Diabetes Associated With Pancreatic Diseases

Juris J. Meier; Arnd Giese


Curr Opin Gastroenterol. 2015;31(5):400-406. 

In This Article

Abstract and Introduction


Purpose of review A relevant number of patients with pancreatic disorders suffer from secondary diabetes. Recent data have shed light on the link between pancreatic damage and subsequent impairments in glucose homeostasis. Furthermore, epidemiological studies provided insights into the relationship between diabetes and the risk of pancreatic carcinoma or pancreatitis. Pancreaticogenic diabetes requires a tailored therapeutic approach taking into account the individual properties of the available glucose-lowering drugs.

Recent findings We review the available literature concerning diabetes in patients with acute or chronic pancreatitis or pancreatic carcinoma. The relationship between the pancreatic damage and alterations in insulin and glucose homeostasis is summarized as well as the effect of diabetes mellitus on the risk of pancreatic cancer and pancreatitis. Caveats in the treatment of pancreaticogenic diabetes with currently available drugs are being discussed.

Summary Patients with pancreatic diseases should be screened for diabetes by means of an oral glucose tolerance test. There is a close inverse relationship between pancreatic β-cell loss and postchallenge hyperglycemia. The risk of hypoglycemia may be increased in patients with pancreaticogenic diabetes. Newly diagnosed diabetes may be a harbinger of pancreatic cancer. There is increasing evidence suggesting an increased risk for (pancreatic) cancer and pancreatitis in patients with diabetes mellitus. Further studies on the ideal glucose-lowering treatment of patients with pancreaticogenic diabetes will be required.


The human pancreas harbors ~1 million islets of ~50–400 μm in diameter each, which are scattered throughout the pancreas.[1] Each islet contains ~2000 insulin-secreting β-cells as well as glucagon-secreting alpha cells, somatostatin-producing delta cells and the pancreatic polypeptide cells secreting pancreatic polypeptide. There is some heterogeneity in the individual distribution of cells within the islets, with pancreatic polypeptide-containing islets being more predominant in the pancreatic head and glucagon-containing islets occurring preferentially in the pancreatic body and tail.

Diabetes has been recognized as a secondary complication of various pancreatic disorders, such as acute and chronic pancreatitis, and pancreatic cancer.[2] Hyperglycemia inevitably results from total pancreatectomy and may develop after partial pancreatectomy. In turn, exocrine pancreatic insufficiency occurs in a substantial proportion of patients with diabetes.[3]

Diabetes in Patients With Chronic Pancreatitis

The evolution of chronic pancreatitis is typically slow and characterized by a progressive loss of acinar tissue over ~10–20 years.[4] Calcifications and general pancreatic atrophy may develop in more advanced disease stages. Histologically, the early stages of the disease are characterized by periductal strains of fibrous tissue.[5] With further disease progression, fibrosis is found increasingly in the inter-lobar septa. In the advanced stages of the disease, the exocrine parenchyma can be almost completely replaced by strains of fibrosis. The alterations of the endocrine islets are rather modest in the early stages of chronic pancreatitis, and even in the advanced stages of the disease, the islets appear to resist the autodigestive process to a much greater extent than the acinar cells.[5] The residual islets in patients with chronic pancreatitis are often enlarged, whereas smaller islets are less frequently found[5] (Fig. 1). In a morphometric examination of human pancreatic tissue, we have found a 29% reduction in the fractional β-cell area of the pancreas from patients with chronic pancreatitis. Taking into account the additional reduction in total pancreatic mass (21%), the resulting deficit in β-cell mass was even more pronounced. Furthermore, an inverse relationship between pancreatic β-cell area and glucose levels has been described in patients with chronic pancreatitis[6,7] (Fig. 2), with a mean β-cell deficit of ~40–50% in patients with overt diabetes.[7] When examined at a clinically quiescent state, no differences in either β-cell apoptosis or new β-cell formation have been described, suggesting that destruction of islet cells may predominantly occur during episodes of overt pancreatitis.[5] The histological changes in the pancreas of patients with cystic fibrosis are largely similar to those observed in chronic pancreatitis patients.[8]

Figure 1.

Pancreatic sections from a patient without overt pancreatic disease (a), chronic pancreatitis (b) and pancreatic cancer (c) stained with hematoxilin and eosine. Arrows point to islets of Langerhans.

Figure 2.

Relationship between plasma glucose concentrations 120 min after oral glucose ingestion and the fractional b-cell area of the pancreas in 82 patients with pancreatic disorders exhibiting normal glucose tolerance (n=21, blue), impaired glucose tolerance and/or impaired fasting glucose (n=26, green) or diabetes (n=35, red). The correlation coefficient (r) was calculated from nonlinear regression analysis. Taken from [6].

Overt hyperglycemia typically occurs late in the clinical course of chronic pancreatitis.[4] In a cross-sectional study, 67% of patients with chronic pancreatitis exhibited impaired glucose tolerance or diabetes mellitus.[9] Loss of pancreatic β-cell function correlates with diminished exocrine pancreatic function.[9] It is likely that the low BMI that typically results from exocrine pancreatic insufficiency masks the manifestation of diabetes in the earlier stages of chronic pancreatitis. Phenotypically, diabetes associated with chronic pancreatitis is characterized by marked impairments in insulin secretion in response to various stimuli.[2,9–11] Also, the contribution of the incretin effect to the overall postprandial insulin responses was found to be impaired in chronic pancreatitis patients.[12] Even though it is likely that the defects in insulin secretion occur as a consequence of reduced β-cell mass,[13] exocrine pancreatic insufficiency may play a role because the secretion of the insulinotropic hormones glucose-dependent insulinotropic polypeptide (GIP) and glucagon-like peptide 1 (GLP-1) from enteroendocrine cells is linked to the intestinal absorption of nutrients.[14]

Glucagon secretion can also be reduced in patients with chronic pancreatitis.[2,15] According to a study examining patients with advanced chronic pancreatitis, glucagon counter-regulation in response to hypoglycemia is markedly reduced.[15] Therefore, patients with chronic pancreatitis may exhibit an increased risk for developing hypoglycemia when treated with insulin or insulinotropic agents. In the advanced stages of the disease, the disturbances in insulin and glucagon secretion may lead to the clinical presentation of an unstable or 'brittle-type' diabetes, characterized by high glucose excursions as well as frequent hypoglycemic episodes.

In line with the predominant insulin deficiency in patients with diabetes secondary to chronic pancreatitis, hyperglycemia in such patients typically manifests in the postprandial state,[10] whereas fasting hyperglycemia develops later in the disease. Therefore, diabetes may easily be overlooked based on determinations of fasting glucose or HbA1c levels. In a study relating changes in pancreatic β-cell area to functional parameters of glucose homeostasis, diabetes was diagnosed based on 2-h glucose levels in an oral glucose tolerance test (OGTT), when β-cell area was reduced by 64% (Fig. 2). In contrast, when fasting glucose or HbA1c levels were used as diagnostic criteria, diabetes was diagnosed at a mean β-cell loss of 93 and 89%, respectively.[6] These findings suggest that oral glucose tolerance tests should be performed more frequently in patients with chronic pancreatitis.

Impaired secretion of pancreatic polypeptide has been recognized as an additional feature of chronic pancreatitis.[16,17] A recent consensus statement has therefore proposed a lack of stimulated pancreatic polypeptide secretion as a specific sign of pancreaticogenic diabetes.[16] However, the sensitivity of measuring pancreatic polypeptide levels for the diagnosis of chronic pancreatitis is low.[18]

In addition to the alterations in islet hormone secretion, gastric emptying is typically accelerated in patients with exocrine pancreatic insufficiency, thereby increasing postprandial rises in glycemia. Supplementation of pancreatic enzymes has been found to largely correct these abnormalities.[19]

Insulin resistance is uncommon in the majority of patients with chronic pancreatitis, likely because of the concomitant intestinal malabsorption and the rather low body weight.[10] However, with increasing BMI, insulin sensitivity can also decline in patients with chronic pancreatitis, thereby increasing the individual risk of the patients for the development of overt hyperglycemia.[6]

Diabetes in Patients With Acute Pancreatitis

Acute pancreatitis is a potentially life-threatening condition, which may cause severe disturbances in glucose metabolism during its initial phase. Glucose concentrations have therefore been integrated in the earliest scoring system used to assess the severity of acute pancreatitis – the Ranson criteria.[20] However, elevated pancreatic enzymes may be a simple bystander of a decompensated diabetes mellitus. In a case series, 21% of patients with diabetic ketoacidosis displayed elevated serum amylase levels, whereas only 11% of them had typical features of acute pancreatitis at computed tomography imaging.[21] A recent meta-analysis revealed a pooled prevalence of diabetes after acute pancreatitis of 23%. The risk of developing diabetes mellitus increased significantly 5 years after an acute pancreatitis [relative risk (RR): 2.7].[22]

Diabetes in Patients With Pancreatic Adenocarcinoma

The islets in the pancreas of patients with pancreatic adenocarcinoma are often reduced in number. The remaining islets are typically embedded in a dense network of fibrous stroma and sometimes accompanied by inflammatory infiltrates (Fig. 1). In an evaluation of 25 pancreatic tissue specimens from patients with pancreatic cancer, no overt reduction in β-cell area was found.[23] However, in these tissue samples, an increased number of replicating β-cells and of isolated β-cells was described.[23] These findings are contrasted by another study examining the islet morphology more specifically within, adjacent to and more distant from the tumor tissue, in which a reduction of both β and alpha-cells was described in the intra and peritumoral tissue.[24]

About 45–65% of patients with pancreatic carcinoma suffer from diabetes,[25–27] and the percentage of patients presenting with new-onset hyperglycemia at the time of diagnosis of pancreatic cancer can be as high as 80%.[28] Short duration of diabetes or hyperglycemia makes the diagnosis of pancreatic cancer more likely.[26,27] On that basis, it has been suggested that new onset maturity diabetes be considered a risk factor for pancreatic cancer that initiates a search for the neoplasm. However, the low prevalence of pancreatic cancer and high prevalence of type 2 diabetes mellitus make this an inefficient screening criterion.[29] The cause of hyperglycemia in patients with pancreatic cancer has not yet been fully elucidated. Possible mechanisms include the destruction of endocrine islets by tumor infiltration, functional changes in insulin secretion induced by the local release of cytokines as well as the induction of insulin resistance. The latter two mechanisms are supported by the observation that diabetes in patients with pancreatic cancer often resolves after surgical removal of the tumor.[10,28]

Diabetes After Pancreas Resection

Complete pancreatectomy inevitably leads to diabetes. The only way to successfully prevent diabetes after total pancreas resection is islet autotransplantation. In a longitudinal study, this procedure has proven to maintain normal glucose homeostasis for up to 13 years after surgery.[30] Unfortunately, islet autotransplantation is only available in a limited number of centers.

Hemipancreatectomy in healthy organ donors results in decreased insulin levels and increased glucose levels.[31] Although various studies in young rodents have demonstrated a marked capacity for β-cell regeneration after partial pancreatectomy,[32,33] such new beta-cell formation is not found in the pancreas of adult humans undergoing partial pancreatic resection.[34]

In a study of 28 healthy organ donors, 25% of the patients developed abnormal glucose tolerance 1 year after hemipancreatectomy.[35] Fifty percent of eight healthy hemipancreatectomized donors developed diabetes during a long-term follow-up (median 13 years). Obesity was found to increase the individual risk of diabetes.[36] In a follow-up study of 74 patients undergoing hemipancreatectomy, we have found a diabetes incidence of 26% 2.5 years after surgery.[7] Modest elevations in fasting glucose, HbA1c and BMI were identified as predictors of postoperative diabetes.[7]

The changes in glucose metabolism after hemipancreatectomy largely depend on the surgical procedure. Thus, whereas resections of the pancreatic head may blunt the rise in glucose levels after oral glucose ingestion, resection of the pancreatic tail may immediately impair glucose metabolism.[10] The transient improvements in postchallenge glucose levels after pancreatic head resection may result from a delay in gastric emptying, impaired nutrient absorption as well as changes in the secretion of the incretin hormones GIP and GLP-1. Interestingly, the reduction in endogenous insulin secretion after partial pancreatectomy may also secondarily impair the insulin-mediated inhibition of α-cell secretion, thereby explaining the paradoxical observation of increased postchallenge glucagon levels after surgery.[37] The alterations in glucose homeostasis provide a strong rationale for performing OGTTs at frequent intervals (e.g. once per year) after partial pancreatectomy.

Effect of Diabetes Mellitus on the Risk of Pancreatic Cancer and Pancreatitis

Diabetes mellitus is a risk factor for different types of cancer.[38,39] There is congruent evidence that diabetes imparts a two-fold increase in risk for pancreatic cancer.[40,41] This risk seems to be increased by a history of pancreatitis[42] or by other hepatobiliar diseases.[43] Also obesity, which often accompanies type 2 diabetes mellitus, has been identified as an independent risk factor for pancreatic cancer. The pathologic link between diabetes and pancreatic cancer has not yet been completely elucidated. There is evidence that hyperglycemia and insulin resistance lead to the formation of reactive oxygen species in acinar cells,[44] which may promote cell dysplasia and cancer formation.

Diabetes is also a risk factor for both acute and chronic pancreatitis. In a meta-analysis, type 2 diabetes mellitus was associated with an increased risk of acute pancreatitis (RR: 1.84).[45] These data are supported by a recent large observational study showing an odds ratio for acute pancreatitis of 1.86.[46] Abnormal exocrine function has long been recognized in patients with type 1 and – to a lesser extent – type 2 diabetes. Histologically, periductal fibrosis is often found in the pancreas of patients with type 1 and type 2 diabetes mellitus.[47,48] Furthermore, transient elevations in circulating lipase and amylase concentrations occur in ~20–30% of patients with type 2 diabetes mellitus.[49]

Treatment of Diabetes Mellitus in Chronic Pancreatitis

Diabetes mellitus secondary to pancreatic diseases occurs less frequently than type 1 or type 2 diabetes mellitus. However, in tertiary hospitals, as many as 8% of all patients with diabetes may suffer from pancreatogenic diabetes.[50] Even though first attempts were made toward specific treatment recommendations,[16] there is still no consensus guideline on the treatment of pancreaticogenic diabetes. Also, prospective studies examining the effects of different glucose-lowering agents in such patients are lacking. Generally speaking, the treatment of pancreaticogenic diabetes follows the therapeutic principles of type 2 diabetes mellitus.[51] However, some aspects require special attention (Table 1). Metformin forms the basis of the pharmacotherapy of type 2 diabetes. However, because insulin resistance is typically not very prominent in patients with pancreatogenic diabetes,[6] the efficacy of the drug in such patients may be limited. Furthermore, metformin may lead to diarrhea and abdominal pain, which are already present in a large number of patients with pancreatic disorders. Sulfonylureas and glinides may be more suitable because of their insulinotropic action. Also, weight gain might be viewed as an advantage in some patients with pancreatogenic malabsorption. However, the risk of hypoglycemia with these drugs might be increased in patients with alterations in glucagon secretion. Alpha-glucosidase inhibitors specifically lower postprandial glucose excursions, but might further aggravate diarrhea, bloating and potentially malnutrition. GLP-1 analogues and dipeptidyl-peptidase-4 (DPP-4) inhibitors enhance insulin secretion without inducing hypoglycemia.[52] Loss of appetite, nausea, diarrhea and weight loss might be particularly unfavorable actions of the GLP-1 analogues in patients with pancreatic disorders.[52] In addition, the induction of pancreatitis and pancreatic cancer by incretin-based therapies has been widely debated.[53] In prospective clinical trials, no increase in the risk of pancreatitis or pancreatic cancer has been found with these drugs, but the number of patients included in these trials is still too small to allow for a final conclusion in this regard.[54] Increases in lipase and amylase levels of yet unknown clinical significance have been consistently seen with incretin-based therapies.[52,53] Therefore, it seems advisable to avoid these drugs in patients with pancreaticogenic diabetes. Glitazones may theoretically be used in patients with pancreatic disorders, but aside from the general safety concerns regarding heart failure, bone fracture risk, etc., improvements in peripheral insulin resistance may not address the pathophysiological needs of these patients. Sodium-glucose cotransporter-2 inhibitors may be an option, but the weight loss of ~2–3 kg should be considered, especially in malnourished patients. In light of the limitations of the above-mentioned glucose-lowering drugs with respect to the treatment of patients with pancreatic disorders, insulin therapy often remains the ultima ratio. The insulin-induced weight gain may be an advantage in patients with malnutrition, and is even an explicit therapeutic goal in the treatment of patients with cystic fibrosis.[55] Clearly, hypoglycemia must be considered as the major limitation of all insulin-based regimens. Which insulin regimen is most favorable in these patients is an open question. Finally, glucose metabolism in patients with chronic pancreatitis may also be improved by pancreatic enzyme substitution.[56]