1458

Diabetes Mellitus and Pancreatic Cancer in a

Population-Based Case-Control Study in the
San Francisco Bay Area, California

Furong Wang,1 Samir Gupta,2 and Elizabeth A. Holly1,3

1
Department of Epidemiology and Biostatistics, School of Medicine, and 2Department of Medicine, Division of Gastroenterology,
University of California San Francisco, San Francisco, California; and 3Department of Health Research and Policy,
School of Medicine, Stanford University, Stanford, California

Abstract

Background: Diabetes has been postulated to be both group, diabetics in the case group had a shorter duration of
a risk factor and a consequence of pancreatic cancer, but diabetes (P = 0.0003) and a larger proportion of insulin users
the degree of risk and associated clinical factors remain (P = 0.002). Risk for pancreatic cancer varied with duration of
unclear. diabetes (OR, 2.4; 95% CI, 1.4-4.0 for 1-4 years; OR, 2.0; 95%

Downloaded from http://aacrjournals.org/cebp/article-pdf/15/8/1458/2265750/1458.pdf by guest on 14 November 2023

Methods: We conducted a population-based case-control CI, 1.2-3.4 for 5-9 years; and OR, 0.86; 95% CI, 0.52-1.4 for z10
study of pancreatic cancer in the San Francisco Bay Area years diabetes duration; P trend = 0.004). Among diabetics, use
between 1995 and 1999. Rapid case ascertainment through of oral diabetes medication or insulin for z5 years was not
the Surveillance, Epidemiology and End Results registry for associated with pancreatic cancer, but insulin use of <5 years
cases and random selection from the general population for was associated with a 6.8-fold risk for pancreatic cancer (95%
controls were employed to identify study participants with CI, 3.7-12).
no proxy interviews. Conclusion: Recent-onset diabetes may be a complication or
Results: Five hundred thirty-two cases and 1,701 controls an early marker of pancreatic cancer. Diabetes of short
were interviewed. Participants with pancreatic cancer were duration with insulin use conferred a substantially elevated
more likely to report a history of diabetes (13%) than were risk for pancreatic cancer and may reflect insulin resistance
controls [9%; odds ratio (OR), 1.5; 95% confidence interval that is elicited by pancreatic cancer. (Cancer Epidemiol
(95% CI), 1.1-2.1]. Compared with diabetics in the control Biomarkers Prev 2006;15(8):1458 – 63)

Introduction
Pancreatic cancer is diagnosed in nearly 34,000 individuals per approximately twice the risk of developing pancreatic cancer
year in the United States, and the overall 5-year survival rate is compared with those without diabetes after censoring of
<4% (1). Although 5-year survival of 15% has been shown in pancreatic cancer diagnosed in the first year of follow-up
highly selected clinic populations among patients who have (5, 13).
localized disease (1), pancreatic cancer is difficult to diagnose Data also suggest that pancreatic cancer may be a cause of
at an early, resectable stage, and candidates for surgery new-onset diabetes. Case-control studies show that patients
comprise <15% of all cases (2). Chemoradiotherapy offers with pancreatic cancer have an increased risk for new
limited benefit (3). Newer molecular therapies, including gene diagnoses of diabetes, especially within 3 years before their
therapy, antiangiogenic agents, immunotherapy, and inhib- cancer diagnosis (6, 23). Other cohort studies have reported
itors of cell signaling potentially, may be effective but still are an association between new-onset diabetes and cancer, but
under development (4). The relative improvement in duration conclusions about the strength of this association are limited
of survival associated with earlier stage of pancreatic cancer by the small number of incident cases of pancreatic cancer in
motivates the search for effective early detection tools. cohort studies (21-23).
Identification of individuals at high risk for pancreatic cancer Overall, prior data suggest that diabetes may be a marker
and prevention by identification of modifiable risk factors may for pancreatic cancer in some individuals and a risk factor for
allow for early interventions that will decrease pancreatic others. If diabetic individuals at highest risk for pancreatic
cancer morbidity and mortality. cancer could be identified through knowledge of epidemio-
Diabetes has been postulated to be both a risk factor for and logic and clinical characteristics, targeted surveillance of this
a consequence of pancreatic cancer (5-22). Meta-analyses of population might potentially identify early-stage, resectable
cohort and case-control studies with >10 years of follow-up pancreatic cancer and thereby improve pancreatic cancer-
have shown that patients with prevalent diabetes have associated survival.
Few large population-based studies have compared the
characteristics (including clinical features) of diabetics with
pancreatic cancer to those without cancer. We conducted the
Received 3/13/06; revised 5/17/06; accepted 5/31/06.
current analyses as part of a large, population-based case-
Grant support: National Cancer Institute, National Cancer Institute grants CA59706, CA89726, control study in the San Francisco Bay Area to examine the
CA108370, and CA109767 (E.A. Holly: Principal Investigator) and The Rombauer Pancreatic relationship between diabetes and pancreatic cancer with
Cancer Research Fund.
particular attention to duration of diabetes and treatment
The costs of publication of this article were defrayed in part by the payment of page charges.
This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. requirements.
Section 1734 solely to indicate this fact.
Requests for reprints: Elizabeth A. Holly, Division of Cancer Epidemiology, Department of
Epidemiology and Biostatistics, University of California San Francisco, Suite 280, 3333
California Street, San Francisco, CA 94118-1944. Phone: 415-476-3355; Fax: 415-563-4602.
Materials and Methods
E-mail: [email protected]
Copyright D 2006 American Association for Cancer Research. Study Population. Detailed methods for this case-control
doi:10.1158/1055-9965.EPI-06-0188 study have been published elsewhere (24-28). A brief summary

Cancer Epidemiol Biomarkers Prev 2006;15(8). August 2006

 Cancer Epidemiology, Biomarkers & Prevention 1459

of methods follows. Cases were individuals with newly regression was applied to compute ORs and 95% CIs. All
diagnosed pancreatic cancer identified by the Northern statistical tests were two sided with a V 0.05 considered as
California Cancer Center between 1995 and 1999 in the San significant and all analyses were done using SAS V8.0 (SAS
Francisco Bay Area using rapid case ascertainment. Diagnoses Institute, Inc., Cary, NC).
of pancreatic cancer were confirmed by participants’ physi- Potential confounders that were considered included
cians and by the Surveillance, Epidemiology and End Results race, adult body-mass index, physical activity, cigarette
abstracts that included histologic confirmation of cancer. All smoking, and alcohol consumption. Race was categorized as
pancreatic cancer patients in the six counties of the San white, black/African American, and Asian/Pacific Islander/
Francisco Bay Area, alive at first contact, 21 to 85 years of Other. Body-mass index was estimated from adult weight
age, and able to complete an interview in English were eligible and height [weight (kg)/height (m2)]. Body-mass index was
to be cases. Additional pancreatic cancer patients were categorized by quartiles among controls: V22.1, 22.2-24.2,
included who were seen in the oncology clinics at the 24.3-26.5, and z26.6; and by WHO categories: V25, 25.1-30,
University of California San Francisco and who met all and >30. Frequency of non-occupation-related physical activity
eligibility criteria except for residence. that lasted at least 30 minutes each episode was recorded as
Control participants were selected from the target pop- never or <1/mo, 1-2/mo, 3-4/mo, 2-3/wk, and daily or almost
ulation using random-digit dial and were frequency matched daily. Participants were defined as smokers if they had
to cases by sex and age within 5-year categories. Controls smoked z100 cigarettes, cigars at least once a month for
older than 65 years were supplemented by random selec- z6 months, or pipes at least once a month for z6 months in
tion from the Health-Care Finance Administration lists. their lifetime. Cigarette smokers were defined as never
Sixty-seven percent of eligible cases and 67% of eligible smoker, former smoker, and current smoker. Because one

Downloaded from http://aacrjournals.org/cebp/article-pdf/15/8/1458/2265750/1458.pdf by guest on 14 November 2023

controls completed an in-person interview and answered can, bottle, or 12-oz. glass of beer, one 4-oz. glass of wine, and
questions in standardized questionnaires. No proxy interviews one shot of liquor contain approximately the same amount
were done. Written consent was obtained from each study of alcohol, each was defined as one drink. Alcohol consump-
participant before the interview. This study was reviewed tion was computed as the average drinks weekly over the past
and approved by the University of California institutional two decades and the current decade, and was categorized as
review board. never drinker, V21, and >21 drinks per week. As none of the
factors mentioned above was a confounder for the association
Data on Diabetes Mellitus. The structured interview
between pancreatic cancer and diabetes mellitus, the final
contained extensive questions about diabetes, including
model was adjusted only for sex and age.
presence or absence of diabetes diagnosed by a physician that
lasted 1 year or longer before pancreatic cancer diagnosis, age
of diabetes diagnosis, use of insulin for glycemic control, first
age, last age, and total duration of insulin use if present, use Results
of oral hypoglycemic medication, first age, last age, and total
Five hundred thirty-two cases and 1,701 control participants
duration of oral hypoglycemic use. Duration of diabetes was
completed the interview and were included in the final
grouped as 1 to 4, 5 to 9, and z10 years. Total years of
analysis. Thirteen percent of cases (68 of 532) and 9% of
treatment was categorized as <5 and z5 years for both insulin
controls (150 of 1,701) reported a history of diabetes. Sex, body
and oral diabetic medication.
mass index, and family history of diabetes were similarly
Statistical Methods. Odds ratios (OR) and 95% confidence distributed among cases and controls with diabetes (Table 1).
intervals (95% CI) were used as an estimate of the relative risk Diabetics in the case group had a shorter duration of diabetes
(hereafter called risk) for pancreatic cancer associated with (mean, 7.6 versus 12.1 years; P = 0.0003) and older age at
diabetes and diabetes-related factors. Unconditional logistic diagnosis of diabetes (mean, 60 versus 57 years; P = 0.05) than

Table 1. Characteristics of diabetes mellitus patients, population-based case control study of pancreatic cancer, San
Francisco Bay Area, California
Characteristic Diabetics with pancreatic cancer (n = 68) Diabetics without pancreatic cancer (n = 150) P

Sex
Men 40 (59%) 91 (61%)
Women 28 (41%) 59 (39%) 0.80
Race
White 47 (69%) 128 (85%)
Black 14 (21%) 10 (7%)
Asian/Pacific Islander/other 7 (10%) 12 (8%) 0.007
Body-mass index, kg/m2 (range)
Mean F SD 27.7 F 4.5 27.3 F 4.6 0.52
Quartile 1 (V22.1) 3 (4%) 20 (13%)
Quartile 2 (22.2-24.2) 14 (21%) 21 (14%)
Quartile 3 (24.3-26.5) 16 (24%) 30 (20%)
Quartile 4 (z26.6) 35 (51%) 79 (53%) 0.17
Diabetes duration, y
Mean F SD 7.6 F 5.6 12.1 F 9.4 0.0003
Diabetes diagnosis age, y
Mean F SD 60.4 F 10.1 57.0 F 12.5 0.05
z50 y old at diagnosis 58 (85%) 110 (73%) 0.05
Insulin use
Yes 39 (57%) 53 (35%) 0.002
Years of use (mean F SD) 2.9 F 4.1 9.2 F 7.5 <0.0001
Years between diabetes diagnosis and 4.2 F 5.2 5.7 F 6.2 0.23
initial use of insulin (mean F SD)
Family history of diabetes 25 (38%) 53 (36%) 0.83

Cancer Epidemiol Biomarkers Prev 2006;15(8). August 2006

1460 Diabetes Mellitus and Pancreatic Cancer

Table 2. ORs and 95% CIs for pancreatic cancer and history of diabetes mellitus, population-based case control study, San
Francisco Bay Area, California
All Men Women

Case Control OR* (95% CI) Case Control OR* (95% CI) Case Control OR* (95% CI)

Diabetes
c
No 455 1,538 1.0 (reference) 245 785 1.0 (reference) 210 753 1.0 (reference)
Yes 68 150 1.5 (1.1-2.1) 40 91 1.4 (0.96-2.2) 28 59 1.7 (1.0-2.7)
Borderline 8 11 2.4 (0.97-6.1) 5 6 2.7 (0.82-9.0) 3 5 2.1 (0.50-8.9)
Diabetes diagnosis age (y)
<50 10 40 0.85 (0.42-1.7) 5 21 0.76 (0.28-2.0) 5 19 1.0 (0.37-2.8)
50-59 20 40 1.7 (0.97-2.9) 15 24 2.0 (1.0-3.9) 5 16 1.1 (0.4-3.0)
z60 38 70 1.8 (1.2-2.8) 20 46 1.5 (0.84-2.6) 18 24 2.5 (1.3-4.8)
Diabetes duration (y)
1-4 25 35 2.4 (1.4-4.0) 14 19 2.4 (1.2-4.9) 11 16 2.4 (1.1-5.3)
5-9 23 38 2.0 (1.2-3.4) 12 26 1.5 (0.75-3.0) 11 12 3.2 (1.4-7.5)
z10 20 77 0.86 (0.52-1.4) 14 46 1.0 (0.54-1.9) 6 31 0.68 (0.28-1.6)
Trend P P = 0.004 P = 0.06 P = 0.03

*Adjusted for age and sex.

cReference group for all comparisons.

Downloaded from http://aacrjournals.org/cebp/article-pdf/15/8/1458/2265750/1458.pdf by guest on 14 November 2023

diabetics in the control group. A greater proportion of diabetic or with oral medication for <5 years (95% CI, 3.7-12) but no
cases (57%) reported insulin use than diabetic controls (35%; elevation of risk among diabetics who took insulin alone or
P = 0.002) and diabetic cases had a shorter duration of insulin with oral medication for z5 years.
use (2.9 years) than diabetic controls (9.2 years; P < 0.0001). Separate analyses for diabetic patients stratified by whether
The interval between diabetes diagnosis and insulin use was or not they were treated with insulin showed that among
shorter for the diabetic case group (mean, 4.2 years) than the insulin-treated diabetics, the risk for pancreatic cancer was
diabetic control group (mean, 5.7 years). 10-fold for 1 to 4 years of diabetes, 4.8 for 5 to 9 years of diabetes,
Compared with participants without diabetes, the risk for and 1.2 for z10 years of diabetes (P trend = 0.0003; Table 5).
pancreatic cancer among participants with diabetes was 1.5 However, among diabetic patients not treated with insulin, no
(95% CI, 1.1-2.1; Table 2). Similar results were seen in men and notable trend was observed based on duration of diabetes.
women. The association between the two diseases was notable
after age 50 in men and after age 60 in women. The risk for
pancreatic cancer was 2.4 for diabetics with 1 to 4 years of Discussion
diabetes, 2.0 for 5 to 9 years, and 0.86 for z10 years of diabetes.
Increasing duration of diabetes was significantly associated Our population-based case-control study has identified that
with decreasing risk for pancreatic cancer (P trend = 0.004). Risk recent-onset diabetes, but not diabetes of z10-year duration,
for the overall association between pancreatic cancer and is associated with an increased risk for pancreatic cancer.
diabetes mellitus may differ by race as a nearly 3-fold Additionally, increased severity of diabetes, as reflected by
increased risk was observed among black/African Americans requirement for insulin therapy alone or insulin plus oral anti-
(95% CI, 1.1-7.3; Table 3). glycemic-agent therapy, conferred an increased risk for
Most participants with diabetes in this study reported oral pancreatic carcinoma, particularly among persons with diabe-
medication use alone or in combination with insulin for tes of recent onset.
treatment of diabetes. The risk for pancreatic cancer among The current estimation of the overall risk of pancreatic
diabetics who reported sole use of oral medication was not cancer among diabetics is consistent with the up to 2-fold
different from participants without diabetes (Table 4). The risk increase in risk reported by other investigators (5-22) and
for diabetics treated with insulin alone was elevated (OR, 1.9) supports previous work that has identified an inverse
but 95% CI overlapped unity. Diabetics who reported the use temporal relationship between diabetes and pancreatic can-
of a combination of oral medication and insulin had 2.5 times cer with cancer-associated risk decreasing with increasing
the risk for pancreatic cancer compared with controls (95% CI, duration of diabetes (5, 13, 16, 29-32) Because of rapid case
1.6-4.0). There was a nearly 7-fold elevation in risk for ascertainment, use of in-person interviews rather than proxy,
pancreatic cancer among diabetics who took insulin alone and population-based sampling, our observations are robust

Table 3. ORs and 95% CIs for pancreatic cancer and prior history of diabetes mellitus by race, population-based case
control study, San Francisco Bay Area, California
White Black/African American Asian/Pacific Islander/other

Case Control OR* (95% CI) Case Control OR* (95% CI) Case Control OR* (95% CI)

Diabetes
c
No 388 1,333 1.0 (reference) 31 65 1.0 (reference) 36 140 1.0 (reference)
Yes 47 128 1.2 (0.87-1.8) 14 10 2.9 (1.1-7.3) 7 12 2.1 (0.76-5.9)
Diabetes duration (y)
1-4 17 28 2.1 (1.1-3.8) 5 3 3.4 (0.75-15) 3 4 2.6 (0.55-13)
5-9 16 31 1.7 (0.93-3.2) 5 3 3.7 (0.81-17) 2 4 1.7 (0.29-9.9)
z10 14 69 0.68 (0.38-1.2) 4 4 2.0 (0.47-8.8) 2 4 2.0 (0.35-12)
Trend P P = 0.007 P = 0.62 P = 0.69

*Adjusted for age and sex.

cReference group for all comparisons.

Cancer Epidemiol Biomarkers Prev 2006;15(8). August 2006

 Cancer Epidemiology, Biomarkers & Prevention 1461

Table 4. ORs and 95% CIs for pancreatic cancer and with insulin, although this study included few cases or
treatment for diabetes mellitus, population-based case controls with diabetes (36). A cohort study of diabetics who
control study, San Francisco Bay Area, California were insulin dependent at baseline identified an increased risk
for pancreatic cancer in the cohort. However, because all study
Case Control OR* (95% CI)
participants were insulin dependent at study outset, the
c investigators could not evaluate insulin treatment as a
No diabetes 455 1,538 1.0 (reference)
Treatment for diabetes potential effect modifier (37). A nested case-control study
Exercise and/or diet alone 2 6 1.1 (0.23-5.6) evaluating the incidence of pancreatic cancer following
Medication alone 27 91 0.98 (0.63-1.5) diabetes diagnosis collected medication data on the 18 diabetic
Insulin alone 4 7 1.9 (0.55-6.5)
Medication and insulin combined 35 46 2.5 (1.6-4.0) case patients but did not report an analysis of these factors (23).
Total years of medication alone Lastly, a hospital-based case-control study estimated a >6-fold
<5 19 48 1.3 (0.77-2.3) increased risk for pancreatic cancer associated with insulin use
z5 8 43 0.62 (0.29-1.3) and a 2-fold increased risk associated with oral diabetic
b
Total years of insulin medication use, as well as evidence of increased risk for
Men + women pancreatic cancer associated with shorter duration of diabetes
<5 33 16 6.8 (3.7-12)
z5 6 37 0.54 (0.23-1.3) (6). Some of these studies contrast with our study that
Men identified a markedly increased risk for cancer among
No diabetes 245 785 1.0 (reference) diabetics with <5-year duration of diabetes.
<5 19 12 5.2 (2.5-11) Our in-person detailed interviews allowed us to identify
z5 5 20 0.81 (0.30-2.2) that the requirement for insulin use, particularly in those who

Downloaded from http://aacrjournals.org/cebp/article-pdf/15/8/1458/2265750/1458.pdf by guest on 14 November 2023

Women
No diabetes 210 753 1.0 (reference)
were newly diabetic, was associated with a substantially
<5 14 4 12 (3.9-37) increased risk for pancreatic cancer. We did not find that use
z5 1 17 0.21 (0.03-1.6) of oral medication alone was associated with increased risk
for pancreatic carcinoma among diabetics. Notably, one oral
*Adjusted for age and sex. diabetes medication, metformin, has been shown in animal
cReference group for all comparisons.
bCombined with or without oral medication.
studies to inhibit pancreatic cancer development (38). Because
metformin is a common diabetic medication, any potential
protective benefit against pancreatic cancer is intriguing but
and lend further support to the hypothesis that diabetes may will need confirmation in other studies.
often be a consequence rather than a cause of pancreatic can- We hypothesize that requirement for insulin treatment is an
cer, or an early marker of the disease (6, 10, 16, 20-23, 33, 34). indicator for more severe diabetes, and that, because of its
Our results extend the results of prior case-control and particular association with pancreatic cancer risk among new
cohort studies by assessing the association between medication diabetics, may reflect insulin resistance that is elicited by
use for diabetes and pancreatic cancer risk. Few studies have pancreatic cancer. Our data provide objective support for the
closely examined the potential interaction of the presence, clinical impression that diabetics with pancreatic cancer tend
type, and intensity of diabetes medication use with risk for to have diabetes of higher severity, and provide another
pancreatic carcinoma. An Italian hospital – based case-control clinical clue that may aid in the identification of diabetic
study with 720 case and 720 control participants documented individuals with pancreatic cancer (34, 39).
similar proportions of participants who used insulin or oral Insights into why early insulin use might be associated with
diabetic medication (16). For oral medications, it is possible pancreatic cancer in new diabetics may come through
that the use of hospital-based controls (who, if diabetic, might examination of prior work. Requirement of insulin treatment
be expected to have more severe diabetes) may have biased may be associated with pancreatic cancer because of factors
results towards the null. A population-based case-control elicited by pancreatic cancer that cause insulin resistance (34,
study with direct interviews in the United States, similar in 40-42) or pancreatic cancer – associated h-cell dysfunction in
scope and methods to our study, reported that treatment for insulin production (42-44). On a cellular level, defects in post-
diabetes, including oral medications or insulin, did not confer insulin receptor glycogen synthesis and storage in skeletal
increased risk for pancreatic cancer diagnosis (35). However, muscle, as well as impaired insulin action on phosphatidyli-
similar to our results, this study reported that shorter duration nositol 3-kinase activity and glucose transport, have been
of insulin use was associated with pancreatic cancer, with a identified as potential contributors to pancreatic cancer –
nearly 3-fold increased risk for <4 years of use, a 2-fold associated insulin resistance (45, 46). Islet amyloid polypeptide
increase risk for 5 to 9 years of use, and no association with and a protein identified through proteome analysis have been
longer use (35), although their risk estimates were not as high cited as candidate pancreatic cancer – associated diabetogenic
as ours. Another population-based case-control study found factors that may induce these functional cellular changes
no association between diabetes and pancreatic cancer, or (47, 48).
interaction between type of diabetes medication and risk of Older age at diagnosis with diabetes was more common
pancreatic cancer, except among men with diabetes treated among diabetic cases than diabetic controls, a finding

Table 5. ORs and 95% CIs for pancreatic cancer and duration of diabetes mellitus by insulin use, population-based case
control study, San Francisco Bay Area, California
Diabetes duration (y) Diabetic patients used insulin Diabetic patients did not use insulin

Case Control OR*(95% CI) Case Control OR* (95% CI)

c
No diabetes 455 1,538 1.0 (reference) 455 1,538 1.0 (reference)
1-4 12 4 10 (3.2-31) 13 31 1.4 (0.73-2.7)
5-9 13 9 4.8 (2.0-11) 10 29 1.1 (0.56-2.4)
z10 14 40 1.2 (0.62-2.1) 6 37 0.54 (0.23-1.3)
Trend P P = 0.0003 P = 0.11

*Adjusted for age and sex.

cReference group for all comparisons.

Cancer Epidemiol Biomarkers Prev 2006;15(8). August 2006

1462 Diabetes Mellitus and Pancreatic Cancer

consistent with other reports and one that may provide 3. Tsai JY, Iannitti DA, Safran H. Combined modality therapy for pancreatic
cancer. Semin Oncol 2003;30:71 – 9.
another clinically important indicator of those new diabetics 4. MacKenzie MJ. Molecular therapy in pancreatic adenocarcinoma. Lancet
at highest risk for cancer (23). Consistent with other inves- Oncol 2004;5:541 – 9.
tigators (23, 35), we did not find that body mass index affected 5. Huxley R, Ansary-Moghaddam A, Berrington de Gonzalez A, Barzi F,
the risk for pancreatic cancer among diabetics. In contrast to Woodward M. Type-II diabetes and pancreatic cancer: a meta-analysis of 36
studies. Br J Cancer 2005;92:2076 – 83.
some (16, 49), but in agreement with others (23, 35), we found 6. Bonelli L, Aste H, Bovo P, et al. Exocrine pancreatic cancer, cigarette
no correlation between presence of family history of diabetes smoking, and diabetes mellitus: a case-control study in northern Italy.
and risk for pancreatic cancer among diabetics. Pancreas 2003;27:143 – 9.
We observed the risk for pancreatic cancer among diabetic 7. Lin Y, Tamakoshi A, Kawamura T, et al. Risk of pancreatic cancer in relation
to alcohol drinking, coffee consumption and medical history: findings from
compared with nondiabetic African Americans to be higher the Japan collaborative cohort study for evaluation of cancer risk. Int J
(OR, 2.9) than the risk for pancreatic cancer among diabetic Cancer 2002;99:742 – 6.
compared with nondiabetic Whites (OR, 1.2) in a small number 8. Stolzenber-Solomon RZ, Pietinen P, Taylor PR, Virtamo J, Albanes D. A
of African Americans. The excess risk among African prospective study of medical conditions, anthropometry, physical activity,
and pancreatic cancer in male smokers (Finland). Cancer Causes Control
Americans persisted after adjustment for tobacco use and 2002;13:417 – 26.
body mass index. Although no statistically significant trend for 9. Gapstur SM, Gann PH, Lowe W, et al. Abnormal glucose metabolism and
decreasing pancreatic cancer risk with increasing duration of pancreatic cancer mortality. JAMA 2000;283:2552 – 8.
diabetes was observed among African Americans, the point 10. Frye JN, Inder WJ, Dobbs BR, Frizelle FA. Pancreatic cancer and diabetes: is
there a relationship? A case-controlled study. Aust N Z J Surg 2000;70:722 – 4.
estimates decreased linearly. We cannot rule out the possibility 11. Coughlin SS, Calle EE, Patel AV, Thun MJ. Predictors of pancreatic cancer
of residual confounding by other factors associated with mortality among a large cohort of United States adults. Cancer Causes
African American race and pancreatic cancer. African Amer-

Downloaded from http://aacrjournals.org/cebp/article-pdf/15/8/1458/2265750/1458.pdf by guest on 14 November 2023

Control 2000;11:915 – 23.
icans have a higher prevalence of diabetes than do Whites 12. Calle EE, Murphy TK, Rodriguez C, Thun MJ, Heath CW, Jr. Diabetes
mellitus and pancreatic cancer mortality in a prospective cohort of United
(13% versus 9%; ref. 50) and a higher annual incidence rate of States adults. Cancer Causes Control 1998;9:403 – 10.
pancreatic cancer (16.6/100,000 versus 10.7/100,000 for 1995- 13. Everhart J, Wright D. Diabetes mellitus as a risk factor for pancreatic cancer.
1999; ref. 51). It is plausible that similar factors may be A meta-analysis. JAMA 1995;273:1605 – 9.
associated with both diseases among African Americans and 14. Chow WH, Gridley G, Nyren O, et al. Risk of pancreatic cancer following
diabetes mellitus: a nationwide cohort study in Sweden. J Natl Cancer Inst
may account for the association that we observed. For 1995;87:930 – 1.
example, in one large population-based case-control study 15. Shibata A, Mack TM, Paganini-Hill A, Ross RK, Henderson BE. A pros-
(52), most of the excess risk for pancreatic cancer among pective study ofpancreatic cancer in the elderly. Int J Cancer 1994;58:46 – 9.
African Americans compared with Whites was explained by 16. Gullo L, Pezzilli R, Morselli-Labate AM. Diabetes and the risk of pancreatic
cancer. Italian Pancreatic Cancer Study Group. N Engl J Med 1994;331:81 – 4.
other established or suspected risk factors for pancreatic 17. Noy A, Bilezikian JP. Clinical review 63: Diabetes and pancreatic cancer:
cancer. Our ability to do detailed adjusted analysis was limited clues to the early diagnosis of pancreatic malignancy. J Clin Endocrinol
by the small number of African American cases. Pooling of our Metab 1994;79:1223 – 31.
data with other similar studies may allow for better delineation 18. Kalapothaki V, Tzonou A, Hsieh CC, et al. Tobacco, ethanol, coffee,
pancreatitis, diabetes mellitus, and cholelithiasis as risk factors for
of the degree and character of any association between pancreatic carcinoma. Cancer Causes Control 1993;4:375 – 82.
diabetes and pancreatic cancer among African Americans. 19. Smith GD, Egger M, Shipley MJ, Marmot MG. Post-challenge glucose
Our study may have several limitations. As with other concentration, impaired glucose tolerance, diabetes, and cancer mortality in
studies of pancreatic cancer, some pancreatic cancer patients men. Am J Epidemiol 1992;136:1110 – 4.
20. Cuzick J, Babiker AG. Pancreatic cancer, alcohol, diabetes mellitus and gall-
who may have been eligible for our study died before bladder disease. Int J Cancer 1989;43:415 – 21.
enrollment. It is possible that, if present, this incomplete case 21. Morris DV, Nabarro JD. Pancreatic cancer and diabetes mellitus. Diabet Med
ascertainment may have differentially affected nondiabetic 1984;1:119 – 21.
participants (who may be less likely to receive regular health 22. Ragozzino M, Melton LJ III, Chu CP, Palumbo PJ. Subsequent cancer risk in
the incidence cohort of Rochester, Minnesota, residents with diabetes
care) than diabetic participants, and possibly biased our risk mellitus. J Chronic Dis 1982;35:13 – 9.
estimates. Additionally, the prevalence of undiagnosed diabe- 23. Chari ST, Leibson CL, Rabe KG, et al. Probability of pancreatic cancer
tes may be as much as half of that of known diabetes in the following diabetes: a population-based study. Gastroenterology 2005;129:
504 – 11.
general population, suggesting that nondifferential misclassi- 24. Chan JM, Wang F, Holly EA. Vegetable and fruit intake and pancreatic
fication of cases and controls as nondiabetics when diabetes cancer in a population-based case-control study in the San Francisco bay
was indeed present may have occurred (53). Such misclassi- area. Cancer Epidemiol Biomarkers Prev 2005;14:2093 – 7.
fication would be expected to bias our results towards the null. 25. Eberle CA, Bracci PM, Holly EA. Anthropometric factors and pancreatic
cancer in a population-based case-control study in the San Francisco Bay
Further, as in all case-control studies, recall of past diagnoses area. Cancer Causes Control 2005;16:1235 – 44.
may be biased. However, diabetes is a serious disease and 26. Holly EA, Eberle CA, Bracci PM. Prior history of allergies and pancreatic
treatment with insulin is unlikely to be confused with oral cancer in the San Francisco Bay area. Am J Epidemiol 2003;158:432 – 41.
diabetic medication use by study participants. 27. Hoppin JA, Tolbert PE, Holly EA, et al. Pancreatic cancer and
serum organochlorine levels. Cancer Epidemiol Biomarkers Prev 2000;
9:199 – 205.
28. Slebos RJ, Hoppin JA, Tolbert PE, et al. K-ras and p53 in pancreatic cancer:
Conclusion association with medical history, histopathology, and environmental
exposures in a population-based study. Cancer Epidemiol Biomarkers Prev
We found that recent-onset diabetes, but not diabetes of z10- 2000;9:1223 – 32.
years duration, was associated with increased risk for 29. Wideroff L, Gridley G, Mellemkjaer L, et al. Cancer incidence in a
pancreatic cancer. Clinical factors that correlated with pancre- population-based cohort of patients hospitalized with diabetes mellitus in
Denmark. J Natl Cancer Inst 1997;89:1360 – 5.
atic cancer – associated diabetes included recent insulin use 30. La Vecchia C, Negri E, Franceschi S, D’Avanzo B, Boyle P. A case-control
and older age, but not body-mass index or family history. study of diabetes mellitus and cancer risk. Br J Cancer 1994;70:950 – 3.
Further efforts to identify diabetics at highest risk for 31. Jain M, Howe GR, St Louis P, Miller AB. Coffee and alcohol as determinants
pancreatic carcinoma may aid in early diagnosis and improved of risk of pancreas cancer: a case-control study from Toronto. Int J Cancer
1991;47:384 – 9.
pancreatic cancer – associated survival. 32. Adami HO, McLaughlin J, Ekbom A, et al. Cancer risk in patients with
diabetes mellitus. Cancer Causes Control 1991;2:307 – 14.
33. Permert J, Ihse I, Jorfeldt L, et al. Improved glucose metabolism after
subtotal pancreatectomy for pancreatic cancer. Br J Surg 1993;80:1047 – 50.
References 34. Permert J, Ihse I, Jorfeldt L, et al. Pancreatic cancer is associated with
1. Jemal A, Siegel R, Ward E, et al. Cancer statistics, 2006. CA Cancer J Clin impaired glucose metabolism. Eur J Surg 1993;159:101 – 7.
2006;56:106 – 30. 35. Silverman DT, Schiffman M, Everhart J, et al. Diabetes mellitus, other
2. Niederhuber JE, Brennan MF, Menck HR. The National Cancer Data Base medical conditions and familial history of cancer as risk factors for
report on pancreatic cancer. Cancer 1995;76:1671 – 7. pancreatic cancer. Br J Cancer 1999;80:1830 – 7.

Cancer Epidemiol Biomarkers Prev 2006;15(8). August 2006

 Cancer Epidemiology, Biomarkers & Prevention 1463

36. Bueno de Mesquita HB, Maisonneuve P, Moerman CJ, Walker AM. Aspects 45. Isaksson B, Strommer L, Friess H, et al. Impaired insulin action on
of medical history and exocrine carcinoma of the pancreas: a population- phosphatidylinositol 3kinase activity and glucose transport in skeletal
based case-control study in The Netherlands. Int J Cancer 1992;52:17 – 23. muscle of pancreatic cancer patients. Pancreas 2003;26:173 – 7.
37. Green A, Jensen OM. Frequency of cancer among insulin-treated diabetic 46. Liu J, Knezetic JA, Strommer L, et al. The intracellular mechanism of insulin
patients in Denmark. Diabetologia 1985;28:128 – 30. resistance in pancreatic cancer patients. J Clin Endocrinol Metab 2000;85:
38. Schneider MB, Matsuzaki H, Haorah J, et al. Prevention of pancreatic cancer 1232 – 8.
induction in hamsters by metformin. Gastroenterology 2001;120:1263 – 70. 47. Basso D, Greco E, Fogar P, et al. Pancreatic cancer-associated diabetes mellitus:
39. Karmody AJ, Kyle J. The association between carcinoma of the pancreas and an open field for proteomic applications. Clin Chim Acta 2005;357:184 – 9.
diabetes mellitus. Br J Surg 1969;56:362 – 4. 48. Permert J, Larsson J, Westermark GT, et al. Islet amyloid polypeptide in
40. Permert J, Adrian TE, Jacobsson P, et al. Is profound peripheral insulin patients with pancreatic cancer and diabetes. N Engl J Med 1994;330:313 – 8.
resistance in patients with pancreatic cancer caused by a tumor-associated 49. Egawa N, Tu Y, Sanaka M, Kamisawa T. Family history of diabetes and
factor? Am J Surg 1993;165:61 – 6; discussion 66 – 7. pancreatic cancer. Pancreas 2005;30:15 – 9.
41. Fogar P, Basso D, Panozzo MP, et al. C-peptide pattern in patients with 50. Centers for Disease Control and Prevention. National diabetes fact sheet:
pancreatic cancer. Anticancer Res 1993;13:2577 – 80. general information and national estimates on diabetes in the United States,
42. Cersosimo E, Pisters PW, Pesola G, et al. Insulin secretion and action in 2005. Atlanta (GA): U.S. Department of Health and Human Services, Centers
patients with pancreatic cancer. Cancer 1991;67:486 – 93. for Disease Control and Prevention; 2005.
43. Nakamori S, Ishikawa O, Ohigashi H, et al. Increased blood proinsulin and 51. Ries LAG, Eisner MP, Kosary CL, et al. SEER cancer statistics review,
decreased C-peptide levels in patients with pancreatic cancer. Hepatogas- 1973-1999. Bethesda: National Cancer Institute; 2002.
troenterology 1999;46:16 – 24. 52. Silverman DT, Hoover RN, Brown LM, et al. Why do Black Americans have
44. Syrigos KN, Konstantoulakis MM, Fyssas I, Katsilambros N, Golematis a higher risk of pancreatic cancer than White Americans? Epidemiology
BC. Autoantibodies against insulin and h-islet cells in pancreatic 2003;14:45 – 54.
adenocarcinoma: a possible explanation for diabetes mellitus. Int J Cancer 53. Harris MI. Diabetes in America: epidemiology and scope of the problem.
1996;66:624 – 6. Diabetes Care 1998;21 Suppl 3:C11 – 4.

Downloaded from http://aacrjournals.org/cebp/article-pdf/15/8/1458/2265750/1458.pdf by guest on 14 November 2023

Cancer Epidemiol Biomarkers Prev 2006;15(8). August 2006