18: Disorders of the Endocrine Pancreas

Insulin and glucagon, the two key hormones that orchestrate fuel storage and utilization, are produced by the islet cells in the pancreas. Islet cells are distributed in clusters throughout the exocrine pancreas. Together, they comprise the endocrine pancreas. Diabetes mellitus, a heterogeneous disorder, is the most common disease of the endocrine pancreas. Affecting 8% of the world’s adult population in 2011, the prevalence of diabetes continues to increase worldwide, having already more than doubled over the past 3 decades. Pancreatic tumors secreting excessive amounts of specific islet cell hormones are far less common, but their clinical presentations underscore the important regulatory roles of each hormone.

Anatomy & Histology

The endocrine pancreas is composed of nests of cells (islets of Langerhans) that are distributed throughout the exocrine pancreas. This anatomic feature allows for their enzymatic isolation from the exocrine pancreas for islet cell transplantation. Although numbering in the millions, the multicellular islets comprise only 1% of the total pancreas. The endocrine pancreas has great reserve capacity; more than 70% of the insulin-secreting β cells must be lost before dysfunction occurs. Each of the four major islet cell types produces a different secretory product. Insulin-secreting β cells are the predominant cell type (60%). The majority of the remaining islet cells, glucagon-secreting α cells (30%) and somatostatin-secreting δ cells (<10%), secrete hormones that counter the effects of insulin. A fourth islet cell type, the pancreatic polypeptide (PP)–secreting cell (<1%), is primarily located in the posterior lobe of the head of the pancreas, an embryologically distinct region receiving a different blood supply.

The islets are much more highly vascularized than the exocrine pancreatic tissues (Chapter 15), with at least one major arteriole supplying each islet. The majority of islet cells are closely apposed to the vasculature and to islet cells of opposing types, suggesting an important role for endocrine (via the microcirculation) and/or intra-islet paracrine (via the interstitium) effects on hormone release (Figure 18–1). Blood from the islets then drains into the hepatic portal vein. Thus, the islet cell hormones pass directly into the liver, a major site of action of glucagon and insulin, before proceeding into the systemic circulation, allowing for much higher hepatic than systemic levels of pancreatic hormones.

The islets are also abundantly innervated. Both parasympathetic and sympathetic axons enter the islets and either directly contact cells or terminate in the interstitial space between the cells. Neural regulation of islet cell hormone release, both directly through the sympathetic fibers and indirectly through stimulation of catecholamine release by the adrenal medulla, plays a key role in glucose homeostasis during stress.

Physiology

Insulin

Synthesis and Metabolism of Insulin

Insulin is a protein composed of two peptide chains (A and B chains) connected by two disulfide bonds (Figure 18–2). The precursor of insulin, preproinsulin (MW 11,500), is synthesized in the ribosomes and enters the endoplasmic reticulum of β cells, where it is promptly cleaved by microsomal enzymes to form proinsulin (MW 9000). Proinsulin, consisting of A and B chains joined by a 31-amino-acid C peptide, is transported to the Golgi apparatus, where it is packaged into secretory vesicles. While in the secretory vesicle, proinsulin is cleaved at two sites to form insulin (51 amino acids; MW 5808) and the C peptide fragment (Figure 18–2). Secretion of insulin is, therefore, accompanied by an equimolar secretion of C peptide and also by small amounts of proinsulin that escape cleavage. In the acidic environment of the secretory granules, stored insulin forms a hexamer in association with zinc atoms, dissociating into active monomers upon secretion. Insulin has a circulatory half-life of 3–5 minutes and is catabolized in both the liver and the kidney. Approximately 50% of insulin is catabolized on its first pass through the liver after it is secreted from the pancreas into the portal vein. In contrast, both C peptide and proinsulin are catabolized only by the kidney and, therefore, have half-lives three to four times longer than that of insulin itself. Recombinant human insulin or related analogs, which either enhance monomer formation (rapid acting) or decrease solubility (longer acting), are used clinically to treat diabetes.

Regulation of Secretion

Glucose is the primary physiologic stimulant of insulin release (Figure 18–3). Glucose entry into β cells is facilitated by one or more glucose transporters (GLUT-1, GLUT-2, and/or GLUT-3), which are in excess to glucose and allow for the bidirectional transport of glucose, thereby creating an equilibrium between extracellular and intracellular glucose concentrations. Once in the cell, the metabolism of glucose—rather than glucose itself—stimulates insulin secretion.

Glucokinase, an enzyme with low affinity for glucose whose activity is regulated by glucose, controls the first and rate-limiting step in glucose metabolism—the phosphorylation of glucose to form glucose 6-phosphate. This enzyme, by determining the rate of glycolysis, is thought to function as the β-cell glucose sensor. Glycolysis produces an increase in adenosine triphosphate (ATP), which is sensed by the sulfonylurea receptor subunit of ATP-dependent K⁺ channels (K_ATP) in the β-cell membrane, resulting in a closure of the channel. The resultant cell depolarization allows Ca²⁺ to enter, triggering exocytosis of insulin-containing granules. Sulfonylurea drugs used to treat type 2 diabetes stimulate insulin secretion in a glucose-independent fashion by binding to the sulfonylurea receptor subunit and blocking K_ATP.

Although glucose is the most potent stimulator of insulin release, other factors such as amino acids ingested with a meal or vagal stimulation also result in insulin release (Table 18–1). Up to 50% of insulin secretion in response to an oral glucose load can be attributed to enteric hormones (incretins) such as glucagon-like peptide-1 (GLP-1), which are released following oral ingestion of nutrients and enhance glucose-stimulated insulin secretion in β cells via activation of cAMP/PKA signaling pathways following binding to their G protein–coupled receptors. Glucagon similarly enhances glucose-stimulated insulin secretion, a counter-regulatory effect that allows for insulin-mediated disposal of glucose following the hepatic production of glucose induced by glucagon. Insulin secretion is inhibited by catecholamines and by somatostatin.

Mechanism of Action

Insulin exerts its effects by binding to insulin receptors present on the surfaces of target cells (Figure 18–4). Insulin receptors are present in liver, muscle, and fat, the classic insulin-sensitive tissues responsible for fuel homeostasis. In addition, insulin can mediate other effects in nonclassic target tissues, such as the ovary, via interaction with insulin receptors or by cross-reactivity with insulin-like growth factor-1 (IGF-1) receptors. Binding of insulin to its receptor causes activation of a tyrosine kinase region of the receptor and autophosphorylation of the receptor. Activation of the insulin receptor initiates a phosphorylation cascade within the cell, beginning with the phosphorylation of a network of docking proteins (insulin receptor substrates [IRSs]) that engage and amplify downstream signaling molecules, ultimately leading to the biologic effects of insulin (eg, translocation of GLUT-4 glucose transporter to the plasma membranes of muscle and fat cells and activation of hepatic glycogen synthase).

Effects

Insulin plays a major role in fuel homeostasis (Table 18–2). Insulin mediates changes in fuel metabolism through its effects on three main tissues: liver, muscle, and fat. In these tissues, insulin promotes fuel storage (anabolism) and prevents the breakdown and release of fuel that has already been stored (catabolism). The total lack of insulin is incompatible with life, and the same is true of excess insulin.

In the liver, insulin promotes fuel storage by stimulation of glycogen synthesis and storage. Insulin inhibits hepatic glucose output by inhibiting gluconeogenesis (glucose synthesis) and glycogenolysis (glycogen breakdown). By also stimulating glycolysis (metabolism of glucose to pyruvate), insulin promotes the formation of precursors for fatty acid synthesis. Insulin stimulates lipogenesis (biosynthesis of fatty acids from glucose) while inhibiting fatty acid oxidation and the production of ketone bodies (ketogenesis), an alternative fuel produced only in the liver that can be used by the brain when glucose is not available.

Although hepatic uptake of glucose, occurring via low-affinity GLUT-2 transporters, is not regulated by insulin, glucose uptake both in muscle and in fat is regulated by insulin, which causes the rapid translocation of an insulin-sensitive glucose transporter (GLUT-4) to the surface of these cells. Uptake of glucose by muscle accounts for the vast majority (85%) of insulin-stimulated glucose disposal. In muscle, insulin promotes the storage of glucose by stimulating glycogen synthesis and inhibiting glycogen catabolism. Insulin also stimulates protein synthesis in muscle.

Insulin stimulates fat storage in adipose tissue by stimulating lipoprotein lipase, the enzyme that hydrolyzes the triglycerides carried in very-low-density lipoproteins (VLDLs) and other triglyceride-rich lipoproteins to fatty acids, which can then be taken up by fat cells. Increased glucose uptake caused by upregulation of the GLUT-4 transporter also aids in fat storage because this increases the levels of glycerol phosphate, a substrate in the esterification of free fatty acids, which are then stored as triglycerides. In fat cells, insulin also inhibits lipolysis, preventing the release of fatty acids, which are potential substrates for hepatic ketone body synthesis and/or hepatic VLDL-triglyceride synthesis. Insulin exerts this effect by preventing phosphorylation of hormone-sensitive lipase, thus inactivating the enzyme that hydrolyzes stored triglycerides to releasable fatty acids. Together, these changes result in increased fat storage in adipose tissue.

Glucagon

Synthesis and Metabolism

Glucagon, a 29-amino-acid peptide, is produced in α cells of the pancreas by the proteolytic processing of proglucagon, a larger precursor protein. In addition to the pancreas, proglucagon is also expressed in the intestine and brain. While glucagon is the major bioactive metabolite produced in the pancreatic α cell, differential processing by L cells in the intestine results in the production of glucagon-like peptide (GLP)-1 and GLP-2 in response to a meal (Figure 18–5). This tissue-specific processing results in peptides with opposing effects on carbohydrate metabolism; pancreatic glucagon opposes the hepatic effects of insulin, whereas GLPs acts as incretins, gut-derived peptides that enhance glucose-stimulated insulin secretion. The circulatory half-life of glucagon is 3–6 minutes. Like insulin, glucagon is metabolized in the liver and kidneys, with kidneys, rather than liver, playing a significant role. Long-acting analogs of GLP-1, which also stimulate β-cell proliferation and increase β-cell mass, or enzyme inhibitors that extend the half-life of endogenous GLIP-1 are a newer and important class of drugs used for the treatment of type 2 diabetes mellitus.

Regulation of Secretion

In contrast to the stimulation of insulin secretion by glucose, glucagon secretion is inhibited by glucose (Table 18–1). However, the relative importance of direct sensing of glucose by the α cell vs. indirect paracrine/endocrine effects of other pancreatic factors in regulating glucagon secretion is a matter of debate. Current evidence suggests that insulin plays a major role in modulating (ie, inhibiting) glucagon secretion. Moreover, the loss of the suppressive effects of increased pancreatic insulin in diabetes in response to hyperglycemia results in an inappropriately high level of glucagon, which contributes to the hyperglycemia of diabetes mellitus. Other pancreatic factors inhibiting glucagon secretion include somatostatin and two additional β-cell secretory products, γ-aminobutyric acid (GABA), and insulin-associated zinc. Like insulin, glucagon secretion is stimulated by amino acids, an important regulatory feature in the metabolism of protein meals. In contrast, fatty acids and ketones inhibit glucagon secretion. Other counter-regulatory hormones such as catecholamines (via a predominating β-adrenergic effect) and cortisol stimulate glucagon release.

Mechanism of Action

The major biological role of glucagon is to maintain normal glucose levels during fasting by inducing hepatic glucose production, thus counteracting the hepatic effects of insulin. Therefore, the liver is the major target organ for glucagon action. Glucagon binds to a G protein–coupled glucagon receptor present on the cell surface of hepatocytes, activating adenylyl cyclase and generating cAMP. Cyclic AMP activates protein kinase A, which activates gene transcription for the enzymes responsible for the biologic activity of glucagon in the liver, and subsequently phosphorylates and activates these same enzymes. There is also some evidence that the glucagon receptor may act via an adenylyl cyclase-independent mechanism by stimulation of phospholipase C.

Effects

The actions of glucagon were first demonstrated in 1921 by Banting and Best when they observed a mild transient hyperglycemia preceding insulin-induced hypoglycemia when testing pancreatic extracts in vivo. Glucagon is a counter-regulatory hormone, acting in a catabolic fashion to oppose the effects of insulin. Indeed, glucagon injections are used clinically to treat severe hypoglycemia. Hepatic effects of glucagon (Table 18–2) include the following: (1) increased hepatic glucose output via the release of glycogen stores (glycogenolysis) and, in concert with other counter-regulatory hormones, stimulation of hepatic glucose synthesis (gluconeogenesis); (2) increased hepatic uptake of amino acids, which fuels gluconeogenesis; and (3) stimulation of fatty acid oxidation and ketogenesis, thus providing an alternative fuel (ketone bodies) that can be used by the brain when glucose is not available. The physiologic significance of glucagon receptors in non-hepatic tissue (kidney, adipose, pancreas) is less certain. For example, glucagon, while less potent, shares with GLIP-1 the ability to enhance glucose-induced β-cell insulin secretion.

Somatostatin

Synthesis, Metabolism, and Regulation of Secretion

Like preproglucagon, preprosomatostatin is synthesized in the pancreas, GI tract, and brain, where it is differentially processed in a tissue-specific fashion to produce several biologically active peptides. Somatostatin-14 (SS-14), the first somatostatin to be isolated, is a 14-amino-acid peptide that was initially discovered in the hypothalamus as the factor responsible for the inhibition of growth hormone release. Only later was it appreciated that δ cells of the pancreas also secrete SS-14. In brain and intestine, somatostatin-28 (SS-28), an amino-terminally extended peptide that includes the 14-amino-acid sequence of SS-14, is also produced from preprosomatostatin and has a range of action comparable to that of SS-14 but a potency that is somewhat greater. The half-life of somatostatin (<3 minutes) is shorter than that of insulin or glucagon. Because somatostatin has been shown to inhibit the synthesis and secretion of most peptide hormones, synthetic somatostatin analogs, such as octreotide, that have a much longer half-life (hours) have been developed for clinical use in inhibiting ectopic peptide hormone production by a variety of tumors. The same secretagogues that stimulate insulin secretion also stimulate somatostatin (Table 18–1). These include glucose, amino acids, enteric hormones, and glucagon.

Mechanism of Action and Effects

Somatostatin exerts its effects via binding to a family of inhibitory G (G_i) protein–coupled receptors (SST1-5) that are distributed in a tissue-specific fashion. In all tissues where somatostatin is produced, it acts primarily in an inhibitory fashion. In the endocrine pancreas, somatostatin is thought to act via paracrine effects on the other islet cells, inhibiting the release of insulin and glucagon (Table 18–1) and of PP. In addition, somatostatin acts in an autocrine fashion to inhibit its own release. In the GI tract, somatostatin retards the absorption of nutrients through multiple mechanisms, including the inhibition of gut motility, inhibition of several enteric peptides, and inhibition of pancreatic exocrine function. Consistent with the multiple inhibitory effects of this peptide, the synthetic somatostatin analog octreotide has multiple clinical uses, including inhibition of hormone production by pituitary adenomas, inhibition of certain types of chronic diarrhea, inhibition of tumor growth, and inhibition of bleeding from esophageal varices.

Pancreatic Polypeptide

Pancreatic peptide (PP), a 36-amino-acid peptide is produced by the PP cells (F cells) in the islets of the posterior lobe of the head of the pancreas, is released in response to a mixed meal, an effect that appears to be mediated by protein and vagal stimulation. While long known to inhibit gastrointestinal motility and pancreatic exocrine secretions, more recent evidence suggests that PP may also control satiety and weight, inhibiting food intake and stimulating energy expenditure. These latter effects of PP (a member of the neuropeptide Y family of peptide hormones) are mediated centrally, via binding to an inhibitory G protein–coupled Y4 receptor and are thought to involve inhibition of hepatic vagal nerve afferent activity.

Hormonal Control of Carbohydrate Metabolism

Carbohydrate metabolism is primarily controlled by the relative amounts of insulin and glucagon produced by the endocrine pancreas (Table 18–2; Figure 18–6). Conversely, dysregulation of both of these hormones contributes to hyperglycemia in diabetes. Under normal conditions, when plasma glucose levels are high, the actions of insulin predominate, including insulin suppression of glucagon secretion. Fuel storage is promoted by insulin stimulation of glycogen storage in the liver; glucose uptake, glycogen synthesis, and protein synthesis by muscle; and fat storage by adipose tissue. Insulin inhibits the mobilization of substrates from peripheral tissues and opposes any effects of glucagon on the stimulation of hepatic glucose output.

In contrast, when glucose levels are low, plasma insulin levels are suppressed and the effects of glucagon predominate in the liver (ie, increased hepatic glucose output and ketone body formation). In the absence of insulin, muscle glucose uptake is markedly decreased, muscle protein is catabolized, and fat is mobilized from adipose tissue. Therefore, with insulinopenia, glucose loads cannot be cleared, and substrates for hepatic gluconeogenesis (amino acids, glycerol) and ketogenesis (fatty acids)—processes that are stimulated by glucagon—are increased.

Fasting State

After an overnight fast, the liver plays a primary role in maintaining blood glucose by producing glucose at the same rate at which it is used by resting tissues. Glucose uptake and utilization occur predominantly in tissues that do not require insulin for glucose uptake, such as the brain. Hepatic glucose output is stimulated by glucagon and is primarily due to glycogenolysis, which can provide, on average, an 8-hour supply of glucose. The low levels of insulin that are present (basal secretion of 0.25–1.0 unit/h) are insufficient to block the release of fatty acids from fat, which provide fuel for muscles (fatty acid oxidation) and substrate for hepatic ketogenesis. However, these levels of insulin are sufficient to prevent excessive lipolysis, ketogenesis, and gluconeogenesis, thus preventing hyperglycemia and ketoacidosis.

With prolonged fasting (>24–60 hours), liver glycogen stores are depleted. Glucagon levels rise slightly, and insulin levels decline further. Gluconeogenesis now becomes the sole source of hepatic glucose production, using substrates such as amino acids that are mobilized from the periphery at a greater rate. With starvation, a switch occurs in the liver from gluconeogenesis to the production of ketones, an alternative fuel source that provides 90% of the energy used by the brain, a critical organ that accounts for 25% of basal metabolic energy needs. In this manner, survival is prolonged as muscle protein is conserved in favor of increased mobilization of fatty acids from adipose tissue, a process made possible by increased insulinopenia. The liver then converts fatty acids to ketone bodies, a process that is stimulated by glucagon. With prolonged fasting or starvation, the kidney also begins to contribute significantly to gluconeogenesis.

Fed State

With ingestion of a carbohydrate load, insulin secretion is stimulated and glucagon is suppressed. Hepatic glucose production and ketogenesis are suppressed by the high ratio of insulin to glucagon. Insulin stimulates hepatic glycogen storage. Insulin-mediated glucose uptake, which occurs primarily in muscle, is also stimulated, as is muscle glycogen synthesis. Fat storage occurs in adipose tissue.

With ingestion of a protein meal, both insulin and glucagon are stimulated. In this way, insulin stimulates amino acid uptake and protein formation by muscle. However, stimulation of hepatic glucose output by glucagon counterbalances the tendency of insulin to cause hypoglycemia.

Conditions of Stress

During severe stress, when fuel delivery to the brain is in jeopardy, counter-regulatory hormones, in addition to glucagon, act synergistically. They maintain blood glucose levels by maximizing hepatic output of glucose and peripheral mobilization of substrates and by minimizing fuel storage. Glucagon and epinephrine act within minutes to elevate blood glucose, whereas the counter-regulatory effects of cortisol and growth hormone are not seen for several hours. Epinephrine, cortisol, and growth hormone stimulate glucagon release, whereas epinephrine inhibits insulin, thus maximally increasing the glucagon-insulin ratio. In addition, these three hormones act directly on the liver to increase hepatic glucose production and peripherally to stimulate lipolysis and inhibit insulin-sensitive glucose uptake. During severe stress, hyperglycemia may actually result from the combined effects of counter-regulatory hormones.

Similar but less marked effects occur in response to exercise when glucagon, catecholamines, and, to a lesser extent, cortisol help meet the several-fold increase in glucose utilization rates due to exercising muscle by increasing hepatic glucose output and lipolysis of fat stores, effects that are made possible by a lowering of insulin levels. Low insulin levels also allow muscles to use glycogen stores for energy.

Role of Renal Gluconeogenesis in Glucose Homeostasis

Kidney and liver both express the enzymes required to augment the glucose pool by gluconeogenesis and the secretion of glucose stored as glycogen. While the kidney contributes little to the glucose pool during an overnight fasting, it contributes approximately 50% of endogenous glucose production during a prolonged (>40 hours) fast. Gluconeogenesis predominates in the kidney as its glycogen stores are minimal, a process that is stimulated by epinephrine, inhibited by insulin, and unaffected by glucagon.

Diabetes Mellitus

Clinical Presentation

Diabetes mellitus is a heterogeneous disorder defined by the presence of hyperglycemia. Diagnostic criteria for diabetes include the following: (1) a fasting plasma glucose of 126 mg/dL or more, (2) classic symptoms of hyperglycemia plus a random plasma glucose of 200 mg/dL or more, or (3) a plasma glucose level of 200 mg/dL or more after an oral dose of 75 g of glucose (oral glucose tolerance test, OGTT). More recently, following the establishment of standardized assays, glycated hemoglobin (Hb_A1C), which correlates with chronic increases in glucose, has been used to diagnose diabetes when Hb_A1C levels 6.5% or more are documented using an appropriate methodology.

Hyperglycemia in all cases is due to a functional deficiency of insulin action. Deficient insulin action can be due to a decrease in insulin secretion by the β cells of the pancreas, a decreased response to insulin by target tissues (insulin resistance), or an increase in the counter-regulatory hormones that oppose the effects of insulin. The relative contributions of these three factors form the basis for the classification of this disorder into subtypes and also helps to explain the characteristic clinical presentations of each subtype (Table 18–3).

Diabetes prevalence worldwide, which has been increasing over the past few decades, reached 8% in 2011 in those 20 years or older (and a prevalence of 11% in the United States). More than 90% of cases of diabetes mellitus are believed to occur in the context of a genetic predisposition and are classified as either type 1 diabetes mellitus (DM) or type 2 DM (Tables 18–3 and 18–4). Type 1 DM is much less common than type 2 DM, accounting for 5–10% of cases of primary diabetes. Type 1 DM is characterized by autoimmune destruction of pancreatic β cells with resultant severe insulin deficiency. In a minority of patients, the cause of type 1 DM is unknown. The disease commonly affects individuals younger than 30 years; a bimodal peak in incidence occurs around age 5–7 years and at puberty. Although autoimmune destruction of the β cells does not occur acutely, clinical symptoms usually do. Patients present after only days or weeks of polyuria, polydipsia, and weight loss with markedly elevated serum glucose concentrations. Ketone bodies are also increased because of the marked lack of insulin, resulting in severe, life-threatening acidosis (diabetic ketoacidosis). Patients with type 1 DM require treatment with insulin.

Type 2 DM differs from type 1 DM in several distinct ways (Table 18–4): It accounts for the overwhelming majority of diabetes (90–95%); has a stronger genetic component; occurs most commonly in adults; increases in prevalence with age (ie, 18% of individuals older than 65 years worldwide, or 27% in the United States); occurs more commonly in Native American, Mexican American, and African American populations in the United States; and is associated with increased resistance to the effects of insulin at its sites of action as well as a decrease in insulin secretion by the pancreas. It is often (85% of cases) associated with obesity, an additional factor that increases insulin resistance. Thus, the rising prevalence of diabetes worldwide has been associated with an increasing prevalence of obesity (12%). Insulin resistance is the hallmark of type 2 DM. Because these patients often have varying amounts of residual insulin secretion that prevent severe hyperglycemia or ketosis, they often are asymptomatic and are diagnosed 5–7 years after the actual onset of disease (frank hyperglycemia) by the discovery of an elevated fasting glucose on routine screening tests. Population screening surveys show that a remarkable 30% of cases of type 2 DM in the United States, or 50% of cases worldwide, remain undiagnosed. Additionally, it is estimated that one-third of the adult population in the United States is insulin-resistant and hence in a pre-diabetic (normoglycemic) state. Once diagnosed with type 2 DM, most individuals (70%) are managed with lifestyle modification (eg, diet, exercise, weight management) alone or in combination with medications that (1) enhance endogenous glucose-independent insulin secretion (sulfonylureas), (2) amplify endogenous glucose-dependent insulin secretion (incretins, such as GLP-1), (3) decrease insulin resistance in hepatic or peripheral tissues (eg, metformin or glitazones, respectively), or (4) interfere with intestinal absorption of carbohydrates (eg, intestinal α-glycosidase inhibitors). A new class of drugs inhibiting the transporter responsible for renal glucose reabsorption (sodium-glucose co-transporter 2 [SGLT2]) is also being developed for use in type 2 DM. Type 2 diabetic patients do not usually require insulin treatment for survival. However, some patients with advanced type 2 DM are treated with insulin to achieve optimal glucose control.

An epidemic of type 2 DM is occurring worldwide, particularly in non-European populations; it has been estimated that 1 in 3 children born after 2000 will develop diabetes, particularly type 2 DM, in their lifetime. Thus, while type 1 DM remains the most common cause of diabetes in children younger than 10 years (regardless of ethnicity) and in older, non-Hispanic white children, type 2 DM accounts for more than 50% of the diagnoses in older children of Hispanic, African American, Native American, and Asian Pacific Islander ancestry. In all age groups and ethnicities, this increased incidence of type 2 DM is associated with obesity.

Other causes of diabetes, accounting for less than 5% of cases, include processes that destroy the pancreas (eg, pancreatitis), specifically inhibit insulin secretion (eg, genetic β-cell defects [MODY]), induce insulin resistance (eg, certain HIV protease inhibitors), or increase counter-regulatory hormones (eg, Cushing syndrome) (Table 18–3, part III). Clinical presentations in these cases depend on the exact nature of the process and are not discussed here.

Gestational diabetes mellitus occurs in pregnant women with an incidence ranging from 3–8% in the general population to up to 16% in Native American women (Table 18–3, part IV), may recur with subsequent pregnancies, and tends to resolve at parturition. The prevalence of gestational diabetes mellitus in a population varies in direct proportion to the prevalence of diabetes. Up to 50% of these women with gestational diabetes mellitus eventually progress to diabetes (predominantly type 2 DM). Gestational diabetes usually occurs in the second half of gestation, precipitated by the increasing levels of hormones such as chorionic somatomammotropin, progesterone, cortisol, and prolactin that have counter-regulatory anti-insulin effects. Because of its potential adverse effects on fetal outcome, gestational diabetes in the United States is currently diagnosed or ruled out by routine screening with an oral glucose load at 24 weeks of gestation in those with average risk or at the first prenatal visit in high-risk populations—obese, age older than 25 years, family history of diabetes, or member of an ethnic group with a high prevalence of diabetes.

Etiology

Type 1 Diabetes Mellitus

Type 1 DM is an autoimmune disease caused by the selective destruction of pancreatic β cells by T lymphocytes targeting ill-defined β-cell antigens. The incidence of type 1 DM, while much lower than that for type 2 DM, appears to be increasing worldwide. In early disease, lymphocytic infiltrates of macrophage-activating CD4+ cells and cytokine-secreting, cytotoxic CD8+ cells surround the necrotic β cells. Autoimmune destruction of the β cell occurs gradually over several years until sufficient β-cell mass is lost to cause symptoms of insulin deficiency. At the time of diagnosis, ongoing inflammation is present in some islets, whereas other islets are atrophic and consist only of glucagon-secreting α cells and somatostatin-secreting δ cells. Autoantibodies against islet cells and insulin, while appearing early in the course of disease, are thought to serve as markers, rather than mediators, of β-cell destruction. As such, they have been used to aid in the differential diagnosis of type 1 DM vs. type 2 DM in children (particularly with the rising incidence of type 2 DM in this population) and to assess the probability for development of type 1 DM in first-degree relatives who are at increased risk for type 1 DM (2–6% incidence vs. 0.3% annual incidence in the general population).

Islet cell antibodies (ICA), which include those directed against insulin (insulin autoantibody [IAA]), glutamic acid decarboxylase (GAD), a β-cell zinc transporter (ZnT8), and tyrosine phosphatase-IA2 protein (IA2), are each present in 50% of newly diagnosed diabetics and are highly predictive of disease onset in first-degree relatives. Overall, 70% of first-degree relatives positive for at least three of these antibodies develop disease within 5 years. Because the appearance of autoantibodies is followed by progressive impairment of insulin release in response to glucose (Figure 18–7), both criteria have been used with great success to identify at-risk first-degree relatives with the ultimate, but as yet unmet, goal of intervening to prevent diabetes. However, because only 15% of individuals newly diagnosed with type 1 DM have a positive family history, these screening methods cannot be used to identify the vast majority of individuals developing this low-incidence type of diabetes.

At least 50% of the genetic susceptibility for type 1 DM has been linked to the genes of the major histocompatibility complex (MHC) that encode class II human leukocyte antigens (HLA) molecules (DR, DQ, and DP) expressed on the surface of specific antigen-presenting cells such as macrophages. Class II molecules form a complex with processed foreign antigens or autoantigens, which then activates CD4 T lymphocytes via interaction with the T-cell receptor. Alleles at the HLA-DR or HLA-DQ loci have the strongest influence on the risk of type 1 DM. While 95% of individuals with type 1 DM have either DR3-DQ2 or DR4-DQ8 haplotypes, they share this genotype with 40% of the general population. In addition, only 6% of children with high-risk HLA types will develop diabetes. Thus, identification of HLA haplotypes remains a research tool.

While genetic susceptibility clearly plays a role in type 1 DM, the 50% concordance rate in identical twins, as well as the continuing increase in the incidence of type 1 DM since World War II, provides additional evidence that environmental factors may also play a critical role. Evidence suggests that viral infections, such as congenital exposure to rubella, may precipitate disease, particularly in genetically susceptible individuals. It is hypothesized that an immune response to foreign antigens may also incite β-cell destruction if these foreign antigens have some homology with islet cell antigens (molecular mimicry). For example, coxsackievirus infections are also associated to the onset of type 1 DM. One particular coxsackie viral protein shares homology with the islet cell antigen, GAD. Vitamin D deficiency also correlates with a greater risk of type 1 DM, which may partially explain the increased incidence of type 1 DM at higher latitudes.

Type 2 Diabetes Mellitus

Given the current obesity-associated epidemic of type 2 DM, environmental factors are clearly critical for the development of this disorder. And yet, the genetic components underlying type 2 DM are even stronger than those associated with type 1 DM. In type 2 DM, in contrast to the absolute lack of insulin in type 1 DM, two metabolic defects are responsible for hyperglycemia: (1) target tissue resistance to the effects of insulin and (2) inadequate pancreatic β-cell insulin secretion in the setting of insulin resistance.

Whether the primary lesion in type 2 DM is insulin resistance or defective β-cell insulin secretion continues to be debated. Several decades before the onset of clinical diabetes, insulin resistance and high insulin levels are present. This has led researchers to hypothesize that insulin resistance could be the primary lesion, resulting in a compensatory increase in insulin secretion that ultimately cannot be maintained by the pancreas (Figure 18–7). When the pancreas becomes “exhausted” and cannot keep up with insulin demands, clinical diabetes results.

Insulin resistance is the key factor linking obesity and type 2 DM. Nutritional excess from any source ultimately leads to increased free fatty acid (FFA) storage as triglyceride in adipose tissue. The increased release of various factors from adipose, particularly central (visceral) adipose tissue, drives insulin resistance. Critical mediators include the following: (1) toxic effects of excess free fatty acids released from adipose by lipolysis (lipotoxicity); (2) dysregulated secretion of fat-specific proteins (adipokines), such as adiponectin, an insulin-sensitizing hormone and the anti-diabetogenic hormone, leptin that acts centrally to control satiety and enhance insulin sensitivity; and 3) increased production of inflammatory cytokines within adipose tissue. For example, tumor necrosis factor (TNF) secretion from hypertropic adipocytes and macrophages attracted into adipose tissue by other inflammatory adipocyte secretory products (eg, macrophage chemoattractant protein-1 [MCP-1]) is thought to block peroxisome proliferator–activated receptor gamma (PPARδ). PPARδ, whose activity is enhanced by the glitzazone class of diabetes drugs, is an adipose transcription factor that decreases insulin resistance by altering adipokine secretion and decreasing FFA release.

Central (visceral) adipose tissue most closely correlates with insulin resistance since it is most susceptible to increased lipolysis due to (1) enhanced sensitivity to the stimulatory effects of counter-regulatory hormones (increased number of β-adrenergic receptors and increased local conversion of inactive cortisone to active cortisol due to high levels of type 1 11β-hydroxysteroid dehydrogenase) and (2) weaker suppressive effect of insulin due to lower insulin receptor affinity. Visceral adipose tissue drains directly into the portal vein, thus exposing the liver to high levels of FFA and altered adipokine levels, resulting in hepatic steatosis and insulin resistance, which manifests as increased hepatic glucose output and elevated fasting glucose levels. Increased FFA flux also results in increased lipid deposition in other insulin-target tissues, such as skeletal muscle, where it is associated with mitochondrial dysfunction and insulin resistance, resulting in impaired insulin-stimulated glucose disposal/transport after a meal due to decreased translocation of GLUT-4. Hyperinsulinemia also contributes to insulin resistance by downregulating insulin receptor levels and desensitizing downstream pathways. Hyperglycemia may lead to increased flux through otherwise minor glucose metabolic pathways that result in products associated with insulin resistance (eg, hexosamines).

The importance of obesity in the etiology of type 2 DM (85% of type 2 DM patients are obese) is underscored by the fact that even a 5–10% weight loss in obese individuals with type 2 DM can ameliorate or even terminate the disorder. However, while the majority of obese individuals are hyperinsulinemic and insulin resistant, most do not develop diabetes. Therefore, alternatively or additionally, a primary pancreatic β-cell defect is also postulated to contribute to the pathogenesis of type 2 DM. Beta-cell mass normally increases with obesity. However, in those who develop impaired glucose tolerance and, later, frank diabetes, β-cell apoptosis results in a decline in β-cell mass. Impairment of the acute release of insulin (first phase insulin release) that precedes sustained insulin secretion in response to a meal occurs well before the onset of frank diabetes. Lipid accumulation in β cells is also thought to contribute to impaired β-cell function by various mechanisms, including activation of the pro-apoptotic unfolded protein response (UPR) in the endoplasmic reticulum. Chronic exposure to hyperglycemia and elevated free fatty acids also contribute to impaired β-cell insulin secretion (glucolipotoxicity).

In the last 2 decades, a great deal of work has been directed toward identifying the genes that account for the strong genetic component of type 2 DM. Initial efforts targeting specific candidate genes have been followed by genome wide approaches, all of which have yielded useful information, including the identification of a small subset of cases of type 2 DM that are monogenic in origin. One monogenetic form of type 2 DM is maturity-onset diabetes of the young (MODY) (Table 18–3). This autosomal dominant disorder accounts for 1–5% of cases of type 2 DM and is characterized by the onset of mild diabetes in lean individuals before the age of 25 years. MODY is caused by mutations in one of six pancreatic genes, glucokinase, the β-cell glucose sensor, or five different transcription factors. In contrast, the vast majority of cases of type 2 DM are thought to be polygenic in origin, due to the inheritance of an interacting set of susceptibility genes. The list of genes linked to increased risk of type 2 DM is extensive and growing. However, genes associated with defects in insulin secretion account for less than 10% of the genetic risk of type 2 DM.

Pathology & Pathogenesis

No matter what the origin, all types of diabetes result from a relative deficiency of insulin action. In addition, glucagon levels can be inappropriately high. This high glucagon-insulin ratio creates a state similar to that seen in fasting and results in a superfasting milieu that is inappropriate for maintenance of normal fuel homeostasis (Table 18–2; Figure 18–6).

The resulting metabolic derangements depend on the degree of loss of insulin action. Adipose tissue is most sensitive to insulin action. Therefore, low insulin activity is capable of suppressing excessive lipolysis and enhancing fat storage. Higher levels of insulin are required to oppose glucagon effects on the liver and block hepatic glucose output. In normal individuals, basal levels of insulin activity are capable of mediating both of these responses, with the liver, in particular, being exquisitely responsive to changes in pancreatic insulin secretion due to its high sensitivity and exposure to elevated levels of insulin in the portal circulation. However, the ability of skeletal muscle to respond to a glucose load with insulin-mediated glucose uptake requires the stimulated secretion of additional insulin from the pancreas.

Mild deficiencies in insulin action are, therefore, frequently manifested by an inability of insulin-sensitive tissues (eg, skeletal muscle which is responsible for 85% of postprandial glucose clearance) to clear glucose loads. Clinically, this results in postprandial hyperglycemia (Figure 18–7). Such individuals, most commonly type 2 diabetics with residual insulin secretion but increased insulin resistance, will have abnormal oral glucose tolerance test results and/or high nonfasting (postprandial) glucose levels. However, fasting glucose levels remain normal because sufficient insulin action is present to counterbalance the glucagon-mediated hepatic glucose output that maintains them. When a further loss of insulin action occurs, glucagon’s effects on the liver are not sufficiently counterbalanced. Individuals, therefore, have both postprandial hyperglycemia and fasting hyperglycemia (Figure 18–7). Interestingly, skeletal tissue remains insulin sensitive in some prediabetic individuals who can present instead with isolated increases in hepatic glucose output and fasting glucose levels. Because of the importance of excessive hepatic glucose output in the pathogenesis of type 2 DM (driven by insulin resistance and inappropriately high levels of glucagon), metformin, a drug that specifically targets hepatic glucose output, is used as a first-line treatment in these individuals.

Although type 2 diabetics usually have some degree of residual endogenous insulin action, type 1 diabetics have none. Therefore, untreated or inadequately treated type 1 diabetics manifest the most severe signs of insulin deficiency. In addition to fasting and postprandial hyperglycemia, they also develop ketosis because a marked lack or absolute deficiency of insulin allows maximal lipolysis of fat stores to supply substrates for unopposed glucagon stimulation of ketogenesis in the liver.

Fatty acids liberated from increased lipolysis, in addition to being metabolized by the liver into ketone bodies, can also be reesterified and packaged into VLDLs. Furthermore, insulin deficiency causes a decrease in lipoprotein lipase, the enzyme responsible for hydrolysis of VLDL triglycerides in preparation for fatty acid storage in adipose tissue, thereby slowing VLDL clearance. Therefore, both type 1 and type 2 diabetics can have hypertriglyceridemia as a result of both an increase in VLDL production and a decrease in VLDL clearance.

Because insulin stimulates amino acid uptake and protein synthesis in muscle, the decrease in insulin action in diabetes results in decreased muscle protein synthesis. Marked insulinopenia, such as occurs in type 1 DM, can cause negative nitrogen balance and marked protein wasting. Amino acids not taken up by muscle are instead diverted to the liver where they are used to fuel gluconeogenesis.

In type 1 DM or type 2 DM, the superimposition of stress-induced counter-regulatory hormones on what is already an insulinopenic state exacerbates the metabolic manifestations of deficient insulin action. The stress of infection, for example, can, therefore, induce diabetic ketoacidosis in both type 1 and some type 2 diabetics.

In addition to the metabolic derangements discussed previously, diabetes causes other chronic complications that are responsible for the high morbidity and mortality rates associated with this disease. Diabetic complications are largely the result of vascular disease affecting both the microvasculature (retinopathy, nephropathy, and some types of neuropathy) and the macrovasculature (coronary artery disease, peripheral vascular disease).

Clinical Manifestations

Acute Complications

1. Hyperglycemia—When elevated glucose levels exceed the renal threshold for reabsorption of glucose, glucosuria results. This causes an osmotic diuresis manifested clinically by polyuria, including nocturia. Dehydration results, stimulating thirst that results in polydipsia. A significant loss of calories can result from glucosuria, because urinary glucose losses can exceed 75 g/d (75 g × 4 kcal/g = 300 kcal/d). Polyphagia also accompanies uncontrolled hyperglycemia. The three “polys” of diabetes—polyuria, polydipsia, and polyphagia—are common presenting symptoms in both type 1 and symptomatic type 2 patients. Weight loss can also occur as a result of both dehydration and loss of calories in the urine. Severe weight loss is most likely to occur in patients with severe insulinopenia (type 1 DM) and is due to both caloric loss and muscle wasting. Increased protein catabolism also contributes to the growth failure seen in children with type 1 DM.

Elevated glucose levels raise plasma osmolality:

Changes in the water content of the lens of the eye in response to changes in osmolality can cause blurred vision.

In women, glucosuria can lead to an increased incidence of candidal vulvovaginitis. In some cases, this may be their only presenting symptom. In uncircumcised men, candidal balanitis (a similar infection of the glans penis) can occur.

2. Diabetic ketoacidosis—A profound loss of insulin activity leads not only to increased serum glucose levels because of increased hepatic glucose output and decreased glucose uptake by insulin-sensitive tissues but also to ketogenesis. In the absence of insulin, lipolysis is stimulated, providing fatty acids that are preferentially converted to ketone bodies in the liver by unopposed glucagon action. Typically, profound hyperglycemia and ketosis (diabetic ketoacidosis) occur in type 1 diabetics, individuals who lack endogenous insulin. However, diabetic ketoacidosis can also occur in type 2 DM, particularly during infections, severe trauma, or other causes of stress that increase levels of counter-regulatory hormones, thus producing a state of profound inhibition of insulin action.

Severe hyperglycemia with glucose levels reaching an average of 500 mg/dL can occur if compensation for the osmotic diuresis associated with hyperglycemia fails. Initially, when elevated glucose levels cause an increase in osmolality, a shift of water from the intracellular to the extracellular space and increased water intake stimulated by thirst help to maintain intravascular volume. If polyuria continues and these compensatory mechanisms cannot keep pace with fluid losses—particularly decreased intake as a result of the nausea and increased losses resulting from the vomiting that accompany ketoacidosis—the depletion of intravascular volume leads to decreased renal blood flow. The kidney’s ability to excrete glucose is, therefore, reduced. Hypovolemia also stimulates counter-regulatory hormones. Therefore, glucose levels rise acutely owing to increased glucose production stimulated by these hormones and decreased clearance by the kidney, an important source of glucose clearance in the absence of insulin-mediated glucose uptake.

In diabetic ketoacidosis, coma occurs in a minority of patients (10%). Hyperosmolality (not acidosis) is the cause of coma. Profound cellular dehydration occurs in response to the marked increase in plasma osmolality. A severe loss of intracellular fluid in the brain leads to coma. Coma occurs when the effective plasma osmolality reaches 330 mOsm/L (normal: 280–295 mOsm/L). Because urea is freely diffusible across cell membranes, blood urea nitrogen is not used to calculate the effective plasma osmolality as:

The increase in ketogenesis caused by a severe lack of insulin action results in increased serum levels of ketones and ketonuria. Insulinopenia is also thought to decrease the ability of tissues to use ketones, thus contributing to the maintenance of ketosis. Acetoacetate and β-hydroxybutyrate, the chief ketone bodies produced by the liver, are organic acids and, therefore, cause metabolic acidosis, decreasing blood pH and serum bicarbonate (Figure 18–8). Respiration is stimulated, which partially compensates for the metabolic acidosis by reducing PCO₂. The presence of unmeasured ketoacid anions in diabetic ketoacidosis (DKA) causes an increased anion gap (the calculated difference between measured cations and anions), which under normal circumstances is primarily due to negatively charged proteins, such as albumin:

Anion Gap (mEq/L) = (Na⁺ + K⁺) − (Cl⁻ + HCO₃⁻)

When the pH level is lower than 7.20, characteristic deep, rapid respirations occur (Kussmaul breathing). Although acetone is a minor product of ketogenesis (Figure 18–8), its fruity odor can be detected on the breath during diabetic ketoacidosis. It should be noted that the ketosis of DKA is much more severe than that appropriately occurring with starvation, because in the latter case, residual insulin action can prevent excessive lipolysis and hepatic ketogenesis while still allowing for peripheral ketone utilization.

Na⁺ is lost in addition to water during the osmotic diuresis accompanying diabetic ketoacidosis. Therefore, total body Na⁺ is depleted. Serum levels of Na⁺ are usually low owing to the osmotic activity of the elevated glucose, which draws water into the extracellular space and in that way decreases the Na⁺ concentration (serum Na⁺ falls approximately 1.6 mmol/L for every 100 mg/dL increase in glucose).

Total body stores of K⁺ are also depleted by diuresis and vomiting. However, acidosis, insulinopenia, and elevated glucose levels cause a shift of K⁺ out of cells, thus maintaining normal or even elevated serum K⁺ levels until acidosis and hyperglycemia are corrected. With administration of insulin and correction of acidosis, serum K⁺ falls as K⁺ moves back into cells. Without treatment, K⁺ can fall to dangerously low levels, leading to potentially lethal cardiac arrhythmias. Therefore, K⁺ supplementation is routinely given in the treatment of diabetic ketoacidosis. Similarly, phosphate depletion accompanies diabetic ketoacidosis, although acidosis and insulinopenia can cause serum phosphorus levels to be normal before treatment. Phosphate replacement is provided only in cases of extreme depletion given the risks of phosphate administration. (Intravenous phosphate may complex with Ca²⁺, resulting in hypocalcemia and Ca²⁺ phosphate deposition in soft tissues.)

Marked hypertriglyceridemia can also accompany diabetic ketoacidosis because of the increased production and decreased clearance of VLDL that occurs in insulin-deficient states. Increased production is due to: (1) the increased hepatic flux of fatty acids, which, in addition to fueling ketogenesis, can be repackaged and secreted as VLDL; (2) increased hepatic VLDL production due to the loss of inhibitory effects of insulin on proteins required for VLDL assembly (apoB and microsomal triglyceride transfer protein [MTP]); and (3) decreased clearance due to decreased lipoprotein lipase activity. Although serum Na⁺ levels can be decreased owing to the osmotic effects of glucose, hypertriglyceridemia can interfere with some common procedures used to measure serum Na⁺. This causes pseudohyponatremia (ie, falsely low serum Na⁺ values, due to overestimation of actual serum volume).

Nausea and vomiting often accompany diabetic ketoacidosis, contributing to further dehydration. Abdominal pain, present in 30% of patients, may be due to gastric stasis and distention. Amylase is frequently elevated (90% of cases), in part because of elevations of salivary amylase, but it is usually not associated with symptoms of pancreatitis. Leukocytosis is frequently present and does not necessarily indicate the presence of infection. However, because infections can precipitate diabetic ketoacidosis in type 1 DM and type 2 DM, other manifestations of infection should be sought, such as fever, a finding that cannot be attributed to diabetic ketoacidosis.

Diabetic ketoacidosis is treated by replacement of water and electrolytes (Na⁺ and K⁺) and administration of insulin. Both treatment modalities are of great importance, as evidenced historically by the marked decrease in mortality from DKA with the advent of insulin therapy (from 100% to 50%) and the further significant decrease (from 50% to 20%) when the importance of hydration was recognized and instituted. With fluid and electrolyte replacement, renal perfusion is increased, restoring renal clearance of elevated blood glucose, and counter- regulatory hormone production is decreased, thus decreasing hepatic glucose production. Insulin administration also corrects hyperglycemia by restoring insulin-sensitive glucose uptake and inhibiting hepatic glucose output. Rehydration is a critical component of the treatment of hyperosmolality. If insulin is administered in the absence of fluid and electrolyte replacement, water will move from the extracellular space back into the cells with correction of hyperglycemia, leading to vascular collapse. Insulin administration is also required to inhibit further lipolysis, thus eliminating substrates for ketogenesis, and to inhibit hepatic ketogenesis, thereby correcting ketoacidosis.

During treatment of diabetic ketoacidosis, measured serum ketones may transiently rise instead of showing a steady decrease. This is an artifact because of the limitations of the nitroprusside test that is often used at the bedside to measure ketones in both serum and urine. Nitroprusside only detects acetoacetate and not β-hydroxybutyrate. During untreated diabetic ketoacidosis, accelerated fatty acid oxidation generates large quantities of NADH in the liver, which favors the formation of β-hydroxybutyrate over acetoacetate (Figure 18–8). With insulin treatment, fatty acid oxidation decreases and the redox potential of the liver shifts back in favor of acetoacetate formation. Therefore, although the absolute amount of hepatic ketone body production is decreasing with treatment of diabetic ketoacidosis, the relative amount of acetoacetate production is increasing, leading to a transient increase in measured serum ketones by the nitroprusside test.

3. Hyperosmolar coma—Severe hyperosmolar states in the absence of ketosis can occur in type 2 DM. These episodes are frequently precipitated by decreased fluid intake such as can occur during an intercurrent illness or in older debilitated patients who lack sufficient access to water and have abnormal renal function hindering the clearance of excessive glucose loads. The mechanisms underlying the development of hyperosmolality and hyperosmolar coma are the same as in diabetic ketoacidosis. However, because only minimal levels of insulin activity are required to suppress lipolysis, these individuals have sufficient insulin to prevent the ketogenesis that results from increased fatty acid flux. Because of the absence of ketoacidosis and its symptoms, patients often present later and, therefore, have more profound hyperglycemia and dehydration; glucose levels often range from 800–2400 mg/dL. Therefore, the effective osmolality exceeds 330 mOsm/L more frequently in these patients than in those presenting with diabetic ketoacidosis, resulting in a higher incidence of coma.

Although ketosis is absent, mild ketonuria can be present if the patient has not been eating. K⁺ losses are less severe than in diabetic ketoacidosis. Treatment is similar to that of diabetic ketoacidosis. Mortality is 10 times higher than in diabetic ketoacidosis because the type 2 diabetics who develop hyperosmolar nonketotic states are older and often have other serious precipitating or complicating illnesses. For example, myocardial infarction can precipitate hyperosmolar states or can result from the alterations in vascular blood flow and other stressors that accompany severe dehydration.

4. Hypoglycemia—Hypoglycemia is a complication of insulin treatment in both type 1 DM and type 2 DM, but it can also occur with oral hypoglycemic drugs that stimulate glucose-independent insulin secretion (eg, sulfonylureas). Hypoglycemia often occurs during exercise or with fasting, states that normally are characterized by slight elevations in counter-regulatory hormones and depressed insulin levels. Under normal circumstances, low insulin levels in these conditions are permissive for the counter-regulatory hormone-mediated mobilization of fuel substrates, increased hepatic glucose output, and inhibition of glucose disposal in insulin-sensitive tissues. In addition, the fall in insulin secretion by the pancreatic β cell in response to low glucose levels is an important stimulus for increased secretion of glucagon. All of these responses would normally restore blood glucose levels. However, in diabetic patients, all of these responses fail when insulin is maintained at excessive levels (relative to plasma glucose) due to excessive exogenous insulin dosing or endogenous glucose-independent insulin stimulation.

The acute response to hypoglycemia is mediated by the counter-regulatory effects of glucagon and catecholamines (Table 18–5). However, the glucagon response can be inadequate in diabetes, increasing the importance of adrenal epinephrine secretion. When counter-regulatory mechanisms fail, initial neurogenic symptoms of hypoglycemia occur secondarily to CNS-mediated sympathoadrenal discharge, resulting in adrenergic (shaking, palpitations, anxiety) and cholinergic (sweating, hunger) responses that encourage carbohydrate-seeking behavior. However, as glucose drops further, neuroglycopenic symptoms also occur from the direct effects of hypoglycemia on CNS function (confusion, coma). A characteristic set of symptoms (night sweats, nightmares, morning headaches) also accompanies hypoglycemic episodes that occur during sleep (nocturnal hypoglycemia).

With symptomatic episodes occurring several times per week, type 1 diabetics are especially prone to hypoglycemia due to a virtually absent glucagon response to hypoglycemia. Moreover, recent episodes of hypoglycemia reduce the adrenal epinephrine response to subsequent hypoglycemia and cause hypoglycemia unawareness by reducing the sympathoadrenal response and associated neurogenic symptoms via unknown mechanisms. This hypoglycemia-induced autonomic failure, which is distinct from diabetic autonomic neuropathy, is reversed by avoidance of hypoglycemia but exacerbated by exercise or sleep, both of which can similarly decrease the sympathoadrenal response to a given level of hypoglycemia.

Acute treatment of hypoglycemia in diabetic individuals consists of rapid oral administration of glucose at the onset of warning symptoms or the administration of exogenous glucagon intramuscularly by others when neuroglycopenic symptoms preclude oral self-treatment. Rebound hyperglycemia can occur after hypoglycemia because of the actions of counter-regulatory hormones (Somogyi phenomenon), an effect that can be aggravated by excessive glucose administration.

Chronic Complications

Over time, diabetes results in damage and dysfunction in multiple organ systems (Table 18–6). Vascular disease is a major cause of most of the sequelae of this disease. Both microvascular disease (retinopathy, nephropathy, neuropathy) that is specific to diabetes and macrovascular disease (coronary artery disease, peripheral vascular disease) that occurs with increased frequency in diabetes contribute to the high morbidity and mortality rates associated with this disease. Neuropathy also causes increased morbidity, particularly by virtue of its role in the pathogenesis of foot ulcers.

Although type 1 DM and type 2 DM both suffer from the complete spectrum of diabetic complications, the incidence varies with each type and with treatment. Macrovascular disease is the major cause of death in type 2 DM. With the advent of intensive glucose control strategies and the use of renin-angiotensin system inhibitors, renal failure secondary to nephropathy is no longer the most common cause of death in individuals with type 1 DM who now, with increased longevity, are increasingly suffering from macrovascular complications. Although blindness occurs in both types, proliferative changes in retinal vessels (proliferative retinopathy) are a major cause of blindness in type 1 DM, whereas macular edema is the most important cause in type 2 DM. Autonomic neuropathy, one of the manifestations of diabetic neuropathy, is more common in type 1 DM.

1. Role of glycemic control in preventing complications—A paradigm shift in diabetes treatment occurred in 1993 with publication of the results of the Diabetes Control and Complications Trial (DCCT), the first major trial to examine the effects of attempted glucose normalization (tight or intensive diabetic control) on the incidence of complications. In this study of individuals with type 1 DM, intensive (vs. conventional) treatment reduced microvascular complications (retinopathy, nephropathy, neuropathy) by 60%. A subsequent study in type 2 DM (United Kingdom Prospective Diabetes Study [UKPDS]) demonstrated a 25% decrease in microvascular complications (retinopathy, nephropathy) with improved glycemic control. In contrast, the role of glycemic control in preventing macrovascular disease, the major cause of death in type 2 DM, is less clear. With the publication in 2008 of three major clinical trials demonstrating either no improvement, or indeed an increase (ACCORD trial), in mortality and macrovascular complications with intensive treatment in type 2 DM, discussions regarding the most appropriate treatment goals (e.g., degree of glucose normalization) and modalities (eg, therapeutics that minimize risk of hypoglycemia and/or weight gain) in type 2 DM continue.

While the importance of glycemic control in influencing the occurrence of microvascular complications is undisputed, genetic factors also clearly play a role. For example, evidence from a variety of studies suggests that approximately 40% of type 1 diabetics are particularly susceptible to the development of severe microvascular complications. This observation suggests that not all individuals with type 1 DM achieve the same benefits from intensive control regimens, which are both inconvenient and associated with an increased risk of hypoglycemia. The identity of genetic factors associated with microvascular disease risk is the subject of ongoing investigations, which have already identified numerous candidate genes coding for the extracellular matrix, transcription factors, growth factor signaling, and/or erythropoietin.

• 2. Microvascular complications—Consistent with clinical evidence defining the critical role of hyperglycemia in microvascular disease, data indicate that high intracellular levels of glucose in cells that cannot down-regulate glucose entry (endothelium, glomeruli, and nerve cells) result in microvascular damage via four distinct, diabetes-specific pathways that were sequentially discovered (Figure 18–9): (1) increased polyol pathway flux, (2) increased formation of advanced glycation end-product (AGE), (3) activation of protein kinase C (PKC), and (4) increased hexosamine pathway flux. More recent information suggests that increased flux through these four pathways is induced by a common factor, overproduction of mitochondrial-derived reactive oxygen species generated by increased flux of glucose through the TCA cycle (Figure 18–9). The end result of these changes in the microvasculature is an increase in protein accumulation in vessel walls, endothelial cell dysfunction, loss of endothelial cells, and, ultimately, occlusion.

The polyol pathway has been extensively studied in diabetic nerve cells and is also present in endothelial cells (Figure 18–9). Many cells contain aldose reductase, an enzyme that converts toxic aldehydes to their respective alcohols (polyol pathway). While aldose reductase has a low affinity for glucose, under conditions of intercellular hyperglycemia, this pathway can account for up to one-third of glucose flux, converting glucose to sorbitol. While excess sorbitol was originally thought to cause osmotic damage, more recent data instead suggest that the real culprit is the consumption of NADPH during glucose reduction. As NADPH is required to regenerate reduced glutathione (GSH), a thiol that detoxifies reactive oxygen species, NADPH consumption prevents the clearance of damaging free radicals. While polyol pathway–mediated damage appears to be a prominent feature in nerve cells, its role in the vasculature is less clear.

The formation of irreversibly glycated proteins called advanced glycosylation end-products (AGEs) also causes microvascular damage in diabetes (Figure 18–9). When present in high concentrations, glucose can react reversibly and nonenzymatically with protein amino groups to form an unstable intermediate, a Schiff base, which then undergoes an internal rearrangement to form a more stable glycated protein, also known as an early glycosylation product (Amadori product) (Figure 18–10). Such a reaction accounts for the formation of glycated HbA, also known as HbA_1c. In diabetics, elevated glucose leads to increased glycation of HbA within red blood cells. Because red blood cells circulate for 120 days, measurement of HbA_1c in diabetic patients serves as an index of glycemic control over the preceding months. Early glycosylation products can undergo a further series of chemical reactions and rearrangements, often involving the formation of reactive carbonyl intermediates, leading to the irreversible formation of AGE. Dicarbonyl formation from direct auto-oxidation of glucose also contributes to AGE formation (Figure 18–10). AGE damage the microvasculature via 3 major pathways: (1) intracellular AGE formation from proteins involved in transcription alters endothelial gene expression; (2) irreversible cross linking of AGE adducts formed from matrix proteins results in vascular thickening and stiffness; and (3) binding of extracellular AGE adducts to AGE receptors (RAGE) on macrophages and endothelium stimulates NF-κB-regulated inflammatory cascades and resultant vascular dysfunction.

Intracellular endothelial hyperglycemia stimulates glycolysis and, with this, an increase in the de novo synthesis of diacylglycerol (DAG) from the glycolytic intermediate, glyceraldehyde-3-phosphate (Figure 18–9). DAG, in turn, activates several isoforms of protein kinase C (PKC) that are present in these cells. This inappropriate activation of PKC alters blood flow and changes endothelial permeability, in part via effects on nitric oxide pathways, and also contributes to thickening of the extracellular matrix.

Last, increased shunting of glucose through the hexosamine pathway via diversion of the glycolytic intermediate, fructose-6-phosphate, is also postulated to play a role in microvascular disease (Figure 18–9). The hexosamine pathway contributes to insulin resistance, producing substrates that, when covalently linked to transcription factors, stimulate the expression of proteins, such as transforming growth factor and plasminogen activator inhibitor, that enhance microvascular damage.

Evidence suggests that all four of these pathways may actually be linked by a common mechanistic element: hyperglycemia-induced oxidative stress. In particular, the increase in electron donors that results from shunting glucose through the tricarboxylic acid cycle increases mitochondrial membrane potential by pumping proteins across the mitochondrial inner membrane. This increased potential prolongs the half-life of superoxide generating enzymes, thus increasing the conversion of O₂ to O₂^–. These increased reactive oxygen species lead to inhibition of the glycolytic enzyme, glyceraldehyde-3-phosphate dehydrogenase (GADPH), and a resultant increase in upstream metabolites that can now be preferentially diverted into the four mechanistic pathways (Figure 18–9).

a. Retinopathy—Diabetes is a leading cause of blindness in developed countries (vs. untreated cataracts in developing nations). Diabetic retinopathy, present after 20 years in more than 95% with type 1 DM and 60% with type 2 DM, occurs in two distinct stages: nonproliferative and proliferative.

Nonproliferative retinopathy has a prevalence of 30% in adults with diabetes in the United States, occurs frequently in both type 1 DM and type 2 DM, and is already present at the time of diagnosis in more than 20% of individuals with type 2 DM. Microaneurysms of the retinal capillaries, appearing as tiny red dots, are the earliest clinically detectable sign of diabetic retinopathy (background retinopathy). These outpouchings in the capillary wall are due to loss of surrounding pericytes that support the capillary walls. Vascular permeability is increased. Fat that has leaked from excessively permeable capillary walls appears as shiny yellow spots with distinct borders (hard exudates) forming a ring around the area of leakage. The appearance of hard exudates in the area of the macula is often associated with macular edema, which is the most common cause of blindness in type 2 DM, occurring in 7% of diabetics. As retinopathy progresses, signs of ischemia appearing as background retinopathy worsen (preproliferative stage). Occlusion of capillaries and terminal arterioles causes areas of retinal ischemia that appear as hazy yellow areas with indistinct borders (cotton wool spots or soft exudates) because of the accumulation of axonoplasmic debris at areas of infarction. Retinal hemorrhages can also occur, and retinal veins develop segmental dilation.

Retinopathy can progress to a second, more severe stage characterized by the proliferation of new vessels (proliferative retinopathy). Neovascularization is more prevalent in type 1 DM than in type 2 DM (25% vs. 15% after 20 years) and is a leading cause of blindness in type 1 DM. It is hypothesized that retinal ischemia stimulates the release of growth-promoting factors, resulting in new vessel formation. However, these capillaries are abnormal, and traction between new fibrovascular networks and the vitreous can lead to vitreous hemorrhage or retinal detachment, two potential causes of blindness.

b. Nephropathy—Diabetes is the most common cause of end-stage renal disease (ESRD) worldwide. Although ESRD occurs more frequently in type 1 DM than in type 2 DM (35% vs. 20% after 20 years), type 2 DM accounts for more than half of the diabetic population with ESRD because of its greater prevalence. ESRD also occurs more frequently in Native Americans, African Americans, and Hispanic Americans than in non-Hispanic whites with type 2 DM.

Diabetic nephropathy results primarily from disordered glomerular function. Histologic changes in glomeruli are indistinguishable in type 1 DM and type 2 DM and occur to some degree in the majority of individuals. Basement membranes of the glomerular capillaries are thickened and can obliterate the vessels; the mesangium surrounding the glomerular vessels is increased owing to the deposition of basement membrane-like material and can encroach on the glomerular vessels; and the afferent and efferent glomerular arteries are also sclerosed. Glomerulosclerosis is usually diffuse but in 50% of cases is associated with nodular sclerosis. This nodular component, called Kimmelstiel-Wilson nodules after the investigators who first described the pathologic changes in diabetic kidneys, is pathognomonic for diabetes but is present in only 30% of patients with microalbuminuria.

In type 1 DM patients, glomerular changes are preceded by a phase of hyperfiltration resulting from vasodilation of both the afferent and efferent glomerular arterioles, an effect perhaps mediated by two of the counter-regulatory hormones, glucagon and growth hormone, or by hyperglycemia. It is unclear whether this early hyperfiltration phase occurs in type 2 DM. It has been proposed that the presence of atherosclerotic lesions in older type 2 DM patients may prevent hyperfiltration and thus account for the lower incidence of overt clinical nephropathy in these individuals.

Early in the course of the disease, the histologic changes in renal glomeruli are accompanied by microalbuminuria, a urinary loss of albumin that cannot be detected by routine urinalysis dipstick methods (Figure 18–11). Albuminuria is thought to be due to a decrease in the heparan sulfate content of the thickened glomerular capillary basement membrane. Heparan sulfate, a negatively charged proteoglycan, can inhibit the filtration of other negatively charged proteins, such as albumin, through the basement membrane; its loss, therefore, allows for increased albumin filtration.

If glomerular lesions worsen, proteinuria increases and overt nephropathy develops (Figure 18–11). Diabetic nephropathy is defined clinically by the presence of more than 300 mg of urinary protein per day, an amount that can be detected by routine urinalysis. In diabetic nephropathy (unlike other renal diseases), proteinuria continues to increase as renal function decreases. Therefore, end-stage renal disease is preceded by massive, nephrotic-range proteinuria (>4 g/d). The presence of hypertension speeds this process. Although type 2 DM patients often already have hypertension at the time of diagnosis, type 1 DM patients usually do not develop hypertension until after the onset of nephropathy. In both cases, hypertension worsens as renal function deteriorates. Therefore, control of hypertension is critical in preventing the progression of diabetic nephropathy.

Retinopathy, a process that is also worsened by the presence of hypertension, usually precedes the development of nephropathy. Therefore, other causes of proteinuria should be considered in diabetic individuals who present with proteinuria in the absence of retinopathy.

c. Neuropathy—Neuropathy (Table 18–6) occurs commonly in about 60% of both type 1 DM and type 2 DM patients and is a major cause of morbidity. Diabetic neuropathy can be divided into three major types: (1) a distal, primarily sensory, symmetric polyneuropathy that is by far the most common (50% incidence); (2) an autonomic neuropathy, occurring frequently in individuals with distal polyneuropathy (>20% incidence); and (3) much less common, transient asymmetric neuropathies involving specific nerves, nerve roots, or plexuses.

Symmetric distal polyneuropathy—Demyelination of peripheral nerves, which is a hallmark of diabetic polyneuropathy, affects distal nerves preferentially and is usually manifested clinically by a symmetric sensory loss in the distal lower extremities (stocking distribution) that is preceded by numbness, tingling, and paresthesias. These symptoms, which begin distally and move proximally, can also occur in the hands (glove distribution). Pathologic features of affected peripheral somatic nerves include demyelination and loss of nerve fibers with reduced axonal regeneration accompanied by microvascular lesions, including thickening of basement membranes. Activation of the polyol pathway in nerve cells is thought to play a major role in inducing symmetric distal polyneuropathy in diabetes. In addition, the microvascular disease that accompanies these neural lesions may also contribute to nerve damage. The presence of antibodies to autoantigens in patients with neuropathy also suggests a possible immune component to this disorder. Last, defects in the production or delivery of neurotrophic factors, such as nerve growth factor (NGF), are hypothesized to play a role in the pathogenesis of symmetric distal neuropathy.

Autonomic neuropathy—Autonomic neuropathy often accompanies symmetric peripheral neuropathy, occurs more frequently in type 1 DM, and can affect all aspects of autonomic functioning, most notably those involving the cardiovascular, genitourinary, and GI systems. Less information is available regarding the morphologic changes occurring in affected autonomic nerves, but similarities to somatic nerve alterations suggest a common pathogenesis.

Fixed, resting tachycardia and orthostatic hypotension are signs of cardiovascular autonomic nervous system damage that can be easily ascertained on physical examination. Orthostatic hypotension can be quite severe. Erectile dysfunction occurs in more than 50% of diabetic men and is due both to neurogenic (parasympathetic control of penile vasodilation) and vascular factors. Sexual dysfunction in diabetic women has not been well studied. Loss of bladder sensation and difficulty emptying the bladder (neurogenic bladder) lead to overflow incontinence and an increased risk of urinary tract infections as a result of residual urine. Motor disturbances can occur throughout the GI tract, resulting in delayed gastric emptying (gastroparesis), constipation, or diarrhea. Anhidrosis in the lower extremities can lead to excessive sweating in the upper body as a means of dissipating heat, including increased sweating in response to eating (gustatory sweating). Autonomic neuropathy can also result in decreased glucagon and epinephrine responses to hypoglycemia.

Mononeuropathy and mononeuropathy multiplex—The abrupt, usually painful onset of motor loss in isolated cranial or peripheral nerves (mononeuropathy) or in multiple isolated nerves (mononeuropathy multiplex) occurs much less frequently than does symmetric polyneuropathy or autonomic neuropathy. Vascular occlusion and ischemia are thought to play a central role in the pathogenesis of these asymmetric focal neuropathies, which are usually of limited duration and occur more frequently in type 2 DM. The third cranial nerve is the most frequently involved, causing ipsilateral headache followed by ptosis and ophthalmoplegia with sparing of papillary reactivity. In contrast to the rare occurrence of these vascular neuropathies, symptomatic compression by entrapment of peripheral nerves (eg, ulnar nerve at elbow; median nerve at the wrist) occurs in 30% of diabetics and usually involves both the nerve and surrounding tissues.

3. Macrovascular complications—Atherosclerotic macrovascular disease occurs with increased frequency in diabetes, resulting in an increased incidence of myocardial infarction, stroke, and claudication and gangrene of the lower extremities. Although macrovascular disease accounts for significant morbidity and mortality in both types of diabetes, the effects of large-vessel disease are particularly devastating in type 2 DM and are responsible for approximately 75% of deaths. The protective effect of gender is lost in women with diabetes; their risk of atherosclerosis is equal to that of men (Figure 18–12).

Reasons for the increased risk of atherosclerosis in diabetes are threefold: (1) the incidence of traditional risk factors, such as hypertension and hyperlipidemia, is increased (50% and 30% incidence at diagnosis, respectively); (2) diabetes itself (likely due to both hyperglycemia and insulin resistance) is an independent risk factor for atherosclerosis; and (3) diabetes appears to synergize with other known risk factors to increase atherosclerosis. The elimination of other risk factors, therefore, can greatly reduce the risk of atherosclerosis in diabetes (Figure 18–12).

Hypertension associated with increased total body extracellular Na⁺ content and volume expansion occurs with increased frequency in type 1 DM and type 2 DM and is responsive to targeted inhibition of the renin-angiotensin system. Despite these similar findings, the epidemiology of hypertension in the two subtypes suggests that different pathophysiologic mechanisms may be operative. In type 1 DM, hypertension usually occurs after the onset of nephropathy (40% incidence after 40 years of type 1 DM), when renal insufficiency impairs the ability to excrete water and solutes. In type 2 DM, hypertension is often already present at the time of diagnosis (70% are hypertensive) in these older, obese, insulin-resistant individuals. Indeed, it has been proposed that insulin resistance plays a central role in both diabetes and hypertension. For example, insulin resistance is associated with activation of the renin-angiotensin system, which leads to hypertension, while renin-angiotensin system activation, in turn, decreases insulin sensitivity.

In contrast to its central role in microvascular disease, the importance of hyperglycemia as a risk factor for macrovascular disease, which occurs in 40% of 40-year-old individuals with type 1 DM (vs. <10% of controls), remains uncertain. However, insulin resistance, a hallmark of type 2 DM that can also develop in response to hyperglycemia in type 1 or type 2 DM, is clearly an important driver of macrovascular complications in diabetes. Insulin resistance is central to the pathogenesis of two obesity-associated syndromes: (1) prediabetes (impaired fasting glucose or glucose tolerance) and (2) metabolic syndrome (a cluster of metabolic abnormalities, including central obesity, elevated glucose, elevated blood pressure, elevated triglycerides, and low high-density lipoprotein [HDL] cholesterol). Both of these syndromes are associated with increased cardiovascular risk, as well as an increased risk for later development of diabetes. At present in the United States, one-third of the adult population is thought to fall into these high-risk categories. Fortunately, an important clinical trial (Diabetes Prevention Program) has demonstrated that significant risk reductions occur in response to lifestyle interventions in this population.

In addition to being a component of metabolic syndrome, hypertriglyceridemia, which is associated with increased risk of cardiovascular disease, is the principal lipid abnormality in poorly controlled type 1 and type 2 DM. Very-low-density lipoprotein-triglyceride (VLDL-TG) levels are increased because of insufficient insulin action in liver and adipose tissue. This results in (1) increased VLDL production due to increased flux of fatty acids from adipose tissue to the liver (ie, increased lipolysis) and loss of insulin suppression of hepatic proteins required for VLDL assembly (ie, loss of phosphoinositide 3-kinase [PI 3-kinase] inhibition of apolipoprotein B [apoB] production and loss of inhibition of the transcription factor FoxO1-induced expression of microsomal triglyceride transfer protein [MTP]); and (2) decreased VLDL clearance as a result of decreased lipoprotein lipase activity. Excessive VLDL levels alter the composition of LDL and HDL, transferring triglycerides to these particles while depleting them of cholesterol, creating small dense LDL particles and low HDL-cholesterol levels, both of which are independent risk factors for cardiovascular disease. LDL cholesterol may also be elevated both because of increased production (VLDL is catabolized to LDL) and decreased clearance (insulin deficiency may reduce LDL receptor activity). Insulin treatment usually corrects lipoprotein abnormalities in type 1 DM. In contrast, treatment of hyperglycemia often does not normalize lipid profiles in obese, insulin-resistant individuals with type 2 DM unless accompanied by weight reduction (ie, by a concomitant reduction in insulin resistance).

Possible reasons why diabetes may be an independent risk factor for atherosclerosis and may also act synergistically with other risk factors include the following: (1) alterations in lipoprotein composition in diabetes that make the particles more atherogenic (eg, increased small dense LDL, increased levels of Lp[a], enhanced oxidation and glycation of lipoproteins); (2) the occurrence of a relative procoagulant state in diabetes, including an increase in certain clotting factors and increased platelet aggregation; (3) proatherogenic alterations in the vessel walls caused either by the direct effects of hyperinsulinemia in type 2 DM or by boluses of exogenously administered insulin (vs. hepatic first-pass clearance of endogenously secreted insulin) in type 1 DM, which include promotion of smooth muscle proliferation, alteration of vasomotor tone, and enhancement of foam cell formation (cholesterol-laden cells that characterize atherogenic lesions); (4) proatherogenic alterations in the vessel walls caused by the direct effects of hyperglycemia, including deposition of glycated proteins, just as occurs in the microvasculature; and, importantly, (5) the proinflammatory milieu that is associated with insulin resistance.

• 4. Diabetic foot ulcers—Diabetic foot ulcers occur in 10% of diabetics, can be complicated by osteomyelitis, and result in amputation in 1%, an event that is associated with high mortality (50% by 3 years). Risk factors for ulcer development include (1) increased injuries in insensate feet due to symmetric polyneuropathy (present in 75–90% of diabetics with foot ulcers), which can be detected clinically by decreased vibratory and cutaneous pressure sensation and absence of ankle reflexes; (2) macrovascular disease (present in 30–40% with foot ulcers) and microvascular disease; (3) infections caused by alterations in neutrophil function and vascular insufficiency; and (4) faulty wound healing caused by unknown factors.

• 5. Infection—Neutrophil chemotaxis and phagocytosis are defective in poorly controlled diabetes. Cell-mediated immunity may also be abnormal. In addition, vascular lesions can hinder blood flow, preventing inflammatory cells from reaching wounds (eg, foot ulcers) or other possible sites of infection. Therefore, individuals with diabetes are more prone to develop infections and may have more severe infections. As a result, certain common infections (eg, candidal infections, periodontal disease) occur more frequently in diabetics. A number of unusual infections also are seen in diabetics (ie, necrotizing papillitis, mucormycosis of the nasal sinuses invading the orbit and cranium, and malignant otitis externa caused by Pseudomonas aeruginosa).

• 6. Skeletal changes in diabetes—Children with type 1 DM have a much lower bone mass, attributed to loss of the anabolic effects of insulin on bone that stimulate the differentiation of bone-forming osteoblasts, and an associated increased in fragility bone fractures. Adults with type 2 DM have an increased fracture risk, perhaps due to subtle microarchitectural changes (eg, increased cortical porosity) since bone mineral density in these typically obese individuals is normal or increased. Emerging evidence suggests that the interactions between carbohydrate homeostasis and the skeleton are bidirectional. Osteocalcin, an insulin-inducible protein secreted by osteoblasts, increases β-cell mass and insulin secretion while also improving adipose insulin sensitivity by stimulating adiponectin expression. The role of osteocalcin in diabetes pathogenesis and treatment is an area of active investigation.

Neuroendocrine Islet Cell Tumors of the Pancreas

While highly prevalent in individuals with multiple endocrine neoplasia type 1 (MEN-1), neuroendocrine tumors arising from the islet cells are otherwise infrequent and account for only 5% of primary pancreatic neoplasms, most of which instead arise from cells of the exocrine pancreas. However, the clinical manifestations associated with islet cell tumor overproduction of a given hormone are illustrative of their normal physiologic functions (Table 18–7). Those tumors associated with inappropriate secretion of hormones regulating carbohydrate metabolism (insulin, glucagon, somatostatin) are highlighted here.

Insulinoma (β-cell tumor)

Clinical Presentation

The occurrence of fasting hypoglycemia in an otherwise healthy individual is usually due to an insulin-secreting tumor of the β cells of the islets of Langerhans (insulinoma; Table 18–7). Although insulinoma is the most common islet cell tumor, it is still a rare disorder. Insulinomas occur most frequently in the fourth to seventh decades, although they can occur earlier, particularly when associated with MEN-1, a neoplastic syndrome characterized by tumors of the parathyroids, pituitary, and endocrine pancreas (see Chapter 17). The diagnosis of hypoglycemia is based on the Whipple triad: (1) symptoms and signs of hypoglycemia, (2) an associated low plasma glucose level, and (3) reversibility of symptoms on administration of glucose.

Etiology

In the great majority of cases, insulinomas are benign solitary lesions composed of whorls of insulin-secreting β cells. Multiple tumors, although infrequent (<10%), are seen most often in patients with MEN-1. Fewer than 10% of the tumors are malignant, as determined by the presence of metastases.

Pathology and Pathogenesis

Inappropriately high levels of insulin in situations normally characterized by a lowering of insulin secretion (eg, fasting and exercise) result in hypoglycemia. Normally, in the postabsorptive and fasting state, insulin levels decline, leading to an increase in glucagon-stimulated hepatic glucose output and a decrease in insulin-mediated glucose disposal in the periphery, which maintains normal serum glucose levels. With exercise, low insulin allows muscles to use glycogen, glucagon, and other counter-regulatory hormones to increase hepatic glucose output and counter-regulatory hormones to mobilize fatty acids for ketogenesis and fatty acid oxidation by muscle. With an insulinoma, insulin levels remain high during fasting or exercise. In this circumstance, glucagon-mediated hepatic glucose output is suppressed while insulin-mediated peripheral glucose uptake continues, and insulin stimulates hepatic fatty acid synthesis and peripheral fatty acid storage while suppressing fatty acid mobilization and hepatic ketogenesis. The result is fasting or exercise-induced hypoglycemia in the absence of ketosis.

Clinical Manifestations

Individuals with insulinomas often are symptomatic for years before diagnosis and are self-treated with frequent food intake. Not all patients experience fasting hypoglycemia in the morning (only 30% of insulinoma patients develop hypoglycemia after a diagnostic 12-hour fast). Often they experience late afternoon hypoglycemia, particularly when precipitated by exercise. Because alcohol, like insulin, inhibits gluconeogenesis, alcohol ingestion can also precipitate symptoms. A high percentage of individuals with insulinoma experience neuroglycopenic as well as autonomic symptoms (Table 18–5). Confusion (80%), loss of consciousness (50%), and seizures (10%) often lead to misdiagnoses of psychiatric or neurologic disorders.

Fasting hypoglycemia can be due either to elevated insulin, as occurs in insulinoma, or to non–insulin-mediated effects such as loss of counter-regulatory hormones (eg, loss of cortisol in Addison disease), severe hepatic damage that prevents hepatic glucose production, loss of peripheral stores of substrates for hepatic glucose production (eg, cachexia), or some states of markedly increased glucose utilization (eg, sepsis, cancer). To distinguish insulin-mediated from non–insulin-mediated fasting hypoglycemia, patients suspected of having insulinoma are subjected to a diagnostic fast during which glucose, insulin, and C peptide levels are measured. An inappropriately elevated insulin level in the setting of hypoglycemia is diagnostic of an insulin-mediated cause of hypoglycemia. Causes of insulin-mediated hypoglycemia other than insulinoma include surreptitious injection of insulin, ingestion of oral hypoglycemic medications that stimulate glucose-independent endogenous insulin (sulfonylureas), and the presence of insulin antibodies. Binding of insulin to the antibodies prevents insulin action, but release of the insulin at an inappropriate time can result in hypoglycemia. Surreptitious insulin administration can be ruled out by C peptide measurements. Because insulin and C peptide are cosecreted, insulinomas will cause elevations in both, whereas elevated levels of exogenous insulin will not be matched by elevations of C peptide in surreptitious injections of insulin. Similarly, insulin antibodies do not result in elevated C peptide levels. Because sulfonylurea drugs stimulate endogenous insulin (and, therefore, C peptide) secretion, insulinoma and inappropriate ingestion of these agents can only be differentiated by measuring drug levels.

Glucagonoma (α-cell tumor)

Glucagonomas are usually diagnosed by the appearance of a characteristic rash in middle-aged individuals, particularly perimenopausal women, with mild diabetes mellitus (Table 18–7). Glucagon levels are usually increased 10-fold relative to normal values but can even be increased 100-fold.

Necrolytic migratory erythema begins as an erythematous rash on the face, abdomen, perineum, or lower extremities. After induration with central blistering develops, the lesions crust over and then resolve, leaving an area of hyperpigmentation. These lesions may be the result of nutritional deficiency, such as the hypoaminoacidemia that occurs from excessive glucagon stimulation of hepatic amino acid uptake and utilization as fuel for gluconeogenesis, rather than the direct effect of glucagon on the skin. Appearance of the rash is a late manifestation of the disease.

Diabetes mellitus or glucose intolerance is present in the vast majority of patients as a result of increased stimulation of hepatic glucose output by the inappropriately high glucagon levels. Insulin levels are secondarily increased. Diabetes is, therefore, mild and is not accompanied by glucagon-stimulated ketosis, because sufficient insulin is present to suppress lipolysis, thus limiting potential substrates for ketogenesis.

Anemia and a variety of nonspecific GI symptoms related to decreased intestinal motility also can accompany glucagonomas.

Although these tumors are solitary and their growth is slow, they are usually large and have often metastasized by the time of diagnosis, making surgical resection difficult. Octreotide, the synthetic somatostatin analog, can be used to ameliorate symptoms via its suppression of glucagon secretion.

Somatostatinoma (δ-cell tumor)

Somatostatinomas present with a variety of GI symptoms in individuals with mild diabetes (Table 18–7). However, these extremely rare tumors are almost uniformly found incidentally during operations for cholelithiasis or other abdominal complaints because the presenting symptoms are both nonspecific and common in an adult population. Documentation of elevated somatostatin levels confirms the diagnosis.

A classic triad of symptoms frequently occurs with excessive somatostatin secretion: diabetes mellitus, because of its inhibition of insulin and glucagon secretion; cholelithiasis, because of its inhibition of gall-bladder motility; and steatorrhea, because of its inhibition of pancreatic exocrine function. Hypochlorhydria, diarrhea, and anemia can also occur.

In type 1 DM and type 2 DM, the effects of insulin insufficiency are aggravated by the occurrence of elevated glucagon levels. In contrast, with somatostatinomas, both insulin and glucagon are suppressed. Therefore, the hyperglycemia resulting from insulinopenia is tempered by the absence of glucagon stimulation of hepatic glucose output. Although low insulin levels are permissive for lipolysis, glucagon deficiency prevents hepatic ketogenesis. The diabetes associated with somatostatinomas is, therefore, mild and not ketosis prone.

Although the majority of somatostatinomas occur in the pancreas, a significant number are found in the duodenum or jejunum. Like glucagonomas, somatostatinomas are often solitary and large and have frequently metastasized by the time of diagnosis.

Yeong Kwok, MD

Case 90

Diabetes Mellitus: Diabetic Ketoacidosis

A 58-year-old homeless man with long-standing insulin-treated type 2 diabetes has been diagnosed with right lower extremity cellulitis. He has taken a prescribed oral antibiotic for the past week but has not noticed much improvement. For the past 2 days, he has complained of intermittent fevers and chills, nausea with poor oral intake, and proximally spreading erythema over his right leg. On the evening of admission, a friend notices that he is markedly confused and calls 911. In the emergency room, he is oriented only to his name. The patient is tachypneic, breathing deeply at a rate of 24/min. He is febrile at 38.8°C. He is normotensive, but his heart rate is elevated at 112 bpm.

On examination, this patient is a delirious, unkempt man with a fruity breath odor. His right lower extremity is markedly erythematous and exquisitely tender to palpation. Serum chemistries reveal a glucose level of 488 mg/dL, potassium of 3.7 mEq/dL, and sodium of 132 mEq/L. Urine dipstick is grossly positive for ketones.

Questions

A. Describe the precipitants of ketoacidosis in this diabetic patient.

Ketoacidosis is caused by a severe lack of insulin seen most commonly in patients with type 1 diabetes mellitus. It may be the initial presentation of this disorder. However, in this patient with a long-standing history of type 2 diabetes and resultant insulin resistance and true insulinopenia, ketosis was precipitated by acute infection. In this case, severe cellulitis induced counterregulatory hormone production, which inhibits insulin’s action. Thus, in the effective absence of insulin, lipolysis generates fatty acids that are preferentially converted to ketone bodies by the liver, resulting in ketoacidosis.

B. What is the cause of his altered mental status?

Altered mental status in diabetic ketoacidosis, as in hyperosmolar coma, most closely correlates with the degree of hyperosmolality induced by hyperglycemia and the associated osmotic diuresis. Profound intracellular dehydration is seen in the brain as fluid shifts in response to elevated plasma osmolality. The effective osmolality in this patient is calculated as follows: 2(132 + 3.7) + 488/18 = 298.5. Coma occurs when the effective plasma osmolality reaches 340 mOsm/L. Although alterations in mental status can occur as plasma osmolality rises above the upper limit of normal (295 mOsm/L), patients usually do not exhibit anything more than mild to moderate drowsiness at the level of osmolarity seen in this patient. Therefore, other possible causes of altered mental status should be considered, including stroke, infection, and drugs.

C. Describe the patient’s respiratory pattern. What is the pathogenetic mechanism?

This patient exhibits Kussmaul breathing (hyperpnea that effectively drops the PCO₂ to partially compensate for the underlying metabolic acidosis). This respiratory pattern is commonly seen with a blood pH less than 7.20. In addition, the fruity odor detected on his breath is due to the keto acid acetone produced in this disorder.

D. What are important issues to consider in replacing electrolytes in this patient?

Mainstays of treatment for diabetic ketoacidosis include concomitant insulin therapy along with free water and electrolyte replacement. The osmotic diuresis results in significant free water loss and total body potassium depletion. However, the serum potassium appears normal because of the shift of K⁺ out of cells and into the extracellular space—induced by acidosis, hyperglycemia, and insulinopenia. Correction of the acidosis and hyperglycemia with insulin therapy shifts potassium back into cells. Unless carefully monitored and replete, serum K⁺ levels can drop dangerously low, leading to potentially fatal cardiac arrhythmias. Phosphate depletion can also be seen, but replacement is considered only in severe cases because of the risks of intravenous phosphate repletion.

Case 91

Insulinoma

A 61-year-old man recently moved to town and is establishing primary care. During a comprehensive review of systems, he reports that he has experienced a 3-year history of “hypoglycemic attacks.” These short periods of light-headedness, confusion, palpitations, and tremor occur more frequently in the late afternoon while jogging. His symptoms are relieved after drinking a sugared sports drink. He has no history of diabetes or cancer. His physical examination is unremarkable, and in the clinic a fasting morning glucose level is 93 mg/dL. Suspecting that an insulinoma-induced hypoglycemic state may be responsible for his symptoms, his physician requests a diagnostic fast period during which glucose, insulin, and C peptide levels are measured.

Questions

A. Describe the Whipple triad in the diagnosis of hypoglycemia.

The Whipple triad sets forth the diagnostic criteria for hypoglycemia: (1) symptoms and signs of hypoglycemia, (2) an associated low plasma glucose level, and (3) improvement in symptoms with the administration of glucose. This patient’s self-diagnosis of hypoglycemic attacks meets these criteria.

B. What patient history clues suggest insulinoma? Discuss the pathogenesis.

The patient’s age and the fasting hypoglycemia are suggestive of insulinoma, an insulin-secreting tumor of the B cells of the islets of Langerhans. Normally during exercise, insulin levels decline, allowing for significant glycogen uptake in the periphery. In addition, glucagon-stimulated hepatic glucose output increases so as to maintain adequate serum glucose levels, and counterregulatory hormones mobilize fatty acids for ketogenesis and fatty acid oxidation by muscle. However, during exercise, an elevated insulin level secreted by an insulinoma suppresses the glucagon-mediated glucose output while insulin-induced peripheral glucose uptake continues. Thus, the patient becomes hypoglycemic and his symptoms recur.

C. How might the tests ordered help identify the cause of the hypoglycemia?

Hypoglycemia in the setting of an elevated serum insulin level essentially rules out examples of non–insulin-mediated causes of hypoglycemia such as Addison disease, sepsis, and severe hepatic injury. The differential diagnosis of insulin-mediated hypoglycemia includes surreptitious insulin injection, oral hypoglycemic use (stimulating endogenous insulin production), and the presence of insulin antibodies. In this patient, a C peptide measurement was elevated, suggesting that this was not due to surreptitious injections or to antibodies. A greater challenge is to distinguish insulinoma from oral hypoglycemic use, both of which show an elevated C peptide and, therefore, require the direct measurement of serum levels of oral hypoglycemic agents to confirm the latter diagnosis.

Case 92

Glucagonoma

A 52-year-old woman with a 3-year history of diet-controlled diabetes presents to her primary care provider complaining of a “stubborn poison ivy rash” over her legs, which she attributed to a possible exposure to the plant during a recent hike. She presented twice to the urgent care center and received high-potency topical steroid cream for this refractory erythematous rash with central blistering. A review of systems reveals intermittent diarrhea and constipation as well as weight loss. Her serum glucagon level is measured and found to be 20 times normal.

Questions

A. What is the rash found in this disease? What is thought to be the cause?

Necrolytic migratory erythema is typically a late manifestation of glucagonoma and may be the result of hypoaminoacidemia stemming from excessive glucagon-mediated hepatic uptake of amino acids. This nutritional deficiency, rather than a direct effect of glucagon itself, is linked to the dermatologic manifestations.

B. Why is the diabetes found in this condition usually mild?

Diabetes or glucose intolerance is usually mild, seen in response to hyperstimulation of hepatic glucose output by supranormal glucagon levels. Subsequently, serum insulin is increased, which prevents lipolysis and an associated ketotic state.

C. What is the typical prognosis?

Glucagonomas are usually malignant, and weight loss and liver metastasis are commonly seen at the time of diagnosis; surgical resection is rarely pursued. Once diagnosed, the median survival is typically less than 3 years.

Case 93

Somatostatinoma

At the time of an elective laparoscopic cholecystectomy for gallstones, a 44-year-old woman with mild diabetes mellitus and chronic diarrhea is noted to have a 3–4 cm solitary mass on the surface of her duodenum. Omental lymphadenopathy is seen. Biopsy demonstrates a high-grade somatostatinoma with lymph node metastasis.

Questions

A. Describe the triad of signs typically seen in patients with somatostatinomas.

Somatostatinomas are very rare tumors, typically associated with a triad of findings, including diabetes, steatorrhea, and cholelithiasis. The latter finding is thought to be due to somatostatin-induced gallbladder hypomotility.

B. Why is the diabetes found in this condition usually mild?

Because somatostatin suppresses both insulin and glucagon secretion, the resulting hyperglycemic state is mild and not accompanied by glucagon-mediated hepatic ketogenesis.