13.4.2 Pharmacological therapy of hyperglycaemia in type 2 diabetes mellitus

The management of type 2 diabetes is complex, due to the diverse, variable, and progressive nature of its pathogenesis, clinical complications, and societal impact (Box 13.4.2.1). Care plans need to be individualized, flexible, and realistic, with provision for patient education and empowerment to enable optimal benefit from the guidance and interventions offered by health care professionals. Relief of acute symptoms and attention to long-term complications and co-morbidities often preoccupy and sometimes overwhelm the treatment process. However, early and sustained remediation of endocrine and metabolic disturbances, plus containment of modifiable cardiovascular risk factors, prevent the onset and limit the severity of chronic pathology. Glycaemic control is a crucial part of the treatment process, and serves as the conventional indicator of metabolic status. This chapter will focus on the treatment of hyperglycaemia, and, particularly, the role of pharmacological therapies.

The importance of metabolic control has been discussed in Chapter 13.4.1, but it is pertinent to re-emphasize that good glycaemic control can reduce both micro- and macrovascular complications. The reductions in microvascular disease have been confirmed in almost all trials of intensive glycaemic control, whereas improved cardiovascular outcomes may only emerge after several years, being more evident in younger patients with better control from the time of diagnosis (1). Moreover, given time, the closer that glycaemic control approaches normoglycaemia, the fewer the complications (Fig. 13.4.2.1), excepting that over-intensification of treatment to cause persistent or severe episodes of hypoglycaemia must be avoided (2).

The post-trial follow-up of the UK Prospective Diabetes Study (UKPDS) has provided clinical confirmation of the so-called ‘glycaemic memory’ or ‘legacy effect’. This effect, which probably reflects the accumulated damage from glucotoxicity, renders individuals with poor glycaemic control during earlier stages of the disease at much higher risk of complications in later life, even if their control is subsequently improved (3). Thus, available evidence supports the early use of blood glucose-lowering therapy to attain and maintain a level of glycaemic control as close to normal as is safe, practicable, and commensurate with the circumstances of the patient.

There is no shortage of guidelines and consensus statements to advise on the use of pharmacological therapies in the treatment of type 2 diabetes (4–6). Despite the nuances of contemporary practices and preferences in different parts of the globe, and the desire for evidence-based medicine, the different guidance documents show a high degree of conformity (Fig. 13.4.2.2).

Lifestyle measures invariably form the foundation for management of type 2 diabetes, especially diet and exercise for weight ontrol, cardiovascular wellbeing, and psychosocial health (7, 8). Pharmacological therapies are added to achieve and sustain the desired glycaemic target, generally starting with one oral agent. If the target is not achieved or maintained by increasing the doses, a differently acting oral agent or injectable agent is added. If this combination is unable to prevent disease progression, a third agent is added. The selection of agents at each stage will be strongly influenced by the presence of co-morbidies, including obesity, renal or hepatic impairment, age, and personal circumstances. Submaximal doses of two or more agents may be preferred to enhance efficacy and reduce side effects that are sometimes encountered with a large dose of one agent. Insulin is customarily reserved for patients in whom a combination of other agents does not provide continuing control, but insulin may be introduced earlier if patients are severely hyperglycaemic and substantially symptomatic with co-morbidities that deter the use of other therapies.

The main orally administered blood glucose-lowering therapies are metformin, sulphonlyureas, meglitinides, thiazolidinediones (TZDs), α-glucosidase inhibitors (AGIs), gliptins, and, most recently, bromocriptine (9, 10). The parenterally administered agents are glucagon-like peptide-1 (GLP-1) analogues, pramlintide, and insulin. The main actions, typical efficacy, and important cautions for agents other than insulin (considered in detail later) are summarized in Table 13.4.2.1.

Not all antidiabetic agents are available in all countries, and they do not always carry the same indications for use, e.g. pramlintide is not available in Europe and gliclazide is not available in the USA. The same drug may be named differently in different countries—e.g. glibenclamide (Europe) is glyburide in the USA—although similarly named drugs may have different formulations, e.g. glipizide. Also, the nomenclature for pre-mixed insulins varies: the percentage of short-acting is numbered first in Europe, but second in USA. The exclusions, precautions, and monitoring may also vary. The same drugs can have different exclusion criteria—e.g. TZDs are excluded for New York Heart Association (NYHA) categories I–V in Europe, but III–IV in the USA—maximum doses are sometimes different, and countries often introduce local restrictions, e.g. GLP-1 analogues are restricted to patients with a high body mass index (BMI) in England. For these reasons, it is not possible to be fully prescriptive or inclusive in this chapter, and prescribers are urged to check national and local formulary restrictions before administering pharmacological therapies.

If hyperglycaemia is severe, or initial lifestyle measures do not achieve desired glycaemic control after one to two months, pharmacological therapy is indicated. Some guidelines advise immediate prescription of metformin (q.v) at the same time as instituting lifestyle change. In theory, a pharmacological agent might be chosen to address a predominant underlying pathogenic factor, such as insulin resistance or β cell dysfunction. In practice, however, it is often difficult to determine the relative impact of different pathogenic factors, and therapies may be prescribed on the basis of local protocols, patient preference, or what is left when preferred options are excluded by contraindications or tolerability (7, 8).

The suitability of a treatment for an individual requires consideration of the risks of the disease, risks of other agents, and risk of the proposed agent compared with the potential benefits for the current and likely future circumstances of that individual.

Starting pharmacotherapy assumes that a chosen agent is not precluded by co-morbidity, potential interactions with other medications, or incompatibility with lifestyle, and that appropriate information and support are offered to the patient. Note baseline glycaemia and begin therapy with a low dose. A summary of individual agents with pharmacokinetic information is given in Table 13.4.2.2. Titrate up at intervals of one to two weeks for metformin, sulphonylureas, meglitinides, and AGIs. Titration is usually monthly for injectable agents other than insulin (considered in detail later in this chapter), and sometimes slower for a TZD. Gliptins do not usually require titration. Continue to titrate until the desired level of glycaemic control is achieved. If intolerance supervenes, reduce a dose level and attempt titration again later. If a titration step does not provide any additional benefit, return to the previous dosage and, if significant adverse events are experienced, consider discontinuation and switching to another class of agent. Switching within class is rarely helpful, except when contraindications develop that can be circumvented by different pharmacokinetics. Appropriate monitoring, which may extend beyond glycaemic parameters, and reinforcement of lifestyle compliance, should undertaken as required.

Single-tablet fixed-dose combinations may be used as a convenient way to reduce the pill burden. Bearing in mind the progressive nature of type 2 diabetes, additional therapy to address deteriorating control, or switching therapy to accommodate emerging comorbidity, is an expected part of the treatment process. Rapidly advancing hyperglycaemia in patients with long-standing diabetes, often with unintentional weight loss, is generally an indication of substantial β cell failure, signalling the need for insulin replacement therapy. Reassessment of diabetes therapy is required when renal or liver function deteriorate, or patients experience cardiovascular events. Also, during certain investigations, such as the use of contrast media, diabetic therapy may require alteration.

Three biguanide drugs (metformin, phenfomrin, buformin) were introduced in the late 1950s. Their origins relate to the glucose-lowering effect of guanidine that was identified in Galega officinalis, a plant used to treat diabetes in traditional herbal medicine. Due to a high incidence of lactic acidosis, phenformin and buformin were withdrawn by the late 1970s. Metformin (Fig. 13.4.2.3) carries negligible risk of lactic acidosis, if appropriately prescribed, and has since become the most used oral antidiabetic agent worldwide.

Metformin lowers blood glucose concentrations without risk of serious hypoglycaemia and without weight gain. This involves several different actions, mostly serving to counter insulin resistance (Table 13.4.2.3). Some of these actions are achieved through enhanced insulin sensitivity, while others are independent of insulin, including activation of adenosine monophosphate-activated protein kinase (AMPK). Metformin does not appear to promote the genomic effects of insulin, and its antidiabetic efficacy requires the presence of at least some circulating insulin (11).

The main glucose-lowering effect of metformin is a reduction in hepatic glucose output, particularly the suppression of gluconeogensis, but with low potency, so that the counterregulatory response is not impeded when glucose levels fall below the normal range (12). Metformin also modestly enhances insulin-stimulated glucose uptake and glycogensis by skeletal muscle, associated with increased deployment of glucose transporters, e.g. the glucose transporter type 4 protein (GLUT4), in the cell membrane (Fig. 13.4.2.4). The effects of metformin contribute to a re-balancing of the glucose-fatty acid cycle, or Randle cycle, to favour glucose utilization. Additionally, anaerobic glucose metabolism, probably due to suppression of respiratory chain activity at complex 1, is increased by the high concentrations of metformin, notably in the walls of the intestine. This increases glucose–lactate turnover, which may contribute to futile cycling and increased energy dissipation that helps to prevent weight gain.

The glucose-lowering efficacy of metformin has been affirmed in many studies, typically reducing HbA_1c by 1–2% (11–22 mmol/mol). The lack of weight gain and low risk of hypoglycaemia have contributed to the general preference for this agent as initial monotherapy, especially in patients who are overweight or more vulnerable to hypoglycaemia. Metformin does not stimulate insulin secretion and generally reduces basal insulin concentrations in hyperinsulinaemic patients. A small improvement in the lipid profile is often seen and reductions in plasma triglyceride and free fatty acids (FFA) are not uncommon. Evidence for a vasoprotective effect of metformin has also contributed to its positioning as initial therapy. Use of metformin reduced the risk of myocardial infarction (by 39% during 10 years) in the UKPDS independently of the glucose-lowering effect, and this benefit persisted during the post-trial follow-up for more than eight years. Metformin has been shown to improve a range of vascular risk markers (e.g. reducing plasminogen activator inhibitor 1 (PAI-1) and increasing fibrinolysis) and surrogate measures such as reducing carotid intima-media thickness and increasing vasoreactivity. Metformin does not appear to affect blood pressure, although a lowering of blood pressure may coincide with reduced body weight (12, 13).

Metformin provides additive efficacy when combined with most other antidiabetic agents, and single-tablet fixed-dose combinations of metformin with several other agents have been developed. When used in conjunction with insulin therapy, metformin can reduce insulin dose requirement, improve the day glucose profile, and reduce hypoglycaemic episodes and weight gain. Hence, metformin is frequently continued when type 2 diabetes patients start insulin therapy.

The main tolerability issue with metformin is development of gastrointestinal symptoms. Diarrhoea may limit dose titration in some patients, although it is usually transient, reduced by temporary dose reduction and by taking after meals. Symptoms can also be reduced by switching to a slow-release formulation, but around 5–15% of patients do not tolerate metformin.

Since metformin is eliminated unchanged in the urine, it is important to check renal function before and at intervals during therapy to avoid drug accumulation, as this may predispose to lactate accumulation (12). The renal exclusion criteria, which vary slightly between countries, are typically set at a serum creatinine greater than 130 µmol/l (1.4 mg/dl), or a creatinine clearance less than 60 ml/min, or an estimated glomerular filtration rate of less than 45 ml/min/1.73 m². Most cases of metformin-associated lactic acidosis are due to failure to respect or recognize deteriorating renal function. Although such cases are rare (with an approximate incidence of 0.03 per 1000 patient years of treatment), about half are fatal. Treatment should be started immediately, usually with intravenous bicarbonate, and haemodialysis may be helpful to remove excess metformin.

While metformin therapy assists in the prevention of cardiovascular events, it is noted that metformin is contraindicated in conditions of hypoxaemia, which include cardiac or respiratory insufficiency, septicaemia, or hypotension. Also, metformin is contraindicated by alcohol abuse, previous history of metabolic acidosis, or severe cirrhotic liver disease. Nevertheless, with suitable caution, metformin can benefit patients with nonalcoholic fatty liver disease (NAFLD). Another use that has emerged for metformin is the treatment of insulin resistance in polycystic ovary syndrome, where it can assist ovulation and conception. Metformin has not been shown to have adverse effects on embryonic or foetal development, and may indeed reduce spontaneous abortion and the risk of maternal gestational diabetes. Since metformin may reduce vitamin B12 absorption, it is advised to check haemoglobin occasionally. Metformin should be stopped temporarily during use of contrast media until normal urine flow returns.

Sulphonylureas were developed in the 1950s following an observation that sulphonamide drugs could cause hypoglycaemia. Since the introduction of sulphonylureas, many members of this class have received extensive use worldwide, and they remain the second most used oral antidiabetic agents (Fig. 13.4.2.5).

The glucose-lowering effect of sulphonylureas is attributed almost entirely to increased insulin secretion resulting from a direct action on the pancreatic β cells. Sulphonylureas bind to the so-called sulphonylurea receptor 1 (SUR1), which is part of a transmembrane protein complex with the ATP-sensitive potassium (K_ATP) channel (14). Binding of a sulphonylurea to SUR1 closes the K_ATP channel. This prevents K+ efflux, depolarizes the membrane, and opens adjacent voltage-dependent long-lasting L-type calcium channels. The resulting increase in cytosolic calcium activates calcium-dependent signalling proteins that regulate the secretion of insulin from preformed granules (Fig. 13.4.2.6).

Sulphonylureas initiate insulin secretion independently of the glucose concentration. Hence, sulphonylureas can increase insulin secretion at all times, including periods when glucose concentrations are low. They therefore predispose to hypoglycaemia. Sulphonylureas may produce a small decrease of glucagon secretion, and there are reports of minor glucose-lowering activity that is independent of effects on the pancreas.

The variety of sulphonylureas (see Table 13.4.2.2, above) with different pharmacokinetic properties enables choice for the onset and duration of action, as well as the mode of metabolism and elimination (15). Typically, sulphonylureas reduce HbA_1c by 1–2% (11–22 mmol/mol) although some desensitization may occur along with disease progression to reduce efficacy during prolonged use. Sulphonylureas are effective as monotherapy, and in combination with other antidiabetic agents (excepting meglitinides). While there is residual β cell function, sulphonylureas can also usefully supplement insulin treatment in type 2 diabetes patients. Thus, by increasing the portal delivery of endogenous insulin, sulphonylureas help to reduce hepatic glucose output, especially during meals: this complements the predominantly peripheral effect of subcutaneously administered insulin. Sulphonylureas have not significantly affected lipids or blood pressure in most studies.

Sulphonylureas are prone to cause weight gain, usually 1–4 kg that stabilizes by about 6 months (16). This is probably due to the anabolic effect of extra insulin and the reduced loss of glucose in the urine. Tolerability is generally good and facilitated by the choice of sulphonylureas available.

Hypoglycaemia is the main serious adverse event associated with sulphonylurea therapy. It is important to start with a low dose and titrate in concert with blood glucose monitoring, harmonizing drug therapy with other aspects of lifestyle. It is also necessary to educate patients to recognize and respond to early warning symptoms of hypoglycaemia, and to prevent hypoglycaemia wherever possible. One or more episodes of hypoglycaemia is likely to occur in up to 20% of sulphonylurea-treated patients annually, although severe episodes occur in only about 1% of patients and the mortality risk has been reported at 0.014–0.033 per 1000 patient years. Severe sulphonylureas-associated hypoglycaemia initially requires intravenous glucose, and glucose supplementation should be continued with appropriate monitoring for more than 24 h to prevent a recurrence, which can occur if longer-acting or more slowly metabolized agents are involved. Glucagon is discouraged in type 2 diabetes patients, as this is itself an insulin secretagogue.

Meglitinides are short-acting (prandial) insulin releasers developed after the observation that the benzamido compound meglitinide, which is a component of some sulphonylurea molecules, can stimulate insulin secretion. Two meglitinide agents (repaglinide and nateglinide) were introduced in the late 1990s/early 2000s (see Fig. 13.4.2.5, above).

Meglitinides bind with the benzamido site on the β cell SUR1 receptor (14), setting in motion the same sequence of events described for stimulation of insulin secretion by sulphonylureas (see Fig. 13.4.2.6, above). The main differences are pharmacokinetic: meglitinides are rapidly absorbed, rapidly acting, but with a shorter duration of action than sulphonylureas, making them suitable for administration with food to enhance insulin secretion to coincide with the period of meal digestion.

The main application of meglitinides is to increase prandial insulin secretion and reduce post-prandial glucose excursions (17, 18). There is usually some carry-over to reduce basal glycaemia, but the reduction in HbA_1c is usually slightly less than with sulphonylureas. However, meglitinides offer the convenience of timing the increase of insulin secretion to coincide with prandial demand. They also reduce inter-prandial insulin concentrations so that the risk of hypoglycaemia is reduced. These agents are appropriate for individuals with irregular lifestyles with unpredictable or missed meals. Meglitinides are conveniently used with metformin or a TZD, and can be used to supplement insulin therapy in type 2 diabetes patients.

Although meglitinides can precipitate hypoglycaemia, such episodes are fewer and less severe than with sulphonylureas. The increase in weight gain is generally less than with sulphonylureas, although there is little noticeable effect when switching between a sulphonylurea and meglitinide, or when combined with metformin. A drug interaction between repaglinide and gemfibrozil should be noted.

TZDs were introduced at the turn of the century with two current members: pioglitazone and rosiglitazone (Fig. 13.4.2.7). A third TZD had been introduced earlier but was withdrawn due to idiosyncratic hepatotoxicity, which is not seen with the current agents.

TZDs produce most of their antidiabetic activity by activation of a nuclear receptor, peroxisome proliferator-activated receptor gamma (PPARγ). This receptor forms a heterodimeric complex with the retinoid X receptor (RXR), and, when a TZD and retinoic acid are bound to the complex, repressors are shed and co-activators recruited (19). The activated receptor locates a nucleotide sequence termed the ‘peroxisome proliferator response element’ within the promoter regions of a range of genes. Many of these genes are insulin sensitive, and others promote complementary effects on glucose and lipid metabolism that improve insulin sensitivity. PPARγ is highly expressed in adipose tissue, and modestly expressed in other key tissues involved in nutrient homeostasis. Stimulation of PPARγ promotes adipogenesis through the differentiation of preadipocytes into new, small, insulin-sensitive adipocytes (Fig. 13.4.2.8). These take up fatty acids, decreasing circulating fatty-acid concentrations and rebalancing the glucose–fatty acid Randle cycle to favour glucose utilization and reduce the availability of fatty acids as a source of energy for hepatic gluconeogensis. TZDs also reduce the accumulation of lipids in muscle and liver, and increase glucose uptake into skeletal muscle and adipose tissue through increased availability of GLUT4.

Additional and diverse (‘pleiotropic’) effects of TZDs include reduced production of several pro-inflammatory cytokines by adipose tissue, notably tumour necrosis factor-alpha (TNFα) (20). TZDs also increase adiponectin production, improve vasoreactivity, and tend to reduce blood pressure, with beneficial effects on a range of cardiovascular risk factors and markers, including a reduction in carotid intima-media thickness.

TZDs exert a slowly generated, glucose-lowering effect, reflecting their predominantly genomic mechanism of action (21). This effect generally requires 2–4 months to achieve full efficacy, which is typically a decrease in HbA_1c of 1% (11 mmol/mol). Thus, dose titration may be a prolonged process, and, since some individuals do not seem to respond to TZDs, an alternative therapy should be considered if no effect is observed in 2–3 months. However, trials have shown the durability of action in responders to extend beyond three years. TZDs do not stimulate insulin secretion, and they do not cause hypoglycaemia. TZDs can be used as monotherapy, or in combination with most other types of agents. They consistently lower fatty-acid concentrations, but have variable effects on other components of the lipid profile: pioglitazone generally reduces triglycerides and often increases HDL. Both TZDs increase the proportion of larger and more buoyant (less atherogenic) LDL particles, and pioglitazone has been shown to reduce long-term cardiovascular disease.

Use of TZDs is typically associated with weight gain, which is often more than that observed with the use of sulphonylureas. This is mainly an increase in subcutaneous adipose tissue, but a proportion may be fluid retention, which should be considered particularly if the weight gain is rapid and marked. Development of oedema after initiation of TZD therapy is usually accompanied by a modest reduction in the circulating haemoglobin concentration which is partly a dilutional effect. Although TZDs reduce several markers of cardiovascular disease, they are contraindicated in individuals with manifest cardiovascular disease (22). TZDs carry a risk of increased heart failure and different countries have excluded their use in patients with New York Heart Association I-IV (Europe) or III-IV (USA). It is advised that liver function be checked and therapy stopped if ALT values exceed 2.5 times ULN. However, TZDs have been used cautiously to treat NAFLD. The improvement of insulin sensitivity with a TZD can restore ovulation in PCOS, but the TZD should be discontinued if pregnancy ensues. Bone fractures are more likely to occur in individuals receiving a TZD, especially post-menopausal women, and individuals with low bone density or osteoporotic disease should be considered as contraindications because TZDs can favour the osteoclast and adipocyte lineages of bone marrow cells development. Rosiglitazone was discontinued in Europe in 2010 because of concerns about adverse cardiovascular outcomes.

Gliptins, or properly dipeptidyl-peptidase-4 (DPP4) inhibitors, were recently introduced and increase the endogenous incretin effect. The incretin effect refers to the enhancement of glucose-stimulated insulin secretion by hormones released from the intestinal tract during meal digestion. The main incretin hormones are GLP-1 and gastric inhibitory polypeptide (or glucose-dependent insulinotropic polypeptide; GIP). These hormones are rapidly degraded by the enzyme DPP4, hence the use of DPP4 inhibitors to enhance the endogenous incretin effect. Agents available at the time of writing in this category are sitagliptin, vildagliptin, and saxagliptin (Fig. 13.4.2.9).

Passage of food through the intestine stimulates the release of GIP from K cells, located mainly in the upper small intestine, and GLP-1 from L cells, located mostly in the ileum (23). These incretin (or entero-insular) hormones act on the pancreatic β cells to enhance the prandial insulin response. They interact with separate G-protein-linked receptors on the β cells to potentiate distal steps in the insulin secretion pathway: they also act via protein kinase A to promote insulin biosynthesis. Other effects of incretins on nutrient homoeostasis are noted in Table 13.4.2.4; in particular, GLP-1 acts on pancreatic α cells to reduce excess glucagon secretion, slow gastric emptying, and augment the meal-induced satiety effect. Both GLP-1 and GIP have been shown to increase β cell mass in animal models, associated with increased neogenesis and reduced apoptosis of β cells, but this has yet to be confirmed in human type 2 diabetes. Possible effects of GLP-1 on the cardiovascular system are currently under investigation.

The meal-stimulated release of GLP-1 appears to be reduced in type 2 diabetes, although the capacity of GLP-1 to potentiate nutrient-induced insulin secretion is retained, provided there is adequate functional β cell mass remaining. Conversely, GIP levels appear to be largely maintained in type 2 diabetes, but the insulin-releasing effect is reduced (24). Hence, the administration of GLP-1, with its retained portfolio of effects in type 2 diabetes, should, in principle, provide a favourable therapeutic approach. However, the rapid degradation of incretins in the circulation (half-life, < 2 min) by the enzyme DPP4 makes this impracticable. DPP4 breaks the N-terminal dipeptide where the second aminoacid is an alanine or proline residue (GLP-1 and GIP each have an alanine residue in this position; Fig. 13.4.2.10). Specific inhibitors of DPP4 were developed to prevent this degradation and prolong the activity of endogenous incretins (25).

Each of the gliptins produces almost complete inhibition of DPP4 activity for at least 12 h, increasing active endogenous incretin concentrations two- to three-fold (26). As monotherapy, this considerably reduces post-prandial glucose concentrations (by about 3 mmol/l or 54 mg/dl) with a modest carry-over to reduce basal glycaemia (by about 1–1.5 mmol/l or 18–27 mg/dl). This typically achieves a reduction in HbA_1c of 0.7–1% (8–11 mmol/mol). Because the increase in insulin secretion is glucose dependent, the extra stimulation of insulin secretion does not occur when glucose levels fall to normal basal values, thereby reducing the risk of interprandial hypoglycaemia. Indeed, monotherapy with gliptins is unlikely to cause severe hypoglycaemia. Gliptins do not appear to reduce the rate of gastric emptying to a clinically significant extent (not sufficiently to cause nausea), and they do not produce a clinically demonstrable satiety effect, but it is evident that gliptins do not cause weight gain and they may assist a small amount of weight loss. Gliptins can be combined with agents that improve insulin sensitivity to give additive efficacy. They can also be used with either a sulphonlyurea or meglitinide to give increased insulin secretion, since incretins act via a different cellular mechanism on the β cell—although, when used in this combination, there is a risk of hypoglycaemia.

There are pharmacokinetic differences between the gliptins that alter their suitability for patients with renal or liver disease, and some minor drug interactions have been noted, but no substantive adverse effects have been reported. It is appreciated that many circulating peptides are degraded by DPP4, including bradykinin, encephalins, neuropeptide Y (NPY), gastrin-releasing polypeptide (GRP), substance P, and monocyte chemoattractant protein-1 (MCP-1). Despite the potential to exert effects on satiety, gastrointestinal motility, and vascular activity, no clinically significant effects of gliptins have been reported. The DPP4 enzyme is also the CD26 T-cell activator, but evidence to date suggests that small molecules that inhibit the dipeptidase activity do not interfere with the immune function of the molecule, and gliptins have not been reported to exert any immunological effects.

The therapeutic exploitation of GLP-1 is compromised by its rapid degradation via DPP4 (25, 27). To circumvent this problem, various modifications of the GLP-1 molecule have been explored, particularly to prevent the action of DPP4. Exendin-4 (exenatide; see Fig. 13.4.2.10, above) was identified in the saliva of the Gila monster Heloderma suspectum. Exendin-4 has 53% homology with GLP-1 and acted as a GLP-1 receptor agonist. It was resistant to DPP4, giving a half-life of more than 2 h, sufficient to enable a therapeutic effect for 4–6 h. Exenatide was introduced in 2005 as a twice-daily, subcutaneously injected peptide. Another GLP-1 receptor agonist, liraglutide, was introduced in 2009 as a once-daily subcutaneous injection (26).

GLP-1 receptor agonists produce the profile of effects described above for GLP-1, and summarized in Table 13.4.2.6. Exenatide is injected before a main meal: it increases post-prandial glucose-induced insulin secretion, reduces post-prandial hyperglucagonaemia in type 2 diabetes, and slows gastric emptying. It also appears to exert sufficient satiety activity that weight loss is a common feature of exenatide therapy. A GLP-1 analogue for once daily administration, liraglutide, is GLP-1 (7–37) with Lys34 replaced with Arg34, and Lys26 attached via a glutamate residue to a cetyl alcohol (C16; 1-hexadecanoyl, palmitoyl alcohol) fatty acid chain. The fatty acid attaches the GLP-1 analogue to circulating albumin, which protects it from degradation by DPP4 and prolongs the half-life to 11–15 h.

GLP-1 receptor agonists reduce HbA_1c by about 1% (11 mmol/mol) or more, associated with substantial reductions (about 4 mmol/l or 72 mg/dl) in post-prandial glycaemia. Durability of the glucose-lowering effect has been observed over two years, but it has yet to be established whether the effects on β cell preservation seen in animals are replicated in human type 2 diabetes. There is some debate at this time as to whether the pharmacokinetic differences associated with twice-daily versus once-daily administration of a GLP-1 analogue will alter the pharmacodynamic effects, but, in very general terms, the two preparations have achieved similar reductions in hyperglycaemia associated with similar reductions in body weight of about 3 kg over 6–12 months. GLP-1 analogues can be combined with any of the other therapies, providing there is adequate β cell function or α cell dysfunction, but they are unlikely to be used with gliptin, since the amount of an injected GLP-1 analogue entering the circulation is much greater than the extra endogenous hormone levels achieved by inhibition of DPP4.

A limiting factor for the use of GLP-1 analogues is initial nausea, presumed to reflect a reduced rate of gastric emptying. This is usually transient and ameliorated by introducing therapy at a low dose for several weeks. Administration to patients with severe gastrointestinal disease, including gastroparesis, should be avoided. GLP-1 analogues carry little risk of severe hypoglycaemia when used as monotherapy, but a risk of hypoglycaemia should be appreciated when used in combination with other types of insulin-releasing agents. Exenatide is mostly cleared by renal proteolysis and glomerular filtration, and a dose reduction or avoidance should be considered in patients with moderate to severe renal disease. It is advised to discontinue these agents in pregnancy. Although some patients develop antibodies to exenatide, these do not usually have a noticeable effect on efficacy, and reactions at the injection site are seldom problematic. Reports of patients developing pancreatitis on exenatide have not been specifically attributable to the drug, and appear to be no more common than in type 2 diabetes patients treated with other therapies. Some adverse effects of liraglutide on thyroid C-cells in animal models have not been seen during clinical studies.

Following evidence that metabolites from cultures of actinomycete fungi could inhibit cell surface glucosidase enzymes, specific AGIs were developed as antidiabetic drugs. The first, acarbose, was introduced in the early 1990s, followed by miglitol. Another, voglibose, is available in some countries (Fig. 13.4.2.11).

AGIs bind to α-glucosidase enzymes with high affinity, acting competitively to prevent the binding and cleavage of disaccharides and oligosaccaharides into monosaccharides (28). This impedes the final step of carbohydrate digestion, resulting in a delayed and slower appearance of glucose in the circulation after a meal rich in complex carbohydrate. AGIs do not specifically affect the glucose absorption process, and they can only exert a clinically significant effect on the post-prandial glycaemic excursion when there is substantial complex carbohydrate in the diet. The activity profiles of acarbose and miglitol vary slightly: acarbose has a greater affinity for glycoamylase than sucrase, whereas miglitol has a stronger inhibitory effect on sucrase.

AGIs act mainly to reduce post-prandial hyperglycaemia, and their effects are generally modest with a reduction in HbA_1c of about 0.5% (5 mmol/mol), although this can be greater in individuals consuming mainly a carbohydrate-rich diet. Usefully, an AGI can be added in to any other therapy, and this is not usually associated with risk of hypoglycaemia. Indeed, by extending the duration of meal digestion, AGIs can reduce the risk of inter-prandial hypoglycaemia in individuals receiving insulin or an insulin initiator. Also, AGIs do not cause weight gain, and some individuals may show a reduced post-prandial triglyceride profile. It has been suggested that, by extending carbohydrate digestion to more distal regions of the intestine, AGIs might increase GLP-1 secretion. However, postprandial insulin concentrations are commonly reduced by an AGI, commensurate with the lowering of post-prandial glycaemia. There have been reports of fewer cardiovascular events in patients receiving an AGI, but it is unclear whether this is accounted for by the reduced post-prandial hyperglycaemia or some other effect of the therapy.

AGIs are prone to cause some carbohydrate malabsorption. Undigested carbohydrate passing into the large bowel is fermented and can create considerable flatulence. Thus, AGIs should be given with appropriate meals and titrated slowly to minimize this effect. Individuals with gastrointestinal conditions are contraindicated: caution is implied with any agents affecting gut motility, and liver function should be checked in individuals receiving a high dose of acarbose, since increased alanine transaminase levels have been noted very occasionally.

Amylin (islet amyloid polypeptide, IAPP) is synthesized and co-secreted with insulin from the pancreatic β cells. Precipitates of IAPP within the islets in type 2 diabetes have been ascribed a pathogenic role in the demise of β cells, although the extent of involvement remains uncertain. While examining the actions of amylin, it became evident that this peptide exerts central effects that independently affect nutrient metabolism. To retain these effects without any detrimental effects on the islets, a non-precipitating amylin analogue (pramlintide) was developed and introduced in a few countries (29).

Pramlintide (Fig. 13.4.2.12) acts centrally to complement the effects of insulin in the control of post-prandial glucose homeostasis. It acts predominantly via the area postrema in the brain stem, which communicates with the hypothalamus and activates neural pathways to suppress glucagon secretion by pancreatic α cells. By reducing glucagon, pramlintide reduces hepatic glucose output. Additionally, via a central route, pramlintide decreases the rate of gastric emptying and reduces the secretion of gastric juice, which slows the rate of digestion with resultant slowing of nutrient absorption. Pramlintide also acts centrally to reduce food intake, which may, in the long term, assist weight control.

Pramlintide is typically used to reduce the insulin dose and prevent the weight gain associated with higher doses of insulin, while improving glycaemic control. It is injected before meals in patients with type 1 or type 2 diabetes who are already receiving insulin therapy, and has been shown to improve the glycaemic profile with reductions in HbA_1c of about 0.3–0.6% (3–7 mmol/mol) and reductions in body weight of 1–2 kg. Since these effects of pramlintide are usually achieved with a reduction in the insulin dose, it is generally advised to reduce the pre-meal short-acting insulin dose by about half when initiating pramlintide therapy, to avoid risk of hypoglycaemia. The pH difference between insulin and pramlintide precludes the combination of these agents in the same syringe.

The most common side effect of pramlintide is nausea, which is usually transient and minimized by gradual dose titration. Since pramlintide is used with insulin, it increases the risk of hypoglycaemia unless appropriately titrated in conjunction with suitable meals and a reduced insulin dose. A drawback to the use of pramlintide is the need for an injection before each main meal, which is additional to the injection required for insulin therapy. Antibodies to pramlintide have been detected in some patients, but this does not appear to have interfered with efficacy.

Bromocriptine received an indication as an anti-diabetic therapy in the USA in 2009. Its glucose-lowering potential had been appreciated through studies decades earlier, and through experience during use in the treatment of Parkinsonism and pituitary tumours.

Bromocriptine is a dopamine D2 receptor agonist that lowers glucose concentrations without stimulating insulin secretion (30). However, its exact mode of action as an antihyperglycaemic agent is not established.

Trials for regulatory approval have shown that bromocriptine reduces HbA_1c by about 0.5–0.6% (5–7 mmol/mol). It can be used as monotherapy or in combination with other oral agents, is unlikely to cause severe hypoglycaemia and is not associated with weight gain.

Experience with bromocriptine during use for other indications suggests that risk of respiratory and pericardial fibrosis, hypotension, and exacerbation of psychotic disorders should be borne in mind, and appropriate monitoring should be undertaken. Also, interactions can occur with dopamine antagonist therapy, drugs that are highly protein bound, and other drugs that are metabolized by or induce the P450 isoform CYP3A4 (cytochrome P450, family 3, subfamily A, polypeptide 4). Use in pregnancy is not recommended.

Sodium-glucose transporter-2 (SGLT-2) is the main transporter that retrieves glucose from the proximal convoluted tubules in the kidneys, preventing glycosuria (31). By specifically inhibiting this class of transporters, it is possible to reduce hyperglycaemia by increasing urinary glucose elimination. SGLT-2 inhibitors are presently advanced in development and have shown promise as adjuncts to other antidiabetic therapies. They can assist weight loss and do not prevent the counterregulatory mechanisms to prevent hypoglycaemia. The main potential adverse effects are fluid depletion and urogenital infections.

Due to the progressive nature of type 2 diabetes, many individuals require insulin to maintain glycaemic control over time, as other anti-diabetic agents are unable to either achieve or maintain glycaemic targets. Data from the UKPDS suggest that 53% of patients will require insulin six years after diagnosis, and 75% of patients will need multiple treatments after nine years (32). Although insulin treatment is very effective in achieving glycaemic control, its use is invariably associated with weight gain and increased risk of hypoglycaemia (33).

The term ‘prandial’ insulin refers to short-acting (regular) or rapid-acting insulin that is injected subcutaneously prior to eating a meal. This serves to control the excursion in blood glucose resulting predominantly from the carbohydrate content of that meal.

Basal insulins are intermediate or long-acting formulations that provide a background level of insulin. They are administered either once or twice daily to control fasting and pre-prandial blood glucose levels.

Animal insulins, made from the pancreatic extracts of cattle and pigs, have been available since shortly after the introduction of insulin in the 1920s. Most of the problems encountered with these insulins are due to impurities, notably allergic skin reactions at injection sites, other immunogenicity problems, and variable rates of absorption and action, sometimes associated with reduced efficacy over time. Animal insulin has been progressively phased out since the advent of human insulin.

Human insulin is synthetic ‘human’ insulin manufactured from recombinant DNA technology. The kinetics of human insulin have important limitations when injected subcutaneously. Short-acting (soluble, neutral) human insulin has a delayed onset of action of around 20–30 min, which means that it needs to be injected at least 30 min before a meal for optimal effect (Table 13.4.2.5). The prolonged duration of action of 6–8 h, and variability in absorption, increases the risk of hypoglycaemia. Basal human insulins (formulations containing protamine and/or zinc, such as neutral protamine Hagedorn (NPH), lente, and ultralente) have a variable onset of action. The peak of activity is between 4–10 h, with a duration of action less than 24 h.

The requirement for insulin formulations that more closely mimic the normal daily pattern of endogenous insulin secretion has prompted the development of so-called insulin analogues. These are grouped under the categories quick-acting and long-acting analogues, and are produced by recombinant DNA technology using either Escherichia coli or yeast (Table 13.4.2.5).

Rapid-acting analogues are formed by minor changes to the amino acid sequence that produce subtle, spatial alterations in the conformation of the insulin molecule, which results in rapid dissociation and formation of stable monomers after subcutaneous injection, allowing rapid absorption.

There are three types of rapid-acting insulin analogues: lispro, aspart, and glulisine. Insulin lispro was the first to be introduced in 1996. Lispro is produced by reversal of amino acid positions: proline at position 28 and lysine at position 29 on the insulin B-chain. It is available in biphasic formulations (see Table 13.4.2.5). Insulin aspart is produced by replacing proline at position 28 on the B-chain of insulin with aspartic acid, and is also available in biphasic formulations (Table 13.4.2.5). Glulisine is synthesized by replacing asparagine with lysine at B3 and glutamic acid for lysine at position B29.

The rapid-acting analogues share very similar pharmacokinetic and pharmacodynamic properties. They are absorbed within 10–15 min of a subcutaneous injection, peak circulating concentrations occur within 30–90 min, and they have a duration of action of 4–6 h. This more closely mimics normal physiological prandial insulin release, which effectively reduces the waiting period from the time of injection to the ingestion of a meal. Individuals can, therefore, inject immediately (or ideally up to 10 min) before eating, and this offers more flexibility and convenience. In patients with variable food intake, the opportunity to inject during the meal offers an advantage but the evidence suggests that post-meal injection must be within 15 min of starting the meal, and the glucose control then approximates a conventional soluble insulin taken immediately before.

Comparison of an analogue combination of lispro/glargine versus the human insulin combination of soluble (regular) insulin/NPH in type 2 diabetes showed that the analogue combination achieved lower post-prandial glucose levels with significantly lower insulin doses and fewer episodes of nocturnal hypoglycaemia. Insulin lispro, aspart, and glulisine have been shown to be effective at lowering post-prandial glucose when added to oral antidiabetic agents. In one study, significant reductions in post-meal glucose levels were observed with mealtime lispro (plus bedtime NPH), compared to mealtime regular human insulin plus NPH (34). However, a meta-analysis that compared rapid-acting analogues with regular insulin and showed only a small beneficial effect on glycaemic control in patients with type 1 diabetes, whereas, in patients with type 2 diabetes, no such improvements were observed (35).

Long-acting insulin analogues are produced by modifications to the insulin molecule that either promote association into small crystals after subcutaneous injection (insulin glargine) or enable attachment to albumin in the circulation (insulin detemir), giving rise to a longer duration of action (Table 13.4.2.5).

Insulin glargine is produced by three amino acid modifications, including addition of two basic amino acids at the C-terminus of the beta-chain. These modifications alter the molecule’s isoelectric point such that it forms micro-precipitates after subcutaneous injection into the slightly alkaline interstitial fluid. This results in slow release from the injection site into the circulation. Glargine is a clear insulin that is maintained in the vial/cartridge at slightly acid pH, and so it cannot be mixed with other insulins. Insulin detemir is produced with a fatty acid (myristic acid) attached to the lysine residue at position B-29. This facilitates self-association and reversible binding to albumin binding, which prolongs the duration of action.

Glargine has a ‘peakless’ profile of action lasting up to about 24 h. Detemir has a mean duration of action of about 20 h after administration of a 0.4 IU/kg dose. Although twice-daily dosing of detemir has been was used in many of the early clinical trials, recent data in patients with type 2 diabetes suggest that many patients require only once-daily dosing (36, 37).

Glargine used as a once daily basal injection type 2 diabetes has been demonstrated to achieve good glycaemic control with marginal weight gain in most patients. In patients with a high (> 9.5% or 80 mmol/mol) HbA_1c, insulin glargine was more effective than triple oral antidiabetic therapy (38). In comparison to NPH, glargine provides at least comparable glycaemic control, with a reduced incidence of hypoglycaemia (39). Several studies have noted that, although the proportion of patients achieving an HbA_1c of 7.0% (53 mmol/mol) or less was similar for glargine and NPH, rates of hypoglycaemia (especially nocturnal) were significantly reduced with glargine (40).

In terms of its effects on fasting blood glucose, detemir was found to exhibit more consistent values, when compared to NPH. In the Treat-to-Target study, twice-daily detemir was added to oral antidiabetic agents for patients with type 2 diabetes with suboptimal control, and similar target HbA_1c levels (< 7% or 53 mmol/mol) were achieved, with fewer hypoglycaemic events in more patients on detemir than in those receiving NPH (41). Studies with detemir have shown that the reduction in intrapatient variability of fasting glucose levels is a major contributor to the reduced risk of hypoglycaemia with detemir, relative to NPH (42). However, a review of long-term trials in patients with type 2 diabetes comparing long-acting analogues to NPH insulin showed only a ‘theoretical advantage’ of improved metabolic control, stating that clinical benefits were related to reduced nocturnal hypoglycaemia.

Both glargine and detemir have shown similar efficacy in terms of HbA_1c reduction, but a higher proportion of individuals need twice-daily dosing with detemir (38). The cost–benefit ratio of these analogues has yet to be established, and this caution is reflected in the UK National Institute for Clinical Excellence (NICE) guidelines, which recommend the use of insulin glargine as an option for people with type 1 diabetes, but not routinely for those with type 2 diabetes, except in certain groups, such as those who require assistance to administer insulin injections, and in patients whose lifestyle is restricted by recurrent symptomatic hypoglycaemia (6).

Most international diabetes guidelines recommend that optimal targets in type 2 diabetes patients are less than 7% (53 mmol/mol) or less than 6.5% (48 mmol/mol) (4–7). In the UKPDS, achieving a target HbA_1c of less than 7% (53 mmol/mol) was associated with a significant reduction in long-term microvascular complications (3, 33). In clinical practice, insulin therapy in type 2 diabetes is still usually initiated too late after use of combinations of oral antidiabetic agents at suboptimal HbA_1c levels. It is incumbent on health care professionals to ensure that insulin is initiated in a timely manner. The selected regimen should be discussed with the patient and their carers, and the insulin dose titrated to achieve optimal glycaemic targets appropriate to the circumstances of the individual patient. Hypoglycaemia and weight gain remain the main barriers to optimizing insulin therapy.

The two most common insulin regimens used in type 2 diabetes (with or without concomitant oral antidiabetic therapy) are basal only and twice-daily biphasic (pre-mixed) insulin regimens (Table 13.4.2.6). Prandial insulin, basal–bolus regimens (or multiple daily injection (MDI) and continuous subcutaneous insulin infusion (CSII)) are less frequently used in type 2 diabetes patients.

The combination of insulin and oral anti-diabetic agents is a useful step when glycaemic control is not achieved with oral agents alone. When dual oral anti-diabetic therapy fails to control hyperglycaemia, the decision to persist with a third oral agent or to initiate insulin is down to clinician preference and circumstances of the patient. Whilst oral agents such as TZDs and gliptins (DPP4 inhibitors) are useful in some patients as the third therapy, insulin can be added to existing therapy as an alternative option. In a study exploring the option of adding either insulin glargine or another oral agent (rosiglitazone) to existing sulphonylurea and metformin therapy, insulin glargine was more cost effective, caused less weight gain, and showed greater reductions in HbA_1c when baseline HbA_1c was greater than 9.5% (80 mmol/mol) (38).

The Treating to Target in Type 2 Diabetes (‘4T’) study was a three-year study of 708 type 2 patients poorly controlled with metformin and a sulphonylurea. These patients were randomized to receive either basal insulin detemir once (or twice if indicated) daily, biphasic insulin aspart twice daily, or prandial insulin aspart three times daily. Interim analysis at the first year noted that the decrease in HbA_1c was significantly greater in the biphasic and prandial groups (7.3% (56 mmol/mol) and 7.2% (55 mmol/mol)) than the basal group (7.6% (60 mmol/mol), p < 0.001), although hypoglycaemia and weight gain were lower with basal insulin. Prandial insulin was associated with a two-fold increase in hypoglycaemic events and a 21% increase in weight. Thus, whilst it is acknowledged that most patients are likely to need more than one type of insulin to achieve target glucose levels, the findings from this and other studies indicate that a basal insulin is a useful first-line add-on to oral antidiabetic agents (36). In the follow-up study, best incremental control was achieved in the group in which prandial insulin was added to basal replacement to maintain control (43).

Intermediate-acting insulins (e.g. NPH) are commonly used in type 2 diabetes, although a number of studies support the use of long-acting analogues (glargine or detemir). In the Treat-to-Target Trial, a single bedtime injection (NPH or glargine) was added to existing oral agents and insulin doses were titrated to a fasting glucose of 5.6 mmol/l (100 mg/dl). The mean HbA_1c declined from 8.6% (70 mmol/mol) at baseline to less than 7% (53 mmol/mol) in both groups in six months; however, less hypoglycaemia (especially nocturnal) was seen in the glargine group (44). The basal analogues offer equal efficacy in terms of glycaemic control as shown in a direct comparison in patients with type 2 diabetes. In a 52-week, multi-national, randomized, controlled trial, insulin-naïve type 2 diabetes patients were randomized to either detemir or glargine once daily, and active dose titration was undertaken to achieve fasting glucose targets. An additional dosing for detemir was permitted. HbA_1c decreased by 1.5% (16 mmol/mol) with both insulins, and 52% participants achieved an HbA_1c of 7% (53 mmol/mol) or less. Weight gain was lower with once-daily detemir, but was comparable between glargine once daily and detemir twice daily. Rates of hypoglycaemia were similar, but higher insulin doses and more injections were needed with detemir to achieve targets and 55% of participants in the detemir arm required twice-daily dosing (37).

Twice-daily biphasic (premixed) insulin regimens can, potentially, target both fasting and post-prandial hyperglycaemia. Using this approach, Raskin et al. compared the efficacy of a biphasic 30/70 (European nomenclature, with short-acting component given first) insulin aspart twice daily or glargine once daily added to metformin ± a TZD in type 2 diabetes (45). At 28 weeks, a significant reduction in HbA_1c was seen in both groups, but greater reductions were seen with the biphasic insulin v. the glargine group (–2.79 ± 0.11 v. –2.36 ± 0.11% or −30±1 vs −26±1 mmol/mol, respectively; p < 0.01) (46). Similar to other studies, weight gain and hypoglycaemia were greater with biphasic insulin, though it can be argued that the exclusion of a sulphonylurea in the glargine arm may have influenced outcomes. However, as seen in these studies, when glucose control deterioriates, basal insulin by itself may be insufficient to achieve tighter glycaemic targets and a biphasic regimen may be more effective.

Prandial insulins with or without the continuation of oral agents can be used to address post-prandial hyperglycaemia, with distinct improvements in HbA_1c levels. In the APOLLO study, the addition of once-daily glargine or prandial insulin in patients poorly controlled with oral agents was equally effective (46). However, basal insulin therapy is usually considered as a safer and simpler option, and is often more acceptable to patients due to a lower incidence of hypoglycaemia, fewer injections, and less weight gain than with prandial insulin therapy (36).

Early initiation of an intensive regimen has been shown to reduce the impact of glucotoxicity and might preserve β cell function. A basal–bolus regimen mimics more closely the normal physiological insulin profile, and has demonstrated good glycaemic control in clinical studies. The efficacy of a basal–bolus regimen versus biphasic insulin was compared in patients with a baseline HbA_1c of 9% (75 mmol/mol) who were previously treated with insulin glargine plus oral antidiabetic agents. At 24 weeks, HbA_1c was significantly reduced in both groups, being marginally lower with a basal–bolus regimen (6.78 vs. 6.95% or 51 vs 53 mmol/mol, p = 0.021). However, a basal–bolus (otherwise referred to as multiple daily injection or MDI regimen) is not considered a practical insulin initiation regimen for the majority of patients with type 2 diabetes.

CSII remains an option in patients who have poor glycaemic control on basal–bolus regimens (or MDI). Studies using CSII in type 2 diabetes have shown comparable glycaemic control versus MDI, but with improved patient satisfaction. However, CSII use is largely confined to those with type 1 diabetes.

It is increasingly recognized that early initiation and subsequent titration of insulin therapy, together with patient education, plays a key role in achieving glycaemic targets and long-term maintenance of glycaemic control in type 2 diabetes. Initiating insulin therapy can be difficult due to a number of barriers as outlined previously. With diabetes care becoming more community based in many countries, it is essential to explore practical options to help primary care and community physicians to deal with the increasing numbers of patients requiring insulin therapy.

Insulin doses can be initiated and titrated using a simple protocol that is easy to follow. For example, basal insulin dosing with glargine is usually started at 10 units or 0.1 to 0.2 units per kg of body weight. This is titrated by 2 units at 3-day intervals until fasting glucose targets (4–6 mmol/l) are achieved with no manifest hypoglycaemia. A larger dose increment of 4 units is advised if the fasting glucose readings remains 10 mmol/l (180 mg/dl) or more. Most patients are eventually likely to require insulin doses, around 0.5 to 1.0 unit/kg. The 4-T study, using insulin determir as the basal insulin, operated a predefined algorithm for insulin titration and a patient-specific insulin starting dose. This study found that a higher starting insulin dose (2–76 IU/day) did not result in severe hypoglycaemia (36).

In the AT-LANTUS study, two insulin titration algorithms were compared in patients with suboptimally controlled type 2 diabetes (47). Algorithm 1 was physician led and involved weekly interventions, whereas algorithm 2 was patient led with adjustments made after every three days. After 24 weeks, there was a significantly greater HbA_1c reduction with algorithm 2 without a significant difference in the incidence of severe hypoglycaemia between the groups. These findings showed that such a regimen can be effectively commenced in the community and in secondary (hospital-based) care. Insulin initiation in groups involves less time and is a cost-effective option, and should be considered both in primary and secondary care.

Insulin analogues and new approaches to initiation and intensification have provided the impetus to make insulin therapy simpler to use, allowing patients with diabetes to have a more flexible lifestyle. Quality of life, hypoglycaemia, and weight gain are important considerations when moving patients from conventional to intensive regimens. This needs open and frank discussions with patients and carers. Insulin regimens can be easy to initiate and titrate with a low risk of hypoglycaemia, particularly with basal insulins. Patient education should form the central plank of our management strategy, empowering patients with diabetes with the requisite skills and knowledge to self-manage their condition more effectively. This is, indeed, more important than the insulin regimen or the specific type of insulin chosen.

Hypoglycaemia is a recognized issue for all individuals receiving antidiabetic therapy, especially those who have achieved a near-normal HbA_1c and/or take insulin or insulin-releasing agents. Unawareness of early hypoglycaemic symptoms is uncommon in type 2, but constitutes an important consideration if patients wish to continue driving or undertake similar attention-dependent activities. The elderly, those living alone, and people with irregular or neglectful lifestyles are also particularly vulnerable to hypoglycaemia, and require appropriate selection of agents, doses, and administration regimens to minimize risk.

Impaired renal function presents a contraindication for antidiabetic drugs that are largely eliminated renally, e.g. metformin. Some other renally eliminated agents can be given in reduced dose, but it may be preferable to use agents that are inactivated by the liver and eliminated in the bile in order to circumvent reliance on the kidneys (Table 13.4.2.2). Conversely, agents metabolized predominantly by the liver should be considered with caution in individuals with impaired liver function or a large pill burden that includes agents metabolized by, or otherwise affecting, the same pathways of hepatic metabolism.

Substantive cardiovascular disease is an exclusion for several oral agents: TZDs are contraindicated where there is evidence of heart failure, and metformin is contraindicated for any hypoxaemic state. In individuals who become pregnant whilst taking oral antidiabetic therapy, it is generally recommended to switch to insulin, although there is no evidence that early embryonic development is adversely affected. There is some evidence that pregnancy outcomes are improved with appropriate continuation of metformin, and metformin may also assist individuals who develop gestational diabetes, provided that there are no pre-existing contraindications.

Managing the progressive and variable natural history of type 2 diabetes with emergent complications and co-morbidities continues to present a formidable therapeutic challenge. Diet and other lifestyle measures are fundamental throughout, supplemented with a series of oral and injectable anti-diabetic agents, as appropriate, eventually requiring insulin in many patients. Individualization and flexibility of therapy to suit the particular needs and circumstances of the patient are important to the management process, and must be complemented with adequate education and empowerment.