Diabetes Volume 66, May 2017 1103

Diabetes, Pancreatogenic Diabetes, and Pancreatic

Cancer
Dana K. Andersen,1 Murray Korc,2 Gloria M. Petersen,3 Guido Eibl,4 Donghui Li,5 Michael R. Rickels,6
Suresh T. Chari,7 and James L. Abbruzzese8
Diabetes 2017;66:1103–1110 | DOI: 10.2337/db16-1477

Downloaded from http://diabetesjournals.org/diabetes/article-pdf/66/5/1103/538525/db161477.pdf

The relationships between diabetes and pancreatic control subjects (1). The increased incidence of pancreatic
ductal adenocarcinoma (PDAC) are complex. Long- cancer in populations with diabetes has been observed
standing type 2 diabetes (T2DM) is a risk factor for repeatedly in epidemiological studies, with a relative risk
pancreatic cancer, but increasing epidemiological or hazard ratio that ranges from 1.5 to 2.0 (2). Impor-
data point to PDAC as also a cause of diabetes due to tantly, among patients with PDAC who also have diabetes,
unknown mechanisms. New-onset diabetes is of partic- the diagnosis of diabetes occurred less than 24 months

PERSPECTIVES IN DIABETES
ular interest to the oncology community as the differen- prior to the diagnosis of PDAC in 74–88% of patients (3).
tiation of new-onset diabetes caused by PDAC as distinct These facts illustrate that diabetes and PDAC demon-
from T2DM may allow for earlier diagnosis of PDAC. To strate “dual causality,” in that both long-standing type
address these relationships and raise awareness of the
2 diabetes (T2DM) is a risk factor for the development
relationships between PDAC and diabetes, a symposium
of PDAC and, conversely, PDAC is a presumed cause of
entitled Diabetes, Pancreatogenic Diabetes, and Pancre-
diabetes in a large number of cases. The mechanisms of
atic Cancer was held at the American Diabetes Associ-
these causal relationships are unclear, as are the diagnos-
ation’s 76th Scientific Sessions in June 2016. This article
summarizes the data presented at that symposium, de- tic criteria for differentiating T2DM from diabetes that
scribing the current understanding of the interrelation- occurs as an early consequence of PDAC. These are im-
ships between diabetes, diabetes management, and portant considerations clinically given that PDAC has an
pancreatic cancer, and identifies areas where addi- overall 5-year survival rate of 7–8% (4) and that it is
tional research is needed. projected to become the second leading cause of cancer-
related deaths by 2020 (5). To explore these issues, a

by guest on 06 April 2024

symposium was held at the 76th Scientific Sessions of
The most common and most lethal form of pancreatic the American Diabetes Association in 2016. The following
cancer, referred to as pancreatic ductal adenocarcinoma represents a summary of the symposium presentations,
(PDAC), has an extraordinary association with diabetes. with the purpose of providing more insight into these
In a recent series of 100 patients diagnosed with cancers important relationships.
of the lung, breast, colon, prostate, or pancreas, 68% of
patients with PDAC had concurrent diabetes, whereas the DIABETES AS A RISK FACTOR FOR PDAC
prevalence of diabetes ranged from 15 to 21% in the other The Centers for Disease Control and Prevention estimates
age-matched cancer cohorts and 24% in 100 age-matched that approximately 29 million people in the U.S. had

1Division of Digestive Diseases and Nutrition, National Institute of Diabetes and 7Division of Gastroenterology and Hepatology, Department of Medicine, Mayo Clinic,
Digestive and Kidney Diseases, National Institutes of Health, Bethesda, MD Rochester, MN
2Division of Endocrinology, Department of Medicine, and Department of Bio- 8Division of Medical Oncology, Department of Medicine, Duke University School of

chemistry and Molecular Biology, Indiana University School of Medicine, and Medicine, Durham, NC
Indiana University Melvin and Bren Simon Cancer Center and Pancreatic Cancer Corresponding author: James L. Abbruzzese, [email protected].
Signature Center, Indianapolis, IN
3Department of Health Sciences Research, Mayo Clinic, Rochester, MN Received 30 November 2016 and accepted 25 January 2017.
4Department of Surgery, David Geffen School of Medicine at UCLA, Los Angeles, CA © 2017 by the American Diabetes Association. Readers may use this article as
5Department of Gastrointestinal Medical Oncology, MD Anderson Cancer Center, long as the work is properly cited, the use is educational and not for profit, and the
Houston, TX work is not altered. More information is available at http://www.diabetesjournals
6Division of Endocrinology, Diabetes and Metabolism, Department of Medicine, .org/content/license.
Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA
1104 Diabetes and Pancreatic Cancer Diabetes Volume 66, May 2017

T2DM in 2014, and about 8 million of these individuals

have not yet been diagnosed. Moreover, approximately
86 million adults in the U.S. are believed to have
prediabetes, defined by a fasting plasma glucose level
of 100–125 mg/dL, a 2-h plasma glucose level of
140–199 mg/dL, or a glycohemoglobin (HbA1c) level of
5.7–6.4% (6). These statistics put into perspective the
difficulties inherent in using glucose intolerance and
diabetes as biomarkers for diagnosing PDAC-associated
diabetes. The global spread of this huge health care bur-
den further underscores the need to better understand
the pathophysiology of T2DM and to distinguish it from
pancreatogenic or type 3c diabetes (T3cDM), which has
been classified as diabetes secondary to pancreatic exo-

Downloaded from http://diabetesjournals.org/diabetes/article-pdf/66/5/1103/538525/db161477.pdf by guest on 06 April 2024

crine disease.
Long-standing T2DM is a risk factor for PDAC (7).
In addition, T2DM is often associated with obesity, and Figure 1—Schematic representation of the islet–acinar–ductal axis.
obesity per se also increases the risk for developing An endocrine islet consisting mostly of insulin-producing b-cells
PDAC (8). T2DM is associated with resistance to the ac- (green) but also other endocrine cell types such as glucagon (yel-
low) is shown receiving an arterial blood supply (red). Some of this
tion of insulin to suppress hepatic glucose release. Con- arterial blood drains into an intrapancreatic (IP) portal circulation
comitantly, there is failure to enhance peripheral that bathes adjacent acinar and ductal cells. The insulin-rich blood
(predominantly skeletal muscle) glucose utilization, with supply exerts trophic effects on these cells that are most evident in
initially increased insulin levels, as the b-cell attempts to the acinar cells making up one such acinus. The acinar cells are
large and have many enzyme-rich zymogen granules (shown as
overcome the insulin resistance (9). With time, there is small yellow circles within each acinar cell).
progressive b-cell failure leading to T2DM including
aberrant activation of the unfolded protein response
pathway, which may induce both apoptosis and senes-
cence leading to a decreased b-cell mass (10). Neverthe- associated proinflammatory profile (12). Conceivably,
less, both obesity and T2DM are likely to lead to long time therefore, excessive RAGE activation may also contribute
periods in which such individuals have high intrapancre- to the higher incidence of PDAC in T2DM, as well as to
atic insulin levels due to an ongoing impetus by the b-cell the increased incidence of colorectal cancer, breast cancer,
to overcome the insulin-resistant state and maintain glu- and other cancers that may occur with T2DM and obesity.
cose homeostasis.
Insulin released by the b-cells is discharged into an GENOMIC ASSOCIATIONS OF DIABETES,
intrapancreatic portal circulation that provides blood to CHRONIC PANCREATITIS, AND PANCREATIC
acinar and ductal cells adjacent to the islets. Both cell CANCER
types can be adjacent to the islets and may be supplied Although there is strong clinical and epidemiological
by blood from the intrapancreatic portal circulation (Fig. 1). evidence that links the risk of PDAC to long-standing
This proximity enables high levels of islet hormones to T2DM or to chronic pancreatitis (CP) (the most frequent
directly reach groups of acinar and ductal cells, exerting cause of T3cDM), the genetic basis of susceptibility among
what has been termed proxicrine effects on insulin recep- these three diseases varies widely, with little overlap.
tors that are present on acinar cells and on IGF-I recep- Indeed, genetic heterogeneity of all three is the rule.
tors in any transformed cells that may arise in this region, Through extensive studies of case series and families
thereby promoting their survival and proliferation. Thus, and using a variety of study designs, catalogs of genes
hyperinsulinemia, particularly intrapancreatic, as a result have been identified for each condition. Interestingly,
of obesity and insulin resistance in prediabetes or early what is known about the mechanism by which these
T2DM may plausibly contribute to the observed increased genes influence susceptibility is uneven because of the
risk in PDAC. different methods used to identify the genes.
Additionally, poor glycemic control is associated with All three diseases share the following characteristics:
increased levels of advanced glycation end products (AGE) 1) all have subsets of patients who report family history or
that activate a receptor for AGE termed RAGE (11). This familial clustering, which are indicators of shared genetic
receptor belongs to the immunoglobulin super family and and/or environmental etiologies, 2) variation in age at di-
binds several ligands in addition to AGE, including certain agnosis has been linked to familial risk in some patients,
inflammatory cytokines and members of the S100 family and 3) Mendelian segregation analyses provide a formal
of proteins that have been implicated in inflammation demonstration that in some families there is evidence for
and cancer, including PDAC (11). Moreover, RAGE activa- a hereditary component (typically due to a major gene). In
tion contributes to the development of obesity and its addition, there are epidemiological risk factors (e.g., obesity
diabetes.diabetesjournals.org Andersen and Associates 1105

[diabetes], alcohol intake [CP], and smoking [PDAC]) that causes pancreatitis, and the majority of patients with re-
may interact with genetic factors to enhance risk. current acute and chronic pancreatitis have multiple vari-
Numerous approaches have been used to discover ants in a gene, or epistatic interactions between multiple
susceptibility genes, from family-based studies to case- genes, coupled with environmental stressors.
control and cohort studies that use a small list of candidate The study of genetic predisposition to PDAC has been
genes; large, agnostic genome-wide association studies particularly challenging because of the logistics of recruiting
(GWAS) that search for associated single nucleotide poly- and collecting biospecimens for analysis from patients with
morphisms (SNPs) that are typically noncoding; or poor survival. Like diabetes and CP, PDAC is genetically
next-generation sequencing. The needed analytic and bio- heterogeneous. Susceptibility genes that have been identified
informatics methods have kept pace and have provided range from rare mutations in genes associated with cancer
the statistical tools for interrogating the resulting data. syndromes to common SNPs. Designs that include family-
The genetic basis of T2DM is best characterized as poly- based and case-control studies have uncovered mutations
genic, with over 50 genes implicated (13). No single major in known syndrome-related genes, such as BRCA1, BRCA2,
gene explains the genetic risk of T2DM except in very lim- CDKN2A, and CFTR. Next-generation sequencing has

Downloaded from http://diabetesjournals.org/diabetes/article-pdf/66/5/1103/538525/db161477.pdf by guest on 06 April 2024

ited subsets (14). In contrast, there are well-documented identified additional mutations such as PALB2 (19) and
reports of CP and PDAC kindred that can be delineated as ATM (20).
due to mutations in major genes, while overall GWAS stud- More recently, patients with apparent sporadic PDAC
ies have identified many low-penetrance common SNPs that have been reported to carry germline mutations in major
are associated with increased risk. genes as well (21); this is likely to change our current
Among the three conditions, T2DM has been the most paradigm for risk assessment. GWAS studies of large
intensively studied, and considerably more resources have numbers of sporadic cases of pancreatic cancer and con-
been invested in uncovering the genetics of diabetes, trol subjects have implicated SNPs in regions that har-
consistent with the high prevalence and economic burden bor ABO, TERT, and CLPTM1L, among dozens of other
of this disease. However, genetic discoveries have not led genes. However, risk modeling that utilizes GWAS SNPs
to ready clinical application. It is clear from dozens of has not shown that the genetic information improves pre-
studies that diabetes is a multifactorial/polygenic disease diction in the general population (22).
and is genetically heterogeneous. Variants of over 50 genes
have been found to confer genetic risk, each with a modest THE ROLE OF OBESITY- AND PANCREATITIS-
effect (e.g., PPARG and KDNJ11, identified through the ASSOCIATED INFLAMMATION IN THE
candidate gene approach; and TCF7L2, WFS1, HDF1B, DEVELOPMENT OF PDAC
FTO, CDKN2A, and SLC20A8, among others, identified There is strong evidence that obesity is associated with an
through association and GWAS approaches). Researchers increased risk of cancer, including pancreatic cancer (23).
have developed new strategies in the search for diabetes In fact, the anticipated increase in pancreatic cancer in-
predisposition genes by further characterizing the genet- cidence and deaths may be at least partially attributed to
ic basis of diabetes through subclinical or related pheno- the obesity endemic. There are many possible mecha-
types (15). The outcome of research to date is that nisms by which obesity leads to (pancreatic) cancer, in-
susceptibility to diabetes fits a polygenic risk model, cluding insulin resistance with resulting hyperinsulinemia
with each genetic variant having a small effect. Interest- and inflammation (24). Nonsteroidal anti-inflammatory
ingly, despite these efforts, the genetic variants do not drugs can attenuate pancreatic cancer development in a
significantly improve risk assessment over common risk genetically engineered mouse model, indicating an impor-
factors such as age, sex, family history, BMI, and clinical tant role of tissue inflammation in this disease (25).
measures (16). Obesity-associated tissue inflammation is thought to create
Family-based strategies and knowledge of the path- a fertile microenvironment conducive to tumor initiation
ophysiology of CP facilitated early success in elucidat- and/or promotion. Recent evidence indicates that in ad-
ing its genetic heterogeneity and established PRSS1, dition to measures of general obesity, e.g., BMI, visceral
SPINK1, CTRC, CFTR, and CASR as susceptibility loci adiposity carries a strong association to metabolic diseases
for CP (17). Much of the variability in susceptibility to and gastrointestinal cancers, including pancreatic cancer
recurrent acute and chronic pancreatitis is related to (26).
genetic differences between patients. A large two-stage The mesenteric adipose tissue adjacent to the pancreas
GWAS analysis by Whitcomb et al. (18) identified and (peri-pancreatic depot) showed a substantially enhanced
replicated PRSS1-PRSS2 and X-linked CLDN2 as suscep- proinflammatory response to a high-fat, high-calorie diet
tibility loci, with the latter gene’s variants potentially compared with the peri-gonadal depot (27). A precise
interacting with alcohol consumption. Alcohol was long knowledge of adipose tissue depot-specific responses to
thought to be the primary causative agent, but genetic diet-induced obesity and their distinct effects on cancer
contributions have been of interest since the discovery development are critically important to explain the asso-
that rare PRSS1, CFTR, and SPINK1 variants were associ- ciation of certain body compositions to cancer risk and to
ated with pancreatitis risk. Thus, a single factor rarely understand possible sex differences. Mechanistically, a
1106 Diabetes and Pancreatic Cancer Diabetes Volume 66, May 2017

high-fat, high-calorie diet accelerates the progression of To date, at least 12 observational studies have investi-
pancreatic intraepithelial neoplasia, a known precursor of gated the association of metformin use and risk of PDAC
pancreatic cancer, and increases the incidence of invasive and with inconsistent data. A meta-analysis of these studies
metastatic pancreatic cancer in the conditional KrasG12D showed a summary risk ratio of 0.73 (95% CI 0.56–0.96,
mouse model (28,29). P = 0.023) among metformin users (33). Notably, two ran-
domized clinical trials comparing metformin with active
ANTIDIABETES MEDICATIONS AND THE RISK glucose-lowering therapy or placebo/usual care in T2DM
OF PDAC patients did not find any significant impact of metformin
Because various antidiabetes medications can directly use on cancer events. These trials were not designed with
affect the key factors mediating the association between cancer as an end point, and the follow-up times were some-
T2DM and PDAC, some of these medications may have an what short. Even if the protective effect of metformin is
impact on PDAC development, progression, and outcome. confirmed in future epidemiological studies, more work
High levels of insulin activate the IGF receptor, thereby needs to be done to identify eligible subpopulations that
acquiring the potential to exert mitogenic and tumor- would benefit from metformin for PDAC prevention.

Downloaded from http://diabetesjournals.org/diabetes/article-pdf/66/5/1103/538525/db161477.pdf by guest on 06 April 2024

promoting effects. Epidemiological studies have found Overall, the causal relationship between antidiabetes
that new use of insulin was associated with an increased therapies and the risk of PDAC cannot be established by
risk of PDAC (30). Sulfonylureas (insulin secretagogues) the current observational studies. Many of the reported
stimulate endogenous release of insulin by inhibiting the observational studies suffered from limitations of in-
K-ATP channel of the pancreatic b-cells. Use of sulfonyl- sufficient power and inadequate outcome validation,
ureas has been associated with an increased risk of pan- incomplete covariate ascertainment, and inadequate con-
creatic cancer but to a less extent than that for use of founding control. Better designed epidemiological inves-
insulin. Thiazolidinediones reduce insulin resistance by tigations and systematic capture of data on pancreatitis
activating peroxisome proliferator–activated receptor g, and PDAC from cardiovascular outcome trials and other
a nuclear receptor regulating glucose and lipid metabo- ongoing clinical trials should facilitate meta-analyses and
lism. Use of thiazolidinediones was not associated with accumulation of further knowledge.
risk of PDAC according to a recent review and meta-
analysis (31). T3CDM: PREVALENCE, DIFFERENTIATION FROM
Incretin-based therapies include glucagon-like pep- T2DM, AND IMPORTANCE OF RECOGNITION
tide 1 (GLP-1) receptor agonists and dipeptidyl pepti- The most common form of T3cDM is diabetes associated
dase 4 (DPP-4) inhibitors. GLP-1 stimulates insulin secretion with recurrent acute or chronic pancreatitis, although the
by the b-cell through activation of the GLP-1 receptor best recognized is cystic fibrosis–related diabetes. Diabe-
resulting in increased production of cAMP. It also increases tes is also associated with pancreatic cancer, and the most
insulin mRNA stability via upregulation of the transcription obvious form of T3cDM is that following partial or com-
factor PDX-1 and, in rodents, promotes b-cell growth and plete surgical resection of the pancreas. In recurrent acute
survival. DPP-4 inhibitors increase the endogenous levels of or chronic pancreatitis and in cystic fibrosis, the develop-
GLP-1 and the second principal incretin hormone glucose- ment of exocrine insufficiency typically predates the de-
dependent insulinotropic polypeptide (GIP). Therefore, velopment of the endocrine insufficiency and diabetes
GLP-1 mimetics and DPP-4 inhibitors may have an indirect (34). Any episode of acute pancreatitis can cause a tran-
trophic effect on pancreatic ductal cells. The association of sient diabetic state and increases the risk of the subse-
incretin-based therapy and acute pancreatitis has been raised quent development of diabetes (35).
in clinical studies, and as a result, the U.S. Food and Drug The pathophysiology of T3cDM most commonly in-
Administration has issued a warning regarding this relation- volves pancreatic glandular inflammation and subsequent
ship. Considering the long latency nature of human malig- irreversible fibrotic damage leading to islet cell loss.
nancies, further studies are required to clarify the association Unlike other types of diabetes, the islet loss involves
of incretin use and PDAC. not only the b-cells but also the pancreatic polypeptide
There is strong experimental evidence to support an (PP) cells early in the disease course and the a-cells late in
antitumor activity of metformin. The fundamental mech- the disease course (Table 1). Because of concomitant pan-
anism of metformin action involves two possible avenues: creatic exocrine insufficiency, there is also maldigestion of
decreased circulating insulin and IGF levels and inhibition nutrients with consequent impairment in incretin secre-
of mitochondrial oxidative phosphorylation, which conse- tion, and pancreatic enzyme replacement can improve
quently leads to energy stress, AMPK activation, and incretin and insulin secretion as well as glucose tolerance.
inhibition of mTOR signaling. A study in 2014 identified Typically, islet b-cell secretory capacity is preserved until
mitochondrial glycerophosphate dehydrogenase as the ma- the majority of pancreatic exocrine function is lost (36);
jor molecular target of metformin (32). Inhibition of this however, given how common overweight and obesity are
enzyme resulted in an altered hepatocellular redox state, now in the population, traditional T2DM risk factors and
reduced conversion of lactate and glycerol to glucose, and insulin resistance may accelerate the presentation of di-
decreased hepatic gluconeogenesis. abetes in the context of pancreatic disease.
diabetes.diabetesjournals.org Andersen and Associates 1107

Table 1—Clinical and laboratory findings in types of diabetes

Parameter T1DM T2DM T3cDM
Ketoacidosis Common Rare Rare
Hypoglycemia Common Rare Common
Peripheral insulin sensitivity Normal or decreased Decreased Normal or increased
Hepatic insulin sensitivity Normal or decreased Decreased Normal or decreased
Insulin levels Low or absent High or “normal” “Normal” or low
Glucagon levels Normal or high Normal or high “Normal” or low
PP levels Normal or low (late) Normal or high Low or absent
GIP levels Normal or low Variable Low
GLP-1 levels Normal Variable Variable
Typical age of onset Childhood or adolescence Adulthood Any

Downloaded from http://diabetesjournals.org/diabetes/article-pdf/66/5/1103/538525/db161477.pdf by guest on 06 April 2024

Typical etiology Autoimmune Obesity, age CP, cystic fibrosis, postoperative
“Normal,” inappropriate in the context of elevated glucose. Adapted from Cui and Andersen (47).

Importantly, pancreatic disease remains underrecogni- Recognition of T3cDM is important for identifica-
zed as an underlying etiology of diabetes. At an academic tion of underlying pancreatic disease that may require
referral center in Germany, almost 10% of all patients with specific intervention. Oral pancreatic enzyme replace-
diabetes could be classified as T3cDM, with CP being the ment to correct maldigestion, particularly of fat, is necessary
most common etiology affecting 80% of cases (37) (Fig. 2). to optimize incretin secretion and enable the absorption of
In most of these cases, T3cDM was initially misclassified as vitamin D and other fat-soluble vitamins. Insulin therapy is
T2DM, leading the authors to propose diagnostic criteria ultimately needed for most patients, and metformin should
for T3cDM (38). As outlined in Table 2, the diagnosis of be considered when concomitant insulin resistance is pre-
T3cDM requires 1) the presence of pancreatic exocrine in- sent. Since metformin may reduce the risk of pancreatic
sufficiency, 2) evidence of pathological pancreatic imaging, cancer in patients with T2DM, its use is strongly encouraged
and 3) the absence of type 1 diabetes (T1DM)-associated when evidence of insulin resistance is present (34).
autoimmune markers; the diagnosis may be further sup- Inaccurate classification of diabetes in clinical prac-
ported by evidence of PP, incretin, or insulin secretory tice has limited our understanding of the prevalence of
defects in the absence of clinical or biochemical evidence T3cDM, and even our current diagnostic criteria still re-
of overt insulin resistance. Even with this guidance, discrim- quire prospective validation before we can appreciate their
inating T3cDM secondary to pancreatitis from T2DM can limitations as well as the health consequences of mis-
still be challenging. When ambiguity remains, confirmation diagnosis. Most importantly, randomized clinical trial
of T3cDM can be made by documentation of an absent PP evidence is needed for specified interventions to prevent
response to mixed-nutrient ingestion, which best discrimi- diabetes from developing in patients with pancreatic
nates the pathological islet response from that of T2DM (39). disease and how to best treat T3cDM in those affected.

Figure 2—Prevalence of types of diabetes. Distribution of types of diabetes (A) and causes of T3cDM (pancreatogenic) diabetes (B) based
on studies of 1,922 patients with diabetes reported by Hardt et al. (49). Reproduced from Cui and Andersen (47).
1108 Diabetes and Pancreatic Cancer Diabetes Volume 66, May 2017

Table 2—Diagnostic criteria for T3cDM

Major criteria (all must be fulfilled)
Presence of exocrine pancreatic insufficiency (according to monoclonal fecal elastase 1 or direct function tests).
Pathological pancreatic imaging (by endoscopic ultrasound, MRI, or computed tomography).
Absence of T1DM-associated autoimmune markers.
Minor criteria
Impaired b-cell function (e.g., as measured by HOMA-B, C-peptide/glucose ratio).
No excessive insulin resistance (e.g., as measured by HOMA of insulin resistance).
Impaired incretin (e.g., GIP) or PP secretion.
Low serum levels of lipid soluble vitamins (A, D, E, or K).
Adapted from Ewald and Bretzel (38). HOMA-B, HOMA of b-cell function.

MECHANISMS OF DIABETES IN PANCREATIC lung, and prostate carcinoma and in age-matched control

Downloaded from http://diabetesjournals.org/diabetes/article-pdf/66/5/1103/538525/db161477.pdf by guest on 06 April 2024

CANCER subjects (1). Further, although insulin resistance is a
The prevalence of diabetes among patients with PDAC common finding in patients with both PDAC and diabe-
is extraordinarily high. In a study of 512 PDAC patients tes, most patients with PDAC report weight loss rather
and 933 control subjects, diabetes was found in 47% of than weight gain. The clinical features of deteriorating
PDAC patients compared with only 7% of control subjects, glycemic control in conjunction with weight loss that
and although a normal fasting glucose was present in 59% accompanies PDAC prior to its diagnosis are atypical
of control subjects, it was found in only 14% of PDAC pa- for T2DM; this combination of features should alert
tients (3) (Fig. 3). In 74% of the PDAC patients with dia- the clinician to the possibility of PDAC-associated
betes, the diagnosis of diabetes was made within 24 diabetes.
months before the diagnosis of PDAC, frequently at a Perhaps most compelling are the data showing that
time when the tumor was radiographically occult (40). new-onset diabetes associated with PDAC may resolve
This suggests that in many patients, new-onset diabetes following tumor resection, as long as there are sufficient
is caused by the tumor and may be a useful “biomarker” islets left in the residual pancreatic tissue. Pancreatic
for the diagnosis of PDAC. The following observations resection is a recognized cause of T3cDM, due to loss of
support this concept. islet mass, and is seen in about half of all subjects after
Despite the fact that canonical risk factors for diabetes proximal pancreatectomy. Earlier reports had observed an
such as age, obesity, and a family history of diabetes are actual improvement in diabetes after resection of some
also risk factors for PDAC, the incidence of diabetes in pancreatic cancers (41), but it was unclear whether this
PDAC is much higher than the incidence of diabetes among was a consequence of nutritional or other factors. In a
other common malignancies. In a study of the prevalence study of 104 patients who underwent PDAC resection, of
of diabetes among patients with the most common solid whom 41 had diabetes at the time of surgery, it was found
tumors, diabetes was found in 68% of patients with PDAC that 57% of the patients with new-onset diabetes had
compared with 14.8–23.5% of patients with breast, colon, resolution of their diabetes postoperatively, whereas all
of the patients with long-standing diabetes had a persis-
tence of diabetes after pancreatic resection (3) (Fig. 4).
These observations strongly suggest that new-onset dia-
betes associated with PDAC may be a paraneoplastic phe-
nomenon, where one or more factors induced by the
malignancy interfere with insulin secretion or insulin
action, leading to diabetes.
Numerous studies have attempted to identify the
mechanism(s) or the genomic and/or protein markers
of the diabetes caused by PDAC. Connexin 26, a gap
junction protein, has been described as being highly
overexpressed in islets of PDAC patients with diabetes
(42), and a PDAC-derived S-100A8 N-terminal peptide
has been described as a diabetogenic agent by Basso
et al. (43). Huang et al. (44) described two upregulated
genes in 27 patients with PDAC associated with new-
onset diabetes, vanin-1 and matrix metalloproteinase 9,
Figure 3—Prevalence of diabetes (DM) in PDAC. The prevalence of
diabetes and impaired fasting glucose in 512 pancreatic cancer
and a variety of microRNA fragments have been sug-
patients and 933 control subjects. Reproduced from Pannala gested as possibly having predictive usefulness (45).
et al. (3). Chari and colleagues (46) at the Mayo Clinic have suggested
diabetes.diabetesjournals.org Andersen and Associates 1109

yet unknown mechanism(s), and the differentiation of

new-onset diabetes caused by PDAC, which is distinct
from long-standing T2DM or T3cDM associated with
CP, may allow for the diagnosis and intervention in
early-stage disease at a point at which curative therapy
is possible. Much information needs to be acquired,
however, to understand the mechanisms of both the
contributions of long-standing T2DM and obesity on
the development of PDAC and the cause of PDAC-
associated diabetes. The ability to distinguish PDAC-
associated diabetes from T2DM is critical for recognition
that new-onset diabetes in patients aged 50 years or older
may be a harbinger of PDAC leading to earlier PDAC de-
tection and improved therapeutic outcome in this deadly

Downloaded from http://diabetesjournals.org/diabetes/article-pdf/66/5/1103/538525/db161477.pdf by guest on 06 April 2024

Figure 4—Prevalence of diabetes (DM) after pancreaticoduodenec- malignancy.
tomy for PDAC. The prevalence of diabetes after resection of PDAC
in new-onset diabetes (<2 years duration), long-standing diabetes
(>2 years duration), impaired fasting glucose (IFG) (100–125 mg/dL)
preoperatively, and normal fasting glucose (NFG) (#99 mg/dL) pre- Funding. This work was supported in part by National Institutes of Health
operatively. Reproduced from Pannala et al. (3). grants RO1-CA-075059 (to M.K.), P01-CA-163200 (to G.E.), R01-DK-97830
(to M.R.R.), U01-DK-108288 (to S.T.C.), and P50-CA-014236 (to J.L.A.). The
authors acknowledge travel support from the Pancreatic Cancer Action
Network.
Duality of Interest. No potential conflicts of interest relevant to this article
that adrenomedullin, carried within PDAC-derived exo- were reported.
somes, acts as a peptide mediator of the impaired insulin
References
sensitivity and secretion seen in PDAC-associated diabe-
1. Aggarwal G, Kamada P, Chari ST. Prevalence of diabetes mellitus in pan-
tes. Adrenomedullin is a multifunctional vasoactive pep-
creatic cancer compared to common cancers. Pancreas 2013;42:198–201
tide that has been implicated in inflammation and sepsis
2. Everhart J, Wright D. Diabetes mellitus as a risk factor for pancreatic
and is highly overexpressed in PDAC (46). Together with
cancer. A meta-analysis. JAMA 1995;273:1605–1609
vanin-1, which is also expressed in response to inflam- 3. Pannala R, Leirness JB, Bamlet WR, Basu A, Petersen GM, Chari ST.
mation, these findings suggest that mediators of inflam- Prevalence and clinical profile of pancreatic cancer-associated diabetes mel-
mation may play a role in altered islet function and litus. Gastroenterology 2008;134:981–987
insulin action in PDAC. 4. Siegel RL, Miller KD, Jemal A. Cancer statistics, 2016. CA Cancer J Clin
A deficiency of PP release has been documented in 2016;66:7–30
T3cDM associated with pancreatic resection, CP, and 5. Rahib L, Smith BD, Aizenberg R, Rosenzweig AB, Fleshman JM, Matrisian
cystic fibrosis (47), and a recent pilot study suggested LM. Projecting cancer incidence and deaths to 2030: the unexpected burden of
that impaired PP release may be a marker of PDAC- thyroid, liver, and pancreas cancers in the United States. Cancer Res 2014;74:
associated diabetes (48). PP release is increased in T2DM, 2913–2921
6. Menke A, Casagrande S, Geiss L, Cowie CC. Prevalence of and trends in
and a deficient PP response to nutrient ingestion may
diabetes among adults in the United States, 1988-2012. JAMA 2015;314:
differentiate new-onset diabetes that is caused by pancre-
1021–1029
atic exocrine disease from T2DM. In the pilot study by 7. Wolpin BM, Bao Y, Qian ZR, et al. Hyperglycemia, insulin resistance, im-
Hart et al. (48), basal and meal-stimulated PP release was paired pancreatic b-cell function, and risk of pancreatic cancer. J Natl Cancer
significantly decreased in patients with diabetes associ- Inst 2013;105:1027–1035
ated with PDAC localized to the head of the pancreas 8. Tang H, Dong X, Hassan M, Abbruzzese JL, Li D. Body mass index and
compared with patients with T2DM. Further studies obesity- and diabetes-associated genotypes and risk for pancreatic cancer.
sponsored by the National Institutes of Health–sponsored Cancer Epidemiol Biomarkers Prev 2011;20:779–792
Chronic Pancreatitis, Diabetes, Pancreatic Cancer consor- 9. Kahn SE. The relative contributions of insulin resistance and beta-cell
tium are currently under way to determine whether a dysfunction to the pathophysiology of type 2 diabetes. Diabetologia 2003;46:3–19
10. Dooley J, Tian L, Schonefeldt S, et al. Genetic predisposition for beta cell
deficient PP response to nutrients can identify a high-
fragility underlies type 1 and type 2 diabetes. Nat Genet 2016;48:519–527
risk group with new-onset diabetes that includes patients
11. Leclerc E, Vetter SW. The role of S100 proteins and their receptor RAGE in
with early-stage PDAC. pancreatic cancer. Biochim Biophys Acta 2015;1852:2706–2711
CONCLUSIONS 12. Song F, Hurtado del Pozo C, Rosario R, et al. RAGE regulates the metabolic
and inflammatory response to high-fat feeding in mice. Diabetes 2014;63:1948–
Long-standing T2DM and CP (that may be associated with 1965
T3cDM) are risk factors for the development of PDAC. 13. Herder C, Roden M. Genetics of type 2 diabetes: pathophysiologic and
Obesity accompanying T2DM may also increase the risk of clinical relevance. Eur J Clin Invest 2011;41:679–692
PDAC due to hyperinsulinemia, the actions of adipokines, 14. Fuchsberger C, Flannick J, Teslovich TM, et al. The genetic architecture of
or other factors. PDAC is also a cause of diabetes due to as type 2 diabetes. Nature 2016;536:41–47
1110 Diabetes and Pancreatic Cancer Diabetes Volume 66, May 2017

15. Zeggini E, Scott LJ, Saxena R, et al.; Wellcome Trust Case Control Con- 33. Wang Z, Lai ST, Xie L, et al. Metformin is associated with reduced risk of
sortium. Meta-analysis of genome-wide association data and large-scale repli- pancreatic cancer in patients with type 2 diabetes mellitus: a systematic review
cation identifies additional susceptibility loci for type 2 diabetes. Nat Genet 2008; and meta-analysis. Diabetes Res Clin Pract 2014;106:19–26
40:638–645 34. Rickels MR, Bellin M, Toledo FGS, et al.; PancreasFest Recommendation
16. Meigs JB, Shrader P, Sullivan LM, et al. Genotype score in addition to common Conference Participants. Detection, evaluation and treatment of diabetes mellitus
risk factors for prediction of type 2 diabetes. N Engl J Med 2008;359:2208–2219 in chronic pancreatitis: recommendations from PancreasFest 2012. Pancreatol-
17. Whitcomb DC. Genetic aspects of pancreatitis. Annu Rev Med 2010;61: ogy 2013;13:336–342
413–424 35. Das SL, Kennedy JI, Murphy R, Phillips AR, Windsor JA, Petrov MS. Re-
18. Whitcomb DC, LaRusch J, Krasinskas AM, et al.; Alzheimer’s Disease Genetics lationship between the exocrine and endocrine pancreas after acute pancreatitis.
Consortium. Common genetic variants in the CLDN2 and PRSS1-PRSS2 loci alter World J Gastroenterol 2014;20:17196–17205
risk for alcohol-related and sporadic pancreatitis. Nat Genet 2012;44:1349–1354 36. Domschke S, Stock KP, Pichl J, Schneider MU, Domschke W. Beta-cell
19. Jones S, Hruban RH, Kamiyama M, et al. Exomic sequencing identifies reserve capacity in chronic pancreatitis. Hepatogastroenterology 1985;32:
PALB2 as a pancreatic cancer susceptibility gene. Science 2009;324:217 27–30
20. Roberts NJ, Jiao Y, Yu J, et al. ATM mutations in patients with hereditary 37. Ewald N, Kaufmann C, Raspe A, Kloer HU, Bretzel RG, Hardt PD. Prevalence
pancreatic cancer. Cancer Discov 2012;2:41–46 of diabetes mellitus secondary to pancreatic diseases (type 3c). Diabetes Metab

Downloaded from http://diabetesjournals.org/diabetes/article-pdf/66/5/1103/538525/db161477.pdf by guest on 06 April 2024

21. Hu C, Hart SN, Bamlet WR, et al. Prevalence of pathogenic mutations in Res Rev 2012;28:338–342
cancer predisposition genes among pancreatic cancer patients. Cancer Epidemiol 38. Ewald N, Bretzel RG. Diabetes mellitus secondary to pancreatic diseases
Biomarkers Prev 2016;25:207–211 (type 3c)–are we neglecting an important disease? Eur J Intern Med 2013;24:
22. Klein AP, Lindström S, Mendelsohn JB, et al. An absolute risk model to 203–206
identify individuals at elevated risk for pancreatic cancer in the general pop- 39. Andersen DK. The practical importance of recognizing pancreatogenic or
ulation. PLoS One 2013;8:e72311 type 3c diabetes. Diabetes Metab Res Rev 2012;28:326–328
23. Renehan AG, Zwahlen M, Egger M. Adiposity and cancer risk: new 40. Pelaez-Luna M, Takahashi N, Fletcher JG, Chari ST. Resectability of pre-
mechanistic insights from epidemiology. Nat Rev Cancer 2015;15:484–498 symptomatic pancreatic cancer and its relationship to onset of diabetes: a ret-
24. Park J, Morley TS, Kim M, Clegg DJ, Scherer PE. Obesity and cancer– rospective review of CT scans and fasting glucose values prior to diagnosis. Am J
mechanisms underlying tumour progression and recurrence. Nat Rev Endocrinol Gastroenterol 2007;102:2157–2163
2014;10:455–465 41. Permert J, Ihse I, Jorfeldt L, von Schenck H, Arnquist HJ, Larsson J. Im-
25. Funahashi H, Satake M, Dawson D, et al. Delayed progression of pancreatic proved glucose metabolism after subtotal pancreatectomy for pancreatic cancer.
intraepithelial neoplasia in a conditional Kras(G12D) mouse model by a selective Br J Surg 1993;80:1047–1050
cyclooxygenase-2 inhibitor. Cancer Res 2007;67:7068–7071 42. Pfeffer F, Koczan D, Adam U, et al. Expression of connexin26 in islets of
26. Vongsuvanh R, George J, Qiao L, van der Poorten D. Visceral adiposity in Langerhans is associated with impaired glucose tolerance in patients with
gastrointestinal and hepatic carcinogenesis. Cancer Lett 2013;330:1–10 pancreatic adenocarcinoma. Pancreas 2004;29:284–290
27. Hertzer KM, Xu M, Moro A, et al. Robust early inflammation of the peri- 43. Basso D, Greco E, Fogar P, et al. Pancreatic cancer-derived S-100A8
pancreatic visceral adipose tissue during diet-induced obesity in the KrasG12D N-terminal peptide: a diabetes cause? Clin Chim Acta 2006;372:120–128
model of pancreatic cancer. Pancreas 2016;45:458–465 44. Huang H, Dong X, Kang MX, et al. Novel blood biomarkers of pancreatic
28. Dawson DW, Hertzer K, Moro A, et al. High-fat, high-calorie diet promotes cancer-associated diabetes mellitus identified by peripheral blood-based gene
early pancreatic neoplasia in the conditional KrasG12D mouse model. Cancer expression profiles. Am J Gastroenterol 2010;105:1661–1669
Prev Res (Phila) 2013;6:1064–1073 45. He XY, Yuan YZ. Advances in pancreatic cancer research: moving towards
29. Philip B, Roland CL, Daniluk J, et al. A high-fat diet activates oncogenic Kras early detection. World J Gastroenterol 2014;20:11241–11248
and COX2 to induce development of pancreatic ductal adenocarcinoma in mice. 46. Sah RP, Nagpal SJS, Mukhopadhyay D, Chari ST. New insights into pan-
Gastroenterology 2013;145:1449–1458 creatic cancer-induced paraneoplastic diabetes. Nat Rev Gastroenterol Hepatol
30. Colmers IN, Bowker SL, Tjosvold LA, Johnson JA. Insulin use and cancer 2013;10:423–433
risk in patients with type 2 diabetes: a systematic review and meta-analysis of 47. Cui Y, Andersen DK. Pancreatogenic diabetes: special considerations for
observational studies. Diabetes Metab 2012;38:485–506 management. Pancreatology 2011;11:279–294
31. Bosetti C, Rosato V, Buniato D, Zambon A, La Vecchia C, Corrao G. Cancer 48. Hart PA, Baichoo E, Bi Y, Hinton A, Kudva YC, Chari ST. Pancreatic poly-
risk for patients using thiazolidinediones for type 2 diabetes: a meta-analysis. peptide response to a mixed meal is blunted in pancreatic head cancer asso-
Oncologist 2013;18:148–156 ciated with diabetes mellitus. Pancreatology 2015;15:162–166
32. Madiraju AK, Erion DM, Rahimi Y, et al. Metformin suppresses gluconeo- 49. Hardt PD, Brendel MD, Kloer HU, Bretzel RG. Is pancreatic diabetes (type 3c
genesis by inhibiting mitochondrial glycerophosphate dehydrogenase. Nature diabetes) underdiagnosed and misdiagnosed? Diabetes Care 2008;31(Suppl. 2):
2014;510:542–546 S165–S169