Chapter 2:  The endocrine pancreas

Glucose turnover

Carbohydrates exist as polysaccharides (such as starch), disaccharides (such as sucrose, maltose and lactose) and monosaccharides (such as galactose, glucose and fructose). Glucose absorbed from the gut is mainly derived from starch that, in Western societies, constitutes about 60% of the daily carbohydrate intake. The rest is in the form of sucrose (30%) and lactose (10%). In starch, straight chains of glucose molecules are held in amylose (approximately 20% of the total starch) whilst branched chains of glucose molecules are held in amylopectin (80% of the total starch). In the gut, these large molecules are broken down by digestion but the polarity of the hexoses requires that absorption across the hydrophobic cell membrane of the gut involves specialized transport proteins (Box 2.3). Such is the importance of glucose, there are five of these glucose transporters (GLUTs) for the absorption and uptake of glucose into cells. These have distinct tissue distributions and features (Box 2.4).

After a meal, some glucose absorbed from the gut is used immediately by cells as a source of energy. Most is either stored in muscle and liver as glycogen or converted to triglyceride in adipocytes. An overview of the uses of glucose is shown in Box 2.5.

In skeletal muscle, absorbed glucose is directly converted to glycogen via glucose 6-phosphate. Whilst the same direct process occurs in the liver, approximately two thirds of the liver glycogen is formed from an indirect pathway. Glucose absorbed in the diet is first converted to lactate (a process termed glycolysis) in non-hepatic tissue (for example, the gut) and, in the liver, is converted to glucose 6-phosphate (via pyruvate), a process known as gluconeogenesis. In the liver, this may also occur from some amino acids (Box 2.5).

The breakdown of glycogen to glucose is known as glycogenolysis (Box 2.5). In the liver pyruvate can be converted to acetyl co-enzyme A (acetyl-CoA) that is not only important in the tricarboxylic acid (TCA) cycle but can also be used to form ketones or fatty acids. In addition, pyruvate can be transaminated to form the amino acid, alanine.

Circulating concentrations of lactate are usually about 1 mmol/l. In conditions such as extreme exercise or when cardiac output is low, glucose is metabolized to lactate. Lactic acid has a pK_a of 3.86 (i.e. in aqueous solution 50% of lactate is ionized at pH 3.86). At a physiological pH of about 7.4, it is completely dissociated leading to the generation of H⁺ ions and to acidosis if large amounts of lactate are produced and the blood buffering capacity is exceeded (resulting in lactic acidosis).

Anabolic and catabolic phases of glucose metabolism

The metabolic use of glucose depends on the nutritional situation existing at any time. Before breakfast after a 10 h fast (termed the post-absorptive state), the metabolic situation is fairly stable. Depending on prior nutritional state, the glucose stores (glycogen) in the liver may have been partly used and glucose production comes from a combination of hepatic gluconeogenesis and glycogenolysis. Glucose production and utilization are approximately equal at about 12 μmol/kg body weight/min. After breakfast, the blood glucose concentration rises. The magnitude of the increase depends not only on the type of carbohydrate ingested (i.e. its glycemic index) but also on the rates of digestion and absorption. The portal vein glucose concentration rises from, say, 4 mmol/l to around 10 mmol/l and the liver takes up some of this glucose. A smaller fraction is taken up by muscle. Thus, in the anabolic phase of glucose metabolism, that fraction of glucose not directly utilized by cells is stored in the liver or muscle in the form of glycogen. In the catabolic phase (i.e. post-absorptive state), glucose is first obtained from glycogen stores (glycogenolysis) and then by gluconeogenesis. Insulin has anabolic actions on glucose metabolism, glucagon has catabolic actions (Box 2.6).

Knowing the average daily intake of carbohydrates, the volume of extracellular fluid and the mean glucose concentration of 5 mmol/l, it can be calculated that the glucose content of the entire extracellular fluid is replaced 25 times daily, i.e. approximately once an hour. Since the venous plasma glucose concentration may only rise by a few mmol/l (from about 4 mmol/l to, say, 6 mmol/l), even after a large meal, it is clear that blood glucose concentrations are tightly regulated.

The liver is able to buffer changes in blood glucose concentrations after a meal because it sits at the head of the portal vein and has a large capacity to take up glucose. There are four reasons for this. Firstly, hepatic cells express the high affinity GLUT 2 transporter protein (Box 2.4). GLUT 2 transporters are not insulin-responsive and they are present at high density. Secondly, glucokinase, which phosphorylates imported glucose, has a high K_m for glucose and does not undergo product inhibition i.e. glucokinase is not inhibited by its product glucose-6-phosphate that can be converted to glycogen. Thirdly, glucose itself can increase glycogen formation by modulating phosphorylation. Lastly, insulin promotes glycogen synthesis by increasing the activity of two key enzymes (see below).

Whilst the liver has a high capacity to store and synthesize glucose, nerve cells can neither synthesize, concentrate nor store more than a trivial amount. Glial cells, however, may have a small store of glycogen. Thus, cells of the nervous system have an absolute requirement for a continuous supply of glucose that is usually oxidized fully to carbon dioxide and water. In the post-absorptive state, the brain uses about 60% of hepatic glucose production, although with time, i.e. during prolonged starvation, the brain can adapt metabolically to use ketones. This reduces but does not abolish the glucose need.

Actions of insulin and glucagon

The major targets for the anabolic actions of insulin are the liver, adipose tissue and muscle. At the cellular level, such target cells possess specific insulin receptors. The actions of insulin and glucagon controlling the overall flow of fuels are summarized in Boxes 2.6 and 2.7. In the liver, insulin promotes glycogen synthesis by stimulating glycogen synthetase and inhibiting glycogen phosphorylase although it has no direct effect on the GLUT 2 transporters and, hence, the uptake of glucose into hepatocytes. In contrast, insulin induces a rapid uptake of glucose in muscle and fat tissue by recruiting intracellular GLUT 4 transporters and, thus, increasing their cell-surface expression. As a consequence, muscle converts glucose to glycogen. In adipose tissue, glucose is converted to fatty acids for storage as triglyceride. Insulin also stimulates the uptake of amino acids into muscle. At the same time, insulin suppresses mobilization of fuels by inhibiting the breakdown of glycogen in the liver, the release of amino acids from muscle and the release of free fatty acids from adipose tissue. This explains, in part, why patients with DM such as Clinical Case 2.1 lose weight despite normal or increased appetite.

Glucagon has opposing actions (termed ‘counter regulatory’) to those of insulin, promoting mobilization of fuels, particularly glucose. Its primary action is on the liver where it binds to a seven-transmembrane G-protein-linked glycoprotein receptor and stimulates the production of cAMP. It may also activate the phosphatidylinositol signaling pathway. Through a subsequent cascade of intracellular events, glucagon ultimately stimulates the breakdown of glycogen to glucose and the production of glucose from amino acids (gluconeogenesis). In addition, it stimulates the release of free fatty acids from adipose tissue and, when these enter the liver, glucagon directs their metabolic fate. Rather than being used for the synthesis of triglycerides, they are shunted towards β-oxidation and the formation of ketoacids (see below). Thus, glucagon is both a hyperglycemic and a ketogenic hormone.

By and large, it is the molar ratio of insulin to glucagon in the portal blood that governs the metabolic state of the liver. As insulin concentrations fall during the post-absorptive state, its inhibitory effect on glycogenolysis is removed whilst the increasing concentration of serum glucagon reduces glycogen synthesis by inactivating (by phosphorylation) glycogen synthetase. In the same way that glycogen synthesis and breakdown is controlled by the molar ratio of insulin to glucagon, so too is gluconeogenesis. Thus, in the post-absorptive state, glycolysis is reduced and gluconeogenesis increases. In the early post-absorptive state, glycogenolysis and gluconeogenesis occur simultaneously but, as the glycogen stores become exhausted, gluconeogenesis becomes the sole source of glucose.

Other hormones that also play major roles in the regulation of blood glucose concentrations include cortisol, adrenaline and growth hormone, all of which act to raise blood glucose concentrations (Box 2.7) and are, thus, considered to be counter-regulatory and in excess diabetogenic.

Lipid metabolism - insulinopenia and diabetic ketosis

Lipids are stored as triglycerides (three fatty acids attached to a glycerol molecule) and transported around the body associated with proteins as lipoproteins. Clinical Case 2.1 had mild ketonuria as a result of disordered lipid metabolism, and these changes that occur in diabetes mellitus cannot be over-emphasized. Indeed, it has been said that, had it been as easy to measure fatty acids and triglycerides as glucose, diabetes would have been better known as a disorder of lipid metabolism rather than that of glucose.

The overall action of insulin on the adipocyte is to stimulate fat storage and inhibit mobilization. The remarkable effects of locally injected insulin on the accumulation of triglyceride into adipocytes are graphically illustrated in Box 2.8. Whilst this was likely to have been caused by de novo lipogenesis from glucose, it is generally believed that on a Western diet, triglycerides are usually accumulated in adipocytes by uptake from plasma (Box 2.9). This process is also stimulated by insulin-mediated activation of key enzymes inducing hydrolysis of verylow-density lipoproteins (VLDL) and chylomicron triglycerides into non-esterified or ‘free’ fatty acids (FFAs), thus making them available for uptake into adipocytes.

Insulin reduces fat mobilization from adipocytes by inhibiting hormone-sensitive tissue lipase and stimulating re-esterification of FFAs (Box 2.8). As insulin concentrations fall, FFAs are released from adipocytes into the circulation. Taken up by the liver, they can be esterified to form triacylglycerol stored in the liver and used either for hepatic energy needs or as the basis of VLDL formation. Alternatively, the FFAs may undergo β-oxidation that also produces energy for hepatic metabolism together with acetoacetate and 3β-hydroxybutyrate. These ketones are exported into the blood. The relative concentrations of insulin and glucagon regulate the rates of esterification and β-oxidation.

The uptake of FFAs into mitochondria, where β-oxidation occurs, is facilitated by the enzyme carnitine O-palmitoyltransferase-1. The activity of this enzyme is inhibited by malonyl-coenzyme A and when insulin concentrations are high so too are the concentrations of malonyl-CoA. Thus, β-oxidation is inhibited and ketone body formation is low. The reverse occurs when the insulin:glucagon ratio is reduced and ketone body formation increases. This, together with a reduction in insulin's suppressive actions on the release of fatty acids from adipocytes and increased gluconeogenesis, fuels ketone body formation (Box 2.10). The presence and degree of ketosis is a sensitive indicator of circulating insulin concentrations; a very important fact for clinical practice.

Ketone bodies are formed as acids (acetoacetic and hydroxybutyric). In cases of mild ketonuria, as seen in Clinical Case 2.1, the blood buffers these acids. When, however, ketone bodies are formed in large quantities, e.g. 7 mmol/min, the buffering capacity of blood (and the ability of the kidney to excrete protons) may be overwhelmed. This leads to a falling serum bicarbonate concentration, falling blood pH and a rising serum potassium concentration (due to the renal K⁺/H⁺ exchange) even though total body potassium is reduced as a result of the diabetic-induced polyuria. The effects of the acidosis on the chemoreceptor control of respiration lead to the characteristic deep, sighing (Kussmaul) respiration said to be characteristic of diabetic ketoacidosis but it may be seen in other forms of severe metabolic acidosis.

The interactions between glucose metabolism and the fatty acid cycle were noted more than 30 years ago. Elevated concentrations of free or non-esterified fatty acids (NEFA) increase lipid oxidation and decrease the activity of pyruvate dehydrogenase and phosphofructokinase, enzymes involved in glycolysis. The resulting increase in glucose 6-phosphate (Box 2.4) leads to a decrease in glucose uptake and oxidation. There is also evidence that increased concentrations of FFAs decrease non-oxidative glucose metabolism, i.e. glucose to lactate.

The changes in the concentrations of FFAs are translated into changes in the expression of genes by a family of nuclear transcription factors, the peroxisome proliferator-activated receptors (PPAR). After binding FFAs, these heterodimerize with the retinoid X receptor and interact with specific response elements in DNA. There are three PPARs (α, β, γ) in the family and they share strong structural similarities with other nuclear receptors such as those for steroids and tri-iodothyronine. PPARγ plays an important role in the adipocyte (Box 2.11).

Protein metabolism and the anabolic actions of insulin

The importance of proteins (or, more correctly, their constituent amino acids) in the control of blood glucose concentrations is that they can be converted to glucose by gluconeogensis or form ketoacids by ketogenesis. Alternatively they can be degraded and the released ammonium incorporated into urea in the liver via a biochemical pathway known as the urea cycle. In turn, this cycle is linked to the energy-producing citric acid cycle. Thus, whatever the catabolic fate of proteins, they are all energy-producing pathways.

Insulin promotes the uptake of amino acids (especially the essential amino acids leucine, valine, isoleucine, tyrosine and phenylalanine) into muscle and stimulates protein synthesis. Simultaneously, it prevents protein breakdown and the release of certain amino acids from muscle. Glucagon, acting predominantly on the liver, stimulates the extraction of amino acids from the circulation and increases the activity of the gluconeogenic enzymes whilst decreasing the activity of the glycolytic enzymes (Box 2.4).

It may be concluded that Clinical Case 2.1 had symptoms resulting from the osmotic effects of hyperglycemia and weight loss and ketonuria as a result of a loss of the metabolic effects of insulin on adipocytes, skeletal muscle and the liver (Box 2.12).

Definition and diagnosis of diabetes mellitus

Diabetes mellitus is defined solely in terms of elevated blood glucose concentrations (Box 2.13) and because of the vital role of insulin in regulating glucose metabolism a reduction in insulin secretion and/or insulin resistance is a common cause of DM. The lability of blood glucose concentrations in the post-prandial state makes it important that diabetes is defined in terms that reflect the timing of the last meal. In most human populations, fasting blood glucose concentrations are unimodally (or normally) distributed. Thus, diabetes mellitus must be diagnosed in a way that encompasses as many ‘true’ diabetics as possible, avoids the inclusion of ‘normal’ people and, yet, remains clinically relevant. The current WHO and the more recent American Diabetes Association criteria are given in Box 2.13.

For many years, the administration of a large oral liquid glucose load (75 g) combined with measurements of the subsequent changes in blood glucose concentration has been used to test pancreatic islet cell function. But a number of factors, apart from loss of insulin secretion, affect the ability of the body to deal with such an unphysiological meal (termed ‘glucose tolerance’). These include increasing age, physical inactivity, reduced carbohydrate intake and intercurrent illness, all of which reduce glucose tolerance. Asymptomatic people with decreased glucose tolerance and not meeting the criteria for DM are said to have impaired glucose tolerance (IGT). About 5–10% of people with IGT progress to DM each year.

Clinical Case 2.1 is used to illustrate the presentation of DM but also because of her very unusual reaction to the diagnosis. She raised questions concerning the etiology of type 1 DM. The biochemical measurements in the presence of classical symptoms confirmed the diagnosis of DM. In response to being told this, she became alternately sullen and monosyllabic and truculently voluble. She refused to give herself insulin and the medical team was forced to admit her to facilitate any treatment, even though physically she was virtually unaffected by her disease.

She demanded antibiotics to treat the newly diagnosed disease and clearly considered that her DM had an infective cause. It became apparent that she thought her diabetes mellitus had been sexually transmitted. Some months prior to the diagnosis, she had returned home early from work to find her husband in flagrante delicto with her younger sister (whom she knew had diabetes). Whilst being aware of a genetic tendency for the disease, she was convinced that her husband had passed the infective agent to her. This proved to be a source of enormous clinical difficulty (not to mention the strain it put on the marriage) but it highlights the possible interaction between genetic and environmental factors.

Etiology of type 1 DM

Type 1 DM is considered to be a T-lymphocyte-dependent autoimmune disease characterized by infiltration and destruction of the islets of Langerhans, the endocrine unit of the pancreas. The initiation and promulgation of the destructive processes remain poorly understood and genetic associations, environmental factors and immune reaction are all considered as causative agents.

Studies in twins have clearly established a major genetic element to type 1 DM, seen in Clinical Case 2.1 whose sister had DM. However, less than half of identical twins both develop the disease. Thus, it is considered that genetic elements form approximately 40% of disease susceptibility. Genetic associations with type 1 DM include those on the short arm of chromosome 6 (in the region of human leukocyte antigen (HLA) molecules, termed IDDM1), and regions of chromosome 11 upstream of the insulin gene itself (IDDM2). Together these putative genes have been numbered IDDM1–IDDM17, though some have only been documented in single families or small studies. Some of the genetic regions identified exert only a small influence and the precise genes remain unknown. More detailed studies have been possible in rodent models of type 1 DM and a number of genes have been implicated in the causation.

Type 1 DM is predominantly a disease of Caucasians living away from the equator and, whilst all ethnic groups may be affected, there are isolated areas of high incidence e.g. Sardinia. It tends to occur in patients <20 years of age; prevalence is approximately 0.2% of 20 year olds. Age of onset is no longer essential to its definition that requires both an insulin deficiency of such degree that ketosis occurs and usually the presence of markers of autoimmunity (Box 2.14). The geographical observations, together with the secular changes in some areas of the world where the incidence of type 1 DM is increasing, support the involvement of environmental factor(s) (Box 2.14). Overall, it occurs with equal frequency in males and females and there are increases in incidence around puberty and before starting school.

Whilst there is a body of experimental work in animal models and circumstantial evidence for infective causes in humans, direct evidence for infective agents as etiological agents of type 1 DM is sparse. Three mechanisms have been proposed by which infective agents may trigger autoimmunity. Damage of the islets and initiation of the ‘innocent bystander’ effect is thought to occur in congenital rubella, the best-known infective precipitant in the human. ‘Molecular mimicry’ (for example, structural similarity between the Coxsackie virus protein p2-C and the islet cell protein glutamic acid decarboxylase, GAD) is also postulated. Finally, production of microbial superantigens (proteins that facilitate the interaction of MHC class II molecules and polyclonal T-cells) as the result of activation of a human endogenous retrovirus has been suggested for some forms of adult-onset type 1 DM.

Clinical Case 2.1 was repeatedly assured that there was no known direct sexually transmitted causative agent but, despite being given considerable emotional support, she remained unconvinced. However, she eventually accepted dietary advice and started treatment with subcutaneous insulin.

Prevention of type 1 DM

It is generally accepted that the autoimmune destruction of the islet β-cells, once initiated, takes many months or years. The availability of genetic markers of predisposition together with serological markers of on-going autoimmunity such as antibodies to islet cells or insulin has led to trials of treatments to prevent type 1 DM in first-degree relatives of patients. None has proved very successful to date and studies continue.

Structure, synthesis and metabolism of insulin and glucagon

Human insulin contains 51 amino acids (molecular weight 5700) and is structurally homologous to insulin-like growth factors 1 and 2 (IGF-1 and -2) and also to the ovarian hormone, relaxin. It is synthesized in the β-cells of the pancreatic islets. The gene for insulin codes for pre-proinsulin which is made up of a signal sequence (approximately 23 amino acids that are rapidly cleaved after hormone synthesis has been directed to the endoplasmic reticulum), and the B chain, connecting (or C) peptide and A chain (Box 2.15). A and B are joined together by two disulfide bonds between common cysteine amino-acid residues. The C peptide is essential to the formation of these disulfide bonds and is cleaved in the Golgi apparatus leaving the joined A and B chains which form the active insulin molecule. It is to be noted that the cleaved C peptide is co-secreted with insulin, a point of great clinical importance (see below). Previously considered to have no physiological role, C peptide is now recognized to have G-protein-coupled cellular receptors and is likely to have some function in regulating blood flow and renal function.

Insulin is secreted in pulses every 10 min or so and has a t_1/2 in the systemic circulation of approximately 3 minutes. About 50% is removed by the liver. This is known as the ‘first-pass’ effect (i.e. the first time insulin passes through the liver). Insulin that has escaped the liver's inactivating activity has, of course, important regulatory actions on peripheral tissues. C-peptide is released in a 1:1 ratio with insulin and since it is not significantly removed by the liver and has a t_1/2 of 30 min its measurement has been used as an index of insulin secretion.

Glucagon contains 29 amino acids (molecular weight 3450) and has no disulfide bonds. The gene for glucagon encodes pre-proglucagon in the α-cells of the pancreas and this gene is a member of a superfamily of structurally similar genes coding for vasoactive intestinal peptide (VIP), gastrointestinal inhibitory peptide (GIP), secretin and growth hormone-releasing hormone (GHRH). After cleavage of the signal peptide, proglucagon is cleaved into active glucagon, glycentin-related pancreatic peptide (GRPP) and the major proglucagon fragment (Box 2.16). The same glucagon gene is also expressed in the gut where, after translation, it yields a different set of hormones from that of the pancreas (Box 2.16). The first-pass clearance of glucagon is also approximately 50% and its t_1/2 in the systemic circulation is approximately 3 min.

Anatomical features of pancreatic islets in relation to hormone secretion and its control

The tight coupling between insulin and glucagon secretion in relation to blood nutrients is possible because of the anatomical arrangement of cells in the islets and the direction of the blood flow in each islet. Surrounded by the exocrine pancreas, there are approximately 1 million islets that constitute 1–1.5% of the total human pancreatic mass (Box 2.17). Each islet contains a central core of insulin-secreting β-cells and a mantle of glucagon-secreting α- and/or somatostatin-releasing δ-cells (Box 2.18 ) or a mantle of δ- and PP cells, so called because they release pancreatic polypeptide. The δ-cells are dendritic in shape (i.e. they have many branches) and send processes into the core of the islet.

Each islet is highly vascularized with small arterioles entering its core. These break up into a network of capillaries that form venules carrying blood to the mantle. Such an arrangement of blood flow allows high concentrations of insulin to bathe the α-, δ- and PP cells (Box 2.18 ). Additionally, the capillaries within the islets are fenestrated, facilitating peptide entry into the blood stream.

Insulin is secreted in response to increases in glucose concentration in extracellular fluid. This metabolic signal requires metabolism of glucose to pyruvate and appears to be detected by the activity of the enzyme glucokinase that catalyzes production of glucose-6-phosphate. This initiates an insulin-releasing signal involving a rise in ATP, closure of a K⁺ channel, depolarization and opening of a Ca²⁺ channel. This process is very rapid and secretion of insulin occurs within one minute of exposure to glucose (Box 2.19). Detection of changes in glucose concentration is facilitated by the presence of canaliculi containing interstitial fluid along the lateral surfaces of neighboring β-cells between the arterioles and venules (Box 2.18 ). Concentrated in the microvilli of these canaliculi are the specific GLUT2 glucose transporters enabling the intracellular concentration of glucose in the β-cells to be essentially the same as that of the interstitial fluid.

Control of insulin and glucagon secretion

The control of insulin secretion is complex (Box 2.20). The most potent metabolic stimuli to insulin secretion are glucose and amino acids that act synergistically. Triglycerides and fatty acids have only a small stimulatory effect on insulin release. In animals, ketoacids may also induce insulin release but this is insignificant in the human. In response to an oral glucose load, insulin secretion occurs in two phases (Box 2.19). The first phase represents the release of insulin stored in secretory granules. Approximately 10 minutes later when pre-formed granules have been depleted, there is a more gradual and sustained increase in insulin release that can last for several hours in normal individuals and is dependent on de novo synthesis of the hormone.

The secretion of glucagon is stimulated by a reduction in blood glucose concentration. It is also stimulated by a protein meal, particularly by the amino acids, alanine and arginine (Box 2.21). These secretory responses are suppressed by the presence of insulin and glucose respectively (reducing transcription of the glucagon gene). Thus, ingestion of an ordinary meal produces much less variation in glucagon secretion than in that of insulin because of insulin's suppressive effect on glucagon release.

The secretion of both insulin and glucagon is potentiated by gastrointestinal hormones (termed incretins) that are released in response to orally ingested nutrients. Thus, an oral glucose load stimulates a greater insulin response than intravenous administration because one or more gastrointestinal hormones are released.

In addition to metabolic stimuli and hormones from the gut, insulin and glucagon secretion are also controlled by neural and paracrine mechanisms (Boxes 2.20 and 2.21). Sympathetic, parasympathetic and peptidergic nerves innervate the islets. Sympathetic nerves stimulate insulin release via β-adrenergic receptors (and inhibit via α-adrenergic receptors) whilst parasympathetic vagal nerves stimulate both insulin and glucagon release. This innervation accounts for the increase in insulin secretion that may occur before the entry of food into the gastrointestinal tract. Finally, somatostatin from the δ-cells inhibits the release of both hormones through a paracrine action (this has importance for clinical practice and is discussed in the context of Clinical Case 2.5).

Type 2 DM

It is clear from the foregoing that the complexity of the regulation of carbohydrate metabolism in general, and insulin and glucagon secretion and their actions in particular, means that there are numerous ways in which hyperglycemia (and thus, DM) can be caused. Whilst type 1 DM is associated with loss of insulin production, type 2 is classically associated with insulin resistance. However, in many populations it is characterized not only by insulin resistance but also by insulinopenia (loss of β-cell function) and the relative roles of each in the etiology of type 2 DM may vary in different populations and remain contentious (Box 2.22). It is noteworthy in rodent models of type 2 DM that even severe insulin resistance is not associated with the development of DM. This suggests that DM does not occur unless there is a failure of β-cell compensatory hypertrophy or hyperplasia.

Elevated circulating concentrations of glucose have an autoregulatory effect in enhancing glucose uptake, decreasing hepatic glucose production and increasing insulin production. However, in DM the prolonged stimulation of the β-cells depletes the insulin granule stores. β-cells are thus unable to secrete ‘pulses’ of insulin and become ‘blind’ to changes in glucose concentration. Hyperglycemia may also cause peripheral insulin resistance as a result of a down-regulation (i.e. decreased numbers) of GLUTs in peripheral tissues.

These effects have been termed, rather loosely, ‘glucotoxicity’ although studies using animal islets of Langerhans suggest that the cells are not actually injured by high glucose concentrations (at least in short-term culture). Histologically, loss of β-cell function in type 2 DM is associated with amyloid deposition within the islets of Langerhans. It is also seen in insulin-secreting tumors of the islets (insulinomas) and the peptide has been termed amylin or islet amyloid polypeptide (IAPP). It is structurally related to calcitonin-gene-related peptide and may act to reduce gastric emptying.

Type 2 DM is a heterogeneous polygenic disorder which is much more common (in the UK) in Afro-Caribbean and Asian populations. It has to be distinguished from type 1 DM presenting in adulthood and rarer forms of inherited DM (see below). The next two clinical cases illustrate acute and chronic presentations respectively of type 2 DM.

Clinical Case 2.2 illustrates a number of issues. The first important clinical observation is that type 2 DM is an insidious disease and patients with type 2 DM do not necessarily know that they have the disease. Metabolically, it is important to note that despite the very high blood glucose there was no ketosis indicating that there was sufficient circulating insulin to suppress the drive to ketone body formation. The extreme hyperglycemia lowered the serum sodium concentration by its osmotic action causing cellular dehydration (attracting water out of cells). The renal failure (evidenced by the marked elevation of serum urea and creatinine) was caused by the osmotic diuresis producing polyuria exacerbated by the fever and general prostration (reducing fluid intake) of the pneumonia. Such severe episodes in which conscious level is affected (due to dehydration of brain cells) are termed hyperosmolar non-ketotic coma (HONK).

Type 2 DM has an inherited component, though the strength of its effect continues to be debated. Twin studies have been variously interpreted as showing that as much as 90% or as little as 30% of disease susceptibility is inherited. In populations such as that of the Pima native Americans or the Pacific island of Narua, as many as 60% of the population may be diabetic or have impaired glucose tolerance. In the UK, the Caucasian population has an overall prevalence of the disease of 4% whilst the Asian and Afro-Caribbean population prevalence is four times higher. Type 2 DM is a complex heterogeneous disease and much work remains to be done to clarify the precise genes involved in its etiology.

Clinical Case 2.2 illustrates the two main environmental or acquired factors increasing the incidence of type 2 DM, age and obesity. The effect of age was reflected in the previous name for type 2 DM, maturity onset DM. Obesity markedly increases insulin resistance and also the likelihood of developing DM in a predisposed population. Other factors include a high-fat diet and ‘Western’ lifestyle. ‘Lipotoxicity’, the effect of increased concentrations of FFAs may also be involved. It was suggested nearly 40 years ago that relative skeletal muscle insulin resistance leading to increased deposition of triglyceride in adipose tissue was an evolutionary adaptation ‘thrifty gene’ allowing survival during periods of famine in hunter-gatherer populations. It is argued that exposure of such populations to cheap, easily available and plentiful high-fat foods in the twentieth century has led to epidemic obesity and type 2 DM.

Insulin resistance can arise as a result of mutations in the receptor itself. These are uncommon and many have unusual phenotypes (e.g. Leprechaunism or Rabson-Mendenhall syndrome) presenting in infancy. Insulin receptors are tetramers made up of two extracellular α-subunits, joined by disulfide bonds, and two transmembrane β-subunits each of which is joined to an α-subunit by disulfide bonds (Box 2.23). The binding of insulin to its receptor activates the intracellular tyrosine kinase on the β-subunits causing autophosphorylation of the receptor. In turn, the fully active tyrosine kinase phosphorylates a number of insulin receptor substrates (numbered IRS-1–5) and these can initiate a cascade of further events which ultimately alter enzyme activity, translocate glucose transporters to the cell membrane and stimulate or inhibit transcription in the nucleus. More than 70 mutations have been described and these have been classified according to the defect caused. Some decrease the rate of receptor synthesis whilst others impair transport of the receptor to the membrane or accelerate its degradation. Some decrease receptor affinity for the hormone and others impair tyrosine kinase signaling.

The vast majority of insulin resistance is, however, manifest by changes in the signalling pathways distal to the receptor. Studies on mice models (such as the IRS knock-outs) have thrown light on the mechanisms of insulin signalling. To date, the only defined defects in type 2 DM have been reductions in insulin receptor kinase or IRS protein tyrosine phosphorylation or PI3 kinase activation.

Epidemiological studies have suggested that people born with low birthweights are more likely to develop type 2 DM, hypertension and heart disease in later life. It has been suggested that poor early nutrition damages islet cell development and in the presence of a ‘Western’ lifestyle leads to impaired glucose tolerance and DM, the ‘thrifty phenotype’ hypothesis. It is to be noted that insulin is the major intrauterine growth factor and that inherited insulin resistance would be predicted to produce impaired intrauterine growth and later type 2 DM. Indeed, the IRS-1 knockout mouse has this phenotype. Thus, the low birthweight may just be an early reflection of insulin resistance rather than an etiological factor per se.

Causes of DM

A complete list of the causes of DM is dauntingly long. It is, however, possible to abbreviate it substantially whilst maintaining pathophysiological relevance and clinical utility (Box 2.24). It is obvious that all processes that damage islets cells are likely to lead to insulin deficiency whether they be trauma, pancreatitis, cystic fibrosis or the deposition of iron (hemochromatosis); these may, therefore, all be lumped together. Note that they do not give rise to type 1 DM since this is by definition autoimmune in origin (and idiopathic when markers of autoimmunity are absent). Note also that processes that damage all the islet cells and not just the β-cells lead to marked decrease in both insulin and glucagon. Such patients are likely, therefore, to need substantially less insulin and be less prone to ketosis.

Similarly, it makes little sense endocrinologically to maintain an inclusive list of all drugs that damage the β-cell unless the mechanism by which the damage is done is informative of the biochemistry of the cells themselves. Equally, long lists of excessively rare syndromes associated with DM serve only to test the memory unless the mechanism(s) (often due to single gene defects) underlying the DM add to our knowledge of β-cell function. This is exemplified in the following group of diseases.

Genetic disorders of β-cell function

As discussed previously, type 2 DM was previously termed maturity-onset DM. It was also termed non-insulin-dependent DM since treatment with insulin was not required to prevent ketosis. Some forms of non-insulin-dependent DM may present at an early age, typically between 15 and 30 years. This has been termed maturity-onset diabetes of youth (MODY), a term used to define a relatively small percentage of diabetic patients who have a strongly familial form of diabetes mellitus (autosomal dominant inheritance). Patients with MODY do not have ketosis or markers of autoimmunity. Genetic studies have defined a number of subtypes of MODY. These include mutations in the glucokinase gene that lead to abnormalities in sensing the circulating concentrations of glucose (MODY2). Mutations in the transcription factor hepatic nuclear factor HNF-1α (MODY3), -4α (MODY1) or HNF-1β (MODY5) could lead to abnormal gene regulation in the liver or pancreas whilst those in insulin promoter factor-1 (IPF-1, MODY4) have been associated with pancreas agenesis in homozygous form (Box 2.24).

Mitochondrial (mt) DNA is double-stranded, circular, highly polymorphic and only inherited from the mother. Because there are potentially many mitochondria in cells, the mutational burden (or percentage of mutant mt DNA) may vary in different tissues. There is a tendency for mutant mt DNA to accumulate in slowly or non-dividing cells (such as neurons). Mutations at np3243 are believed to account for 1–2% of all non-insulin-dependent DM patients and may be found in MELAS syndrome (Mitochondrial Encephalopathy, Lactic Acidosis and Stroke-like episodes). Such forms of DM may be misdiagnosed unless the link of maternal inheritance and associated neurological features is noted.

Counter-regulatory hormones and DM

Whilst most DM is caused by a loss of β-cell function and/or insulin resistance, it may also result from an excess secretion of the hyperglycemic (counter-regulatory) hormones - cortisol (see Clinical Case 4.1), catecholamines (see Clinical Case 4.7) and somatotrophin (see Clinical Case 7.4). The clinical features of these cases are discussed elsewhere but the effects of such hormone excess on carbohydrate tolerance have not been emphasized. Patients with these diseases frequently develop secondary diabetes or impaired glucose tolerance because the excess hormone-induced hyperglycemia may lead to insulinopenia and/or peripheral resistance to insulin. The mechanisms by which these hormones induce hyperglycemia are summarized in Box 2.25.

Although DM is compatible with a normal life span, on average it shortens life expectancy by about 30%. It causes damaging complications, the treatment of which absorbs approximately 10% of the NHS budget.

Complications of DM

Diabetes is the most common cause of blindness in adults of working age, the most common cause of end-stage renal failure requiring dialysis or transplantation and the most common cause of non-traumatic amputation. It is an important factor in the etiology of heart attack and stroke. Clinical Cases 2.3 and 2.4 illustrate the range and scope of complications and discussion will focus on the mechanisms of their generation.

Clinical Case 2.3 like Case 2.1 illustrates the rapid onset of the clinical features in type 1 DM once more than about 80% of islet β-cell function has been lost. It is reasonably certain that she only had diabetic symptoms for a few days or weeks before she was diagnosed in the autumn (the most common time of the year in the UK for type 1 DM to present). In less than 3 months, her diabetes had made her virtually blind. Such acute diabetic cataracts are rare. More usual are the chronic cataracts that typically develop over the age of 40 years and are more common in diabetics. Acute cataractogenesis may arise from an increased production of sorbitol through the polyol pathway (Box 2.28) that may cause osmotic swelling and damage to the lens.

The underlying diagnosis of schizophrenia made it more difficult to be certain of the duration of type 2 DM in Clinical Case 2.4, but the complications have clearly taken some years to become apparent.

Macrovascular circulatory changes

Most diabetic complications arise from damage to blood vessels. Those arising from accelerated atherosclerosis particularly affect the coronary, carotid and femoral arteries (and are termed macrovascular, Box 2.27). Those more specific to diabetes affecting the retina, kidney and nervous system (and termed microvascular) give rise to retinopathy, nephropathy and neuropathies respectively (see below).

The atheromatous plaques in diabetic macrovascular disease are no different from those in non-diabetics though they tend to be more common and more extensive. The processes through which diabetes is thought to induce atherosclerosis are, in part, due to the hyperglycemic effects on endothelial cell structure, platelet adhesion and stimulatory factors in plaque formation (Box 2.27). DM also has a prothrombotic effect.

Microvascular changes - diabetic retinopathy, nephropathy and neuropathy

A number of hypotheses (that are not mutually exclusive) have been proposed to account for the etiology of microvascular complications (Box 2.28). An increase in the polyol pathway that reduces glucose to fructose increases the NADH/NAD⁺ ratio and decreases the intracellular concentration of myoinositol, a precursor of phosphatidylinositols, has been suggested. The increased NADH:NAD⁺ ratio (reflecting ‘oxidative stress’) leads to increased protein kinase C activation and increased intracellular glyceraldehyde concentration. Non-enzymic glycation of proteins occurs (the same chemical reaction that causes the browning of a peeled apple). Such protein glycation may affect protein function and, thus, alter the structure and function of the microvasculature or lens of the eye. Glycation of proteins such as hemoglobin (hemoglobin A_1c) may be used as an index of diabetic control.

The structural changes in the microvasculature include thickening of the basement membrane of capillaries, affecting both their permeability and structural integrity, probably via increased synthesis and reduced catabolism of glycoproteins. Clinical Case 2.4 had retinopathy and renal failure as a result of his DM. Whilst microvascular changes can be seen in almost all tissues of the body, the poorly supported capillaries of the retina and kidney are very vulnerable in DM. Examination of the microcirculation of the retina with an ophthalmoscope provides a valuable way of assessing the degree of structural and functional microangiopathy.

Retinopathy is usually graded into background and proliferative retinopathy (Box 2.29). Tiny hemorrhages and areas of capillary exudate are characteristic of background retinopathies. They do not cause visual impairments, unless they encroach on the macula. When they do aggregate on the macula, visual acuity is reduced and the retinopathy is classified as a maculopathy. Background retinopathy may progress to proliferative retinopathy because the ischemia, resulting from the damaged capillaries, induces growth of new capillaries (angiogenesis). These new vessels are fragile and bleed readily into the vitreous humor and, if the proliferation is left untreated, will lead to total blindness within 5 years. Proliferative retinopathy is prevented by laser photocoagulation of the retina. This is believed to remove the source of growth factors released from the leukocytes that accumulate in the venule end of the capillaries as a result of their structural changes.

The first clinical sign of nephropathy in a diabetic patient is the presence of small amounts of protein in the urine (termed microalbuminuria). Once established, it leads, over a period of years, to larger amounts of proteinuria and chronic renal failure. Hypertension develops at about the same time in type 1 DM but may already be established in type 2 DM. Uncontrolled hypertension is a major factor in the progression of DM complications, particularly nephropathy.

Diabetes has many effects on the nervous system (Box 2.30) and whilst the majority of these are thought to arise from metabolic effects, some such as the cranial neuropathies or the mononeuropathies may arise as a result of microvascular occlusion which removes the blood supply to individual nerve bundles. Several studies have shown that careful glycemic control may reduce the prevalence and severity of diabetic complications.

The insidious nature of the onset of type 2 DM means that about 20% of patients have microvascular changes at the time of diagnosis indicating that they had elevated blood glucose concentrations for a considerable period prior to presentation. This has implications for the screening for type 2 DM in predisposed populations.

Diabetes and the neuropathic foot

It is clear from the attempt by Clinical Case 2.4 to leave the Emergency Room barefoot that his right foot must have been painless, despite its appearance (Box 2.31). This patient had a neuropathic foot, one of the most common neurological sequelae of the disease; it is usually bilateral and symmetrical. The problem in this man's right foot arose because of a progressive distal sensory neuropathy (Box 2.30) resulting from the metabolic effects of DM on the nervous system. This predisposed the foot to unnoticed injury. His poor footwear led to an ulcer that was secondarily infected.

Angiography confirmed that macrovascular disease, which causes arterial ischemia and makes tissues vulnerable, played no role in the pathology of this man's foot problems (Box 2.31). He was treated with bed rest, local chiropodial surgery to the ulcer, intravenous antibiotics and improved control of blood glucose concentrations. He refused major surgery on his leg because his belief in the dinosaur remained unwavering, despite medication for his schizophrenia. He needed many months in hospital and his mental state raised medicolegal issues regarding his ability to give consent (website).

Diabetes and insulin resistance of pregnancy

One of the dominant metabolic effects of normal pregnancy is an increase in insulin resistance, probably induced by placental hormones including progesterone and placental lactogen. This leads to higher postprandial glucose concentrations that are considered to improve fetal growth; it is termed ‘facilitated anabolism’. Fasting glucose concentrations decrease as a result of placental glucose transfer and in the later stages of pregnancy, there is also enhanced maternal lipolysis. This is considered to spare glucose for the fetus and is termed ‘accelerated starvation’.

In genetically predisposed women, the normal insulin resistance of pregnancy may lead to the diagnosis of DM for the first time, termed ‘gestational diabetes’. This may disappear within hours of giving birth depending on individual factors such as islet β-cell function and predisposing factors such as obesity. Women with pre-existing DM require higher doses of insulin during pregnancy and patients who are usually controlled using oral hypoglycemic agents are transferred to insulin at this time.

The effects of pregnancy-induced insulin resistance in women with DM lead to poorer control of blood glucose and also an increased likelihood of ketoacidosis. The hyperglycemia in early pregnancy has considerable effects on the development of the fetal pancreas. Maternal ketoacidosis leads to fetal loss.

Development of the pancreas: effects of DM on organogenesis

The pancreas develops from two buds of gut endoderm termed the ventral and dorsal initially though they are affected by gut rotation (Box 2.32). The islets arise from specialized buds of the cells that give rise to the pancreatic ducts and exocrine cells. Insulin is present within the fetal pancreas by 9 weeks gestation. Islet development in both animal models and in humans is affected by the metabolic environment. Poor maternal diabetic control in early pregnancy leads to an increase in fetal islet β-cell mass. This programming may continue during later pregnancy so that the increase in β-cell mass may continue to have fetal effects even though maternal DM control has improved.

Poor DM control at the time of fetal organogenesis is teratogenic, producing malformations of the heart, nervous system and skeleton in particular. The teratogenicity is related to the degree of diabetic control and has been termed ‘fuel-mediated teratogenesis’.

The patient in Clinical Case 2.4, who became pregnant a year after her diabetic cataracts had been removed, subsequently delivered a very large and physically disproportionate baby due to fetal overgrowth (macrosomia - Box 2.26). This was caused by her poorly controlled hyperglycemia which, by transplacental transfer, caused increased fetal insulin secretion. Insulin, like IGF-1 and -2 is a growth factor and is the major factor for fetal growth in utero. In addition, diabetic pregnancies are associated with an increased incidence of miscarriage, intrauterine death, pre-eclamptic toxemia and neonatal respiratory distress syndrome. Tight control of blood glucose during pregnancy can prevent many of these problems but the high incidence of congenital malformations (approximately 10%) means that efforts to achieve good glucose control should be established prior to conception.

Treatment of DM - rationale and practical considerations

It is important to emphasize basic principles.

Treatment targets should be clinically relevant. Thus, the degree of glycemic control set for a 20-year-old pregnant woman will not be the same as that, say, for her 80-year-old grandmother. The monitoring of glucose control should be practical and reflect these clinical targets. Thus, measuring glucose concentrations in urine may be all that is required for some patients whilst others may need to check capillary blood glucose concentrations several times daily. Within these limitations, it is to be emphasized that DM complications are less in those with better DM control.

Insulinopenia of such a degree as to cause ketosis requires immediate insulin therapy. A corollary of this is that keto(acido)sis is always a medical emergency. The degree of ketosis and acidosis should be carefully documented and monitored during therapy. In an established diabetic, the cause of the ketoacidosis should always be sought and particular attention should be paid to the possibility of poor DM management, infection and vascular events (such as myocardial infarct or stroke)

The genetic implications of the diagnosis for the rest of the family should also be given consideration. Relatives of diabetic probands should be screened on a regular basis. Prevention of DM will receive greater attention as the worldwide epidemic threatens. For type 2 DM, prevention or reduction of obesity and the use of exercise programs has been shown to prevent or delay disease onset. Large-scale (and expensive) public health measures to reduce obesity have not proved successful to date. In type 1 DM, a variety of preventative measures are under consideration or trial.

Dietary modification is fundamental to the long-term treatment of all forms of DM. The generally recommended diet, the so-called ‘healthy-eating diet’, contains >55% carbohydrate, 10–15% protein and <30% fat (<10% saturates) and, in general, needs no modification other than to add an injunction to reduce sources of simple sugars (<25 g per day added sucrose) and to replace them with complex carbohydrates. In the case of type 1 DM there is a requirement to balance the amount of carbohydrate with the insulin dose at any meal, with the proviso that this may be altered by the amount of exercise to be done. The need for between meal or bedtime carbohydrate snacks is dependent on the time course of action of the insulin used. Alcohol should be taken in moderation and smoking should be scrupulously avoided.

Treatable causes of insulin resistance should be removed or minimized. The main one is obesity. There is good evidence that even modest weight loss (say 5–10 kg in an adult) results in improved diabetic control and, long-term, in a decrease in cardiovascular events and deaths. Regular exercise (that improves insulin sensitivity) should be encouraged. Generally, this succeeds best if it is incorporated into normal life; examples include cycling or walking to work and using the stairs rather than the elevator. It is important to emphasize that significant weight loss cannot easily be achieved by increased exercise alone. Calculating the exercise required to consume 1 kg fat (based on the assumption that this contains 7000 kcal/ 30 000 kJ energy) one would need to jog for about 14.5 h, play football for 12.7 h or cycle 175 km.

Drug therapies (Box 2.33, Box 2.34, and Box 2.35) should be kept to the minimum that is effective. They should not be used as an alternative to changes in eating patterns, but as adjuncts. Glucosidase inhibitors, such as acarbose, may help reduce post-prandial peaks of serum glucose, but have major gastrointestinal side effects. The effects of the soluble form of amylin pramlintide on gastric emptying (and, thus, slowing glucose absorption) in type 1 DM have been studied. Agents such as the pancreatic lipase inhibitor orlistat may aid the reduction in obesity. For the obese, metformin or the recently introduced PPARγ agonists thiazolidinediones e.g. rosiglitazone may aid the improvement in insulin resistance. It is to be emphasized that adjunctive therapies may be needed for additional metabolic problems such as hyperlipidemia or for the treatment of systemic hypertension that is so often an accompaniment to type 2 DM.

Specific treatments are being developed to prevent the complications of DM. These include orally active inhibitors of aldose reductase, inhibitors of non-enzymatic glycation such as aminoguanidine or the protein kinase C inhibitor LY333531.

Insulin and sulfonylureas should be used with due circumspection. In Western countries, the majority of insulins now available are human. However, it will be immediately clear that there is virtually nothing else physiological about current insulin replacement therapy. Physiologically, insulin is secreted from β-cell granules (probably as hexamers and dimers) and enters the circulation through specialized fenestrations in the capillary endothelium that increase permeability 5–10 fold. The insulin is carried directly to the liver in the portal vein. Its short-lived pulses (t_1/2 approximately 3 min) are exquisitely related to the metabolic milieu that is regulated within fine limits.

In contrast, replacement insulin is injected subcutaneously into the systemic circulation. Absorption is slow, extremely variable and dependent on multiple factors including the site in the body, capillary density, temperature, blood flow and, not least, the method used to slow its absorption. The vast majority of modifications of insulin have, to date, involved the use of materials such as zinc or proteins such as protamine to slow absorption (Box 2.33). As a result, after a meal all patients taking insulin are inadequately replaced early and over-replaced some hours after. Recently, molecular techniques have been used to alter the structure of human insulin. Site-directed mutagenesis has been used to create novel human insulins (e.g. human insulin lispro) with structures that have a decreased tendency to form dimers and hexamers. The absorption of these is much more rapid, less variable and as a result improves post-prandial control of glucose.

A wide variety of sulfonylureas is available (Box 2.34). These act on the sulfonylurea receptor of the K⁺-ATPase channel (Box 2.19) to increase insulin secretion. They all bind strongly to albumin. They vary in cost and duration of action and are best used in those in whom insulin resistance due to obesity has been addressed. They have the serious side-effect of weight gain and the potentially fatal one of hypoglycemia (see Clinical Case 2.6). A recent development has been a non-sulfonylurea agonist repaglinide that has a shorter duration of action and may improve post-prandial glucose excursions. Newer sulfonylureas have greater potency (i.e. effect per milligram) but there is little evidence that they have any greater maximal effect on insulin secretion and improved clinical benefit. Older drugs are often cheaper and available from generic manufacturers.

The treatment of DM is bedeviled by the problem of hypoglycemia, the sword of Damocles that hangs over every patient attempting to normalize his or her blood glucose. Were it not for hypoglycemia, the treatment of DM would be child's play.

Hypoglycemia

Most cases of hypoglycemia seen in clinical practice are diabetics over-treated with insulin or sulfonylureas.

The brain with its absolute dependency on a continuous supply of glucose is central to discussions on hypoglycemia and on how hypoglycemia is defined. Glucose crosses the blood-brain barrier by facilitated diffusion via endothelial GLUT 1 receptors (Box 2.4) and this is the major rate-limiting step. At normal blood glucose concentrations, the rate of supply is approximately twice that of neuronal glucose utilization. As the arterial plasma glucose concentration falls below approximately 3.6 mmol/l, this transfer becomes rate limiting to neuronal glucose metabolism. Clearly what matters, therefore, in determining hypoglycemia is cerebral capillary glucose concentration. However, in ordinary clinical practice the only available measure is venous serum glucose concentration and there has been long discussion as to what this value should be in order to define hypoglycemia. A commonly used value is 2.2 mmol/l but this value can only be used to diagnose hypoglycemia when two other criteria have also been established viz. the symptoms experienced should be compatible with hypoglycemia and they should be improved when the hypoglycemia is corrected. These three criteria form Whipple's triad. However, the rigid use of a single value of glucose concentration in the definition of hypoglycemia is difficult to justify both physiologically and clinically.

The symptoms and signs of hypoglycemia have been divided into those due to lack of glucose to the brain (neuroglycopenia) and those due to increased activity of the sympathetic nervous system (Box 2.36). The causes of hypoglycemia are given in Box 2.37. Mechanistically, it is clear that the causes can be divided into those that increase its removal from the blood (i.e. hormones and drugs) and those that reduce its entry into the blood (i.e liver failure, hormone deficiencies). The hypoglycemia seen in Clinical Case 2.5 was caused by both of these.

She was taking regular glibenclamide, the long-acting second-generation sulfonylurea responsible for most cases of hypoglycemia reported in the medical literature. On the night she became ill she also drank the best part of 180 g ethanol. If taken over, say, a 2 h period, (assuming total body water to be 39 l with a clearance rate of alcohol at 8 g/h) her serum ethanol concentration would have been approximately 400 mg/d - approximately 5 times the current U.K. legal drink-drive limit. Ethanol is not taken up by tissues and stored. It, thus, forces its metabolism by hepatic alcohol dehydrogenase. This changes the hepatic NAD⁺/NADH ratio that reduces hepatic gluconeogenesis, particularly from lactate.

Ethanol is an important dietary component in a significant proportion of the population and discussion continues as to whether it is a macronutrient or a poison. It can cause either hyper- or hypoglycemia. This depends on the nutritional state of the imbiber and the carbohydrate intake with the ethanol. The original descriptions of ethanol-induced hypoglycemia were in patients in poor nutritional state (with poor hepatic glycogen stores) who were fasted. In the well-nourished, non-fasting subject ethanol may well cause hyperglycemia by a peripheral action to increase insulin resistance. Clinical Case 2.5 was not in a poor nutritional state and it is unlikely that ethanol alone would have caused her hypoglycemic presentation with continuous generalized fits. What is more likely is that she took her glibenclamide in the evening and did not eat her usual meal; ethanol would have reduced awareness of the resultant hypoglycemia. The absence of oral carbohydrate intake and the alcohol-induced inhibition of gluconeogenesis meant that the drug-induced hyperinsulinemic hypoglycemic state was not correctable. Prolonged hypoglycemia, particularly when followed by epileptic activity, leads to neuronal death (through mechanisms that remain ill-understood). Indeed, this patient never recovered and remained dependent on full nursing care.

Whilst alcohol was pivotal to this patient's demise, the question sometimes arises as to whether episodes of hypoglycemia are accidental or deliberately induced by the patient (suicide) or by someone else (murder). To understand these important medicolegal possibilities it is important to reconsider the synthesis and secretion of insulin (Box 2.15). Cleaved C-peptide is co-secreted with insulin but is not cleared by the liver and has a longer t_1/2 in the systemic circulation. Commercially available human insulin contains no C-peptide, so in the case of exogenous hyperinsulinemia circulating concentrations of insulin would be high but the C-peptide low or absent. If the hyperinsulinemia were induced by sulfonylurea stimulated β-cell function, then the concentrations of both immunoreactive insulin and C-peptide would be high.

Physiological responses to hypoglycemia and its treatment

As arterial blood glucose falls below about 4.5 mmol/l, serum insulin concentrations fall. With a continued lowering of blood glucose to below about 3.8 mmol/l there is secretion of the counter-regulatory hormones, glucagon, catecholamines, cortisol and somatotrophin (of which the most important in the acute situtuation are glucagon and epinephrine). With further reductions of blood glucose (at about 3.0 mmol/l) there is symptomatic awareness and at around 2.6 mmol/l cognitive dysfunction is apparent. However, it is clear that recent episodes of hypoglycemia can alter (increase) these thresholds so that greater reductions in serum glucose concentrations are required before hormonal responses and symptoms are manifest. This is true both for normal volunteers and diabetics. The reverse occurs (decreased thresholds) in diabetics who have had a period of high serum glucose concentrations. Thus, depending on preceding blood glucose concentrations, a diabetic may be symptomatically hypoglycemic at a ‘normal’ serum glucose concentration or asymptomatic with a serum glucose of, say, 2.5 mmol/l.

The immediate treatment of a hypoglycemic emergency is to restore circulating glucose concentrations using a glucose infusion if necessary (Box 2.38). Additional therapy is used to reduce circulating insulin concentrations. Clinical Case 2.5 was additionally given an octapeptide analog of somatostatin, octreotide, to try to reduce the glibenclamide-induced insulin secretion. In this case it had little demonstrable effect, probably because of the delay in the patient's clinical presentation. Glucagon may be given to stimulate hepatic gluconeogenesis, but in sulfonyl-urea-induced hypoglycemia it may cause additional insulin release. Other drugs used in this situation include diazoxide (Box 2.38) this drug opens the K⁺-ATPase channel in the β-cell and this inhibits insulin secretion (see Box 2.19).

Hypoglycemia and insulinoma

The next case, though rare, illustrates the fact that not all cases of hypoglycemia are caused by diabetic drugs.

This patient has severe symptomatic spontaneous hypoglycemia with low venous serum glucose (1.0 mmol/l). The serum insulin (78 pmol/1) and C-peptide (916 pmol/1) concentrations were high at the time of hypoglycemia. From the foregoing, it is clear that this patient has an endogenous source of hyperinsulinism and an insulinoma was top of the diagnostic list.

These are rare tumors (0.5 per 100 000 per annum) occurring most frequently between the ages of 30 and 60 years and with equal incidence in both sexes. The vast majority (>99%) occur in the pancreas and about 10% are malignant. Thirty per cent are less than 1 cm diameter at diagnosis. The diagnosis of insulinoma may well be achieved late in the course of the disease. Once diagnosed, the problem is to localize it within the pancreas (Box 2.39) and to remove it surgically. In Clinical Case 2.6, a CT scan of the pancreas and the liver was normal. However, a ^IIIIn-labeled octreotide (somatostatin analog) scan showed a clear, but small, midline area of increased uptake (binding) of this somatostatin analog (Box 2.39). After initial medical treatment with diazoxide, the small benign adenoma of β-cells causing the insulinoma was removed with resolution of the symptoms.

Hypoglycemia in infancy

There are some causes of hypoglycemia that come to light early in childhood. These include transient hypoglycemia in the neonatal period (more common in pre-term or small-for-gestational-age infants), hyperinsulinism and congenital enzyme defects. Some of these such as hereditary fructose intolerance and galactosemia present with post-prandial hypoglycemia. Such hypoglycemia has been termed ‘reactive hypoglycemia’. This diagnosis was overused in the 1960s and 1970s and fell into disrepute. It is probably preferable to use the term post-prandial hypoglycemia, to remember to enquire as to the types of food that precipitate symptoms and to insist on the criteria in Whipple's triad being met.

The causes of hypoglycemia in infancy and childhood form another rather daunting list (Box 2.40). However, it is possible to clarify the subject somewhat and make some generalizations. The first is that all infants are prone to hypoglycemia, probably as a result of greater brain mass relative to body size. In addition, there may be some immaturity of gluconeogenic pathways in the pre-term and small-for-gestational-age infants. Second, the most common cause of hyperinsulinism is maternal DM (as seen in the child of Clinical Case 2.3). Fetal hormonal adjustments to maternal pathology may cause problems neonatally. In the case of fetal hyperglycemia, this leads to hyperinsulinism and hypoglycemia in the early neonatal period. Third, the majority of enzyme defects leading to hypoglycemia will be those disrupting glycogenolysis and gluconeogenesis, fatty acid oxidation or amino acid supply for gluconeogenesis. Examples of each of these are given in Box 2.40.

Mutations in the K⁺-ATP channel in β-cells may lead to inactivation of the channel and membrane depolarization and insulin release. These may be inherited in autosomal dominant or recessive forms and give rise to persistant hyperinsulinemic hypoglycemia of infancy (PHHI) previously termed residioblastosis.

Disorders of the α, γ and PP cells of the islets

Both insulin deficiency and its excess have been discussed. Functional glucagon deficiency may be seen in diabetics, increasing with the duration of the disease and being virtually universal after about 20 years of diabetes mellitus. Though α-cells are present within the islets they become functionally insensitive to hypoglycemia through mechanisms that are unknown. As would be predicted from the actions of glucagon, this results in markedly slower recovery from hypoglycemia.

Clearly, a state of clinical glucagon excess might be expected to result in DM. Glucagon excess may be seen in rare tumors (glucagonomas) of the α-cells. These are seen in the middle aged and older and the catabolic effects on muscle result in weight loss and wasting. The increased uptake of amino acids leads to a reduction in circulating concentrations. The DM is usually mild (resulting from increased hepatic gluconeogenesis) and is not associated with ketosis. For reasons that are not understood glucagonomas produce a necrolytic rash that migrates to different areas of skin and to thrombosis in blood vessels.

An excess of somatostatin (seen with extremely rare somatostatinomas of the γ-cells of the islet) also causes mild DM and circulating concentrations of both glucagon and insulin are low. The gut effects are noteworthy with malabsorption of fat (termed steatorrhea) and the formation of gallstones. Treatment of patients with octreotide or lanreotide results in similar gut effects that can limit its use. Tumors of the islet PP cells secreting pancreatic polypeptide occur but seem not to be associated with clinical sequelae. The principles of treatment of these tumors of the islets are the same as those in insulinoma.

Clinical case questions

The following are examples of applied pathophysiology and these clinically based questions can be answered with the information provided in this chapter.