, celecoxib and tipifarnib
This article is protected by copyright. All rights reserved.
strongly inhibits pancreatic cancer cells and xenograft pancreatic tumors. Int J Oncol
2014;44:2139-2145.
10. Liao J, Chung YT, Yang AL, Zhang M, Li H, Zhang W, Yan L, Yang GY.
Atorvastatin inhibits pancreatic carcinogenesis and increases survival in LSL-KrasG12D-
LSL-Trp53R172H-Pdx1-Cre mice. Mol Carcinog 2013;52:739-750.
11. Gbelcova H, Sveda M, Laubertova L, Varga I, Vitek L, Kolar M, Strnad H, Zelenka
J, Bohmer D, Ruml T. The effect of simvastatin on lipid droplets accumulation in human
embryonic kidney cells and pancreatic cancer cells. Lipids Health Dis 2013;12:126.
12. Corcos L, Le Jossic-Corcos C. Statins: perspectives in cancer therapeutics. Dig
Liver Dis 2013;45:795-802.
13. Mohammed A, Qian L, Janakiram NB, Lightfoot S, Steele VE, Rao CV.
Atorvastatin delays progression of pancreatic lesions to carcinoma by regulating PI3/AKT
signaling in p48Cre/+ LSL-KrasG12D/+ mice. Int J Cancer 2012;131:1951-1962.
14. Mistafa O, Stenius U. Statins inhibit Akt/PKB signaling via P2X7 receptor in
pancreatic cancer cells. Biochem Pharmacol 2009;78:1115-1126.
15. Hussein D, Mo H. d-delta-Tocotrienol-mediated suppression of the proliferation of
human PANC-1, MIA PaCa-2, and BxPC-3 pancreatic carcinoma cells. Pancreas
2009;38:e124-136.
16. Gbelcova H, Lenicek M, Zelenka J, Knejzlik Z, Dvorakova G, Zadinova M,
Pouckova P, Kudla M, Balaz P, Ruml T, Vitek L. Differences in antitumor effects of various
statins on human pancreatic cancer. Int J Cancer 2008;122:1214-1221.
17. Issat T, Nowis D, Legat M, Makowski M, Klejman MP, Urbanski J, Skierski J,
Koronkiewicz M, Stoklosa T, Brzezinska A, Bil J, Gietka J, et al. Potentiated antitumor effects
This article is protected by copyright. All rights reserved.
of the combination treatment with statins and pamidronate in vitro and in vivo. Int J Oncol
2007;30:1413-1425.
18. Bocci G, Fioravanti A, Orlandi P, Bernardini N, Collecchi P, Del Tacca M, Danesi
R. Fluvastatin synergistically enhances the antiproliferative effect of gemcitabine in human
pancreatic cancer MIAPaCa-2 cells. Br J Cancer 2005;93:319-330.
19. Kusama T, Mukai M, Iwasaki T, Tatsuta M, Matsumoto Y, Akedo H, Inoue M,
Nakamura H. 3-hydroxy-3-methylglutaryl-coenzyme a reductase inhibitors reduce human
pancreatic cancer cell invasion and metastasis. Gastroenterology 2002;122:308-317.
20. Kusama T, Mukai M, Iwasaki T, Tatsuta M, Matsumoto Y, Akedo H, Nakamura H.
Inhibition of epidermal growth factor-induced RhoA translocation and invasion of human
pancreatic cancer cells by 3-hydroxy-3-methylglutaryl-coenzyme a reductase inhibitors.
Cancer Res 2001;61:4885-4891.
21. Muller C, Bockhorn AG, Klusmeier S, Kiehl M, Roeder C, Kalthoff H, Koch OM.
Lovastatin inhibits proliferation of pancreatic cancer cell lines with mutant as well as with
wild-type K-ras oncogene but has different effects on protein phosphorylation and induction
of apoptosis. Int J Oncol 1998;12:717-723.
22. Sumi S, Beauchamp RD, Townsend CM, Jr., Pour PM, Ishizuka J, Thompson JC.
Lovastatin inhibits pancreatic cancer growth regardless of RAS mutation. Pancreas
1994;9:657-661.
23. Sumi S, Beauchamp RD, Townsend CM, Jr., Uchida T, Murakami M, Rajaraman
S, Ishizuka J, Thompson JC. Inhibition of pancreatic adenocarcinoma cell growth by
lovastatin. Gastroenterology 1992;103:982-989.
This article is protected by copyright. All rights reserved.
24. Bang UC, Watanabe T, Bendtsen F. The relationship between the use of statins
and mortality, severity, and pancreatic cancer in Danish patients with chronic pancreatitis.
Eur J Gastroenterol Hepatol 2018;30:346-351.
25. Yadav D, Timmons L, Benson JT, Dierkhising RA, Chari ST. Incidence,
prevalence, and survival of chronic pancreatitis: a population-based study. Am J
Gastroenterol 2011;106:2192-2199.
26. Schmidt M, Schmidt SA, Sandegaard JL, Ehrenstein V, Pedersen L, Sorensen
HT. The Danish National Patient Registry: a review of content, data quality, and research
potential. Clin Epidemiol 2015;7:449-490.
27. Pottegard A, Schmidt SA, Wallach-Kildemoes H, Sorensen HT, Hallas J, Schmidt
M. Data Resource Profile: The Danish National Prescription Registry. Int J Epidemiol
2016;46:798-798f.
28. Gjerstorff ML. The Danish Cancer Registry. Scand J Public Health 2011;39:42-
45.
29. Schmidt M, Pedersen L, Sorensen HT. The Danish Civil Registration System as a
tool in epidemiology. Eur J Epidemiol 2014;29:541-549.
30. Raimondi S, Lowenfels AB, Morselli-Labate AM, Maisonneuve P, Pezzilli R.
Pancreatic cancer in chronic pancreatitis; aetiology, incidence, and early detection. Best
Pract Res Clin Gastroenterol 2010;24:349-358.
31. Kamisawa T, Egawa N, Nakajima H. Autoimmune pancreatitis is a systemic
autoimmune disease. Am J Gastroenterol 2003;98:2811-2812.
32. Stone JH, Zen Y, Deshpande V. IgG4-related disease. N Engl J Med
2012;366:539-551.
This article is protected by copyright. All rights reserved.
33. Munigala S, Kanwal F, Xian H, Agarwal B. New diagnosis of chronic pancreatitis:
risk of missing an underlying pancreatic cancer. Am J Gastroenterol 2014;109:1824-1830.
34. Ryan DP, Hong TS, Bardeesy N. Pancreatic adenocarcinoma. N Engl J Med
2014;371:1039-1049.
35. Edge SB, Byrd DR, Compton CG, Fritz AG, Greene FL, Trotti A. AJCC Cancer
Staging Manual, 7th ed. New York: Springer, 2010.
36. Lipid-lowering drugs [Lipidsænkende lægemidler]: Danish Board of Health, 2011.
37. Gagne JJ, Glynn RJ, Avorn J, Levin R, Schneeweiss S. A combined comorbidity
score predicted mortality in elderly patients better than existing scores. J Clin Epidemiol
2011;64:749-759.
38. Pottegard A, Friis S, Sturmer T, Hallas J, Bahmanyar S. Considerations for
Pharmacoepidemiological Studies of Drug-Cancer Associations. Basic Clin Pharmacol
Toxicol 2017.
39. Lund JL, Richardson DB, Sturmer T. The active comparator, new user study
design in pharmacoepidemiology: historical foundations and contemporary application. Curr
Epidemiol Rep 2015;2:221-228.
40. Yadav D, Whitcomb DC. The role of alcohol and smoking in pancreatitis. Nat Rev
Gastroenterol Hepatol 2010;7:131-145.
41. Bang UC, Benfield T, Hyldstrup L, Bendtsen F, Beck Jensen JE. Mortality,
cancer, and comorbidities associated with chronic pancreatitis: a Danish nationwide
matched-cohort study. Gastroenterology 2014;146:989-994.
42. Duncan RE, El-Sohemy A, Archer MC. Statins and cancer development. Cancer
Epidemiol Biomarkers Prev 2005;14:1897-1898.
This article is protected by copyright. All rights reserved.
43. Gong J, Sachdev E, Robbins LA, Lin E, Hendifar AE, Mita MM. Statins and
pancreatic cancer. Oncol Lett 2017;13:1035-1040.
44. Mulligan NJ, Yang S, Andry C, Klein M, O'Brien MJ. The role of p21ras in
pancreatic neoplasia and chronic pancreatitis. Hum Pathol 1999;30:602-610.
45. Logsdon CD, Ji B. Ras activity in acinar cells links chronic pancreatitis and
pancreatic cancer. Clin Gastroenterol Hepatol 2009;7:S40-43.
46. Kirkegård J, Mortensen MR, Johannsen IR, Mortensen FV, Cronin-Fenton D.
Positive predictive value of acute and chronic pancreatitis diagnoses in the Danish National
Patient Registry: A validation study. Scand J Public Health 2018.
47. Statistics of Prescription Drug Use in Denmark [Medstat]: Danish Board of Health,
2019.
48. Suissa S. Immortal time bias in observational studies of drug effects.
Pharmacoepidemiol Drug Saf 2007;16:241-249.
49. Tramacere I, Scotti L, Jenab M, Bagnardi V, Bellocco R, Rota M, Corrao G, Bravi
F, Boffetta P, La Vecchia C. Alcohol drinking and pancreatic cancer risk: a meta-analysis of
the dose-risk relation. Int J Cancer 2010;126:1474-1486.
50. Zou L, Zhong R, Shen N, Chen W, Zhu B, Ke J, Lu X, Zhang T, Lou J, Wang Z,
Liu L, Qi L, et al. Non-linear dose-response relationship between cigarette smoking and
pancreatic cancer risk: evidence from a meta-analysis of 42 observational studies. Eur J
Cancer 2014;50:193-203.
51. Sogaard M, Heide-Jorgensen U, Norgaard M, Johnsen SP, Thomsen RW.
Evidence for the low recording of weight status and lifestyle risk factors in the Danish
National Registry of Patients, 1999-2012. BMC Public Health 2015;15:1320.
This article is protected by copyright. All rights reserved.
Table 1. Baseline characteristics of 8,311 patients diagnosed w
ith chronic pancreatitis in Denm
ark during 1996-2012.
Patients
Person-years
Statin ever users
(%)
Statin never users
(%)
Statin users (%)
Non-users (%
)
Overall
2,318 (27.9%)
5,993 (72.1%)
8,278 (13.7%)
52,087 (86.3%)
Age
≤45 years
400 (17.3%)
1,865 (31.1%)
1,525 (18.5%)
19,411 (37.3%)
46-55 years
632 (27.3%)
1,734 (28.9%)
2,518 (30.4%)
15,531 (29.8%)
56-65 years
709 (30.6%)
1,194 (19.9%)
2,526 (30.6%)
9,729 (18.7%)
>65 years
577 (24.9%)
1,200 (20.0%)
1,689 (20.4%)
7,416 (14.2%)
Sex
M
en 1,489 (64.2%
) 4,009 (66.9%
) 5,345 (64.6%
) 34,385 (66.0%
)
W
omen
829 (35.8%)
1,984 (33.1%)
2,922 (35.4%)
17,702 (34.0%)
Socioeconomic status
E
mployed
681 (29.4%)
1,692 (28.2%)
2,777 (33.5%)
18,149 (34.8%)
U
nemployed
160 (6.9%)
462 (7.7%)
560 (6.8%)
4,685 (9.0%)
R
etired 1,447 (62.4%
) 3,580 (59.7%
) 4,807 (58.1%
) 26,997 (51.8%
)
This article is protected by copyright. All rights reserved.
O
ther/unknown
30 (1.3%)
259 (4.3%)
133 (1.6%)
2,257 (4.3%)
Period of diagnosis
1996-2003
977 (42.2%)
2,996 (50.0%)
4,381 (52.9%)
34,252 (65.8%)
2004-2012
1,341 (57.9%)
2,997 (50.0%)
3,896 (47.1%)
17,835 (34.2%)
Gagne C
omorbidity Index score
Low
(≤ 0) 1,275 (55.0%
) 3,119 (52.0%
) 5,036 (60.8%
) 24,708 (62.2%
)
M
oderate (1-2) 801 (34.6%
) 2,287 (38.2%
) 2,599 (31.4%
) 21,730 (32.6%
)
S
evere (>2) 242 (10.4%
) 587 (9.8%
) 643 (7.9%
) 5,649 (5.2%
)
Com
orbidity diagnoses
A
lcohol-related disease 799 (34.5%
) 2,744 (45.8%
) 1,015 (12.3%
) 22,129 (42.5%
)
S
moking-related disease
748 (32.3%)
2,406 (40.2%)
1,065 (12.9%)
17,567 (33.7%)
O
besity 129 (5.6%
) 148 (2.5%
) 284 (4.3%
) 1,470 (2.8%
)
D
iabetes mellitus
862 (37.2%)
1,088 (18.2%)
1,581 (19.1%)
10,337 (19.8%)
C
erebrovascular disease 298 (13.9%
) 398 (9.0%
) 655 (7.9%
) 2,909 (5.6%
)
C
ardiovascular disease 609 (26.3%
) 542 (18.3%
) 1,421 (17.2%
) 5,204 (10.0%
)
Liver disease
265 (11.4%)
1,097 (4.0%)
351 (4.2%)
7,647 (14.7%)
R
enal disease 154 (6.6%
) 238 (4.5%
) 268 (3.2%
) 1,667 (3.2%
)
H
yperlipidemia
400 (17.3%)
111 (1.9%)
1,214 (14.7%)
1,830 (3.5%)
P
revious malignancy
137 (5.9%)
448 (7. 5%)
271 (3.3%)
2,752 (5.5%)
This article is protected by copyright. All rights reserved.
U
lcer 315 (13.6%
) 882 (14.7%
) 477 (5.8%
) 7,699 (14.8%
)
C
hronic hepatitis B or C
<10*
99 (1.7%)
<10* 602 (1.2%
)
Medication use at baseline
S
tatins 654 (28.2%
) -
2,443 (29.5%)
781 (1.5%)
O
ther cholesterol-lowering drugs
31 (1.3%)
22 (0.4%)
172 (2.1%)
198 (0.4%)
N
on-steroidal anti-inflamm
atory drugs 286 (12.3%
) 599 (10.0%
) 1,061 (12.8%
) 4,851 (9.3%
)
P
roton pump inhibitors
666 (28.7%)
1,330 (22.2%)
2,169 (26.2%)
9,635 (18.5%)
P
ancreatic enzymes
104 (4.5%)
252 (4.2%)
346 (4.2%)
1,869 (3.6%)
* Num
bers collapsed for confidentiality purposes.
This article is protected by copyright. All rights reserved.
Table 2. Risk of pancreatic cancer in 8,311 patients diagnosed w
ith chronic pancreatitis in Denm
ark during 1996-2012 according to statin exposure.
Events
Person-years Incidence rate
(per 1,000 years) C
rude rate ratio A
djusted HR
6
Main analysis
1
N
on-users 132
52,087 2.53 (2.14-3.01)
1.00 (reference) 1.00 (reference)
S
tatin users 21
8,278 2.54 (1.65-3.89)
1.00 (0.60-1.60) 0.90 (0.56-1.44)
1-month lag period
2
N
on-users 129
51,224 2.52 (2.12-2.99)
1.00 (reference) 1.00 (reference)
S
tatin users 24
9,140 2.63 (1.76-3.92)
1.04 (0.64-1.62) 0.94 (0.60-1.47)
2-year lag period3
Non-users
130 53,001
2.45 (2.07-2.91) 1.00 (reference)
1.00 (reference)
S
tatin users 23
7,363 3.12 (2.08-4.70)
1.27 (0.78-1.99) 1.14 (0.72-1.79)
50% grace period
4
N
on-users 132
52,126 2.53 (2.14-3.00)
1.00 (reference) 1.00 (reference)
S
tatin users 21
8,238 2.55 (1.66-3.91)
1.01 (0.60-1.60) 0.90 (0.56-1.44)
Non-prevalent users
5
N
on-users 77
26,874 2.87 (2.29-3.58)
1.00 (reference) 1.00 (reference)
S
tatin users <10*
2,951 2.0* (0.91-4.53)
0.7* (0.25-1.61) 0.86 (0.36-2.02)
1 Statin prescriptions included as a tim
e-varying covariate, lagged by one year. 2 S
tatin prescriptions included as a time-varying covariate, lagged by one m
onth.
3 Statin prescriptions included as a tim
e-varying covariate, lagged by two years.
This article is protected by copyright. All rights reserved.
4 Statin prescriptions included as a tim
e-varying covariate, lagged by one year. Grace period as 50%
of the previous statin prescription instead of
30 days. 5 S
tatin prescriptions included as a time-varying covariate, lagged by one year. E
xclusion of patients with a statin prescription in the five years
prior to chronic pancreatitis diagnosis. Restricted to patients w
ith at least five years of prescription history available. 6 A
djusted for age (restricted cubic spline with three knots), sex, socioeconom
ic status, and Gagne C
omorbidity Index score.
* Num
bers collapsed for confidentiality purposes.
This article is protected by copyright. All rights reserved.