Endocrine system

Diabetes mellitus 582

• Physiological principles of glucose and

insulin metabolism 582

• Epidemiology and classification 587

• Aetiology and pathogenesis 589

• Natural history 591

• Clinical features 593

• Complications 593

• Management 604

• Monitoring 628

Thyroid disease 630

• Physiological principles 630

• Hypothyroidism 633

• Hyperthyroidism 637

• References and further reading 643

Endocrine control of physiological functions represents broadly targeted, slow acting but funda-

mental means of homeostatic control, as opposed to the rapidly reacting nervous system. In

endocrine disease there is usually either an excess or a lack of a systemic hormonal mediator,

but the cause may be at one of a number of stages in the endocrine pathway. Thyroid disease

and diabetes mellitus represent contrasting extremes of endocrine disease and its management.

Diabetes is one of the most serious and probably the most common of multisystem diseases.

Optimal control of diabetes requires day-to-day monitoring, and small variations in medication

dose or patient activity can destabilize the condition. Therapy requires regular review and

possible modification. Furthermore, long-term complications of diabetes cause considerable

morbidity and mortality.

Thyroid disease is a disorder of thyroid hormone production that has, compared to diabetes,

equally profound overall effects on metabolic and physiological function. However, it causes few

acute problems and has far fewer chronic complications. Moreover, management is much easier,

requiring less intensive monitoring and few dose changes. Furthermore, control is rarely

disturbed by short-term variations in patient behaviour.

9Endocrine system

Diabetes mellitus is primarily a disorder ofcarbohydrate metabolism yet the metabolicproblems in properly treated diabetes are notusually troublesome and are relatively easy tocontrol. It is the long-term complications ofdiabetes that are the main causes of morbidityand mortality. People with diabetes suffer farmore from cardiovascular and renal diseasethan other people, and diabetes is the principalcause of acquired blindness in the West. Mostpeople with diabetes do not die from metaboliccrises such as ketoacidosis but from stroke, MIor chronic renal failure.

Diabetes is associated with obesity and lack ofexercise, and the steady increase in prevalencein the West is being reproduced in large parts ofthe developing world as they adopt that life-style. Diabetes is in danger of becoming almostpandemic. Particularly worrying is the rise in theincidence of diabetes of both types in everyounger patients. This threatens to put an intol-erable strain on health services, particularly indeveloping countries.

Physiological principles of glucose andinsulin metabolism

Insulin action

Insulin is the body’s principal anabolichormone. It expands energy stores during timesof adequate nutrition against times of foodshortage. Opposing this action are several cata-bolic ‘counter-regulatory’ or ‘stress’ hormonesthat mobilize glucose for use when increasedenergy expenditure is necessary. The mostimportant of these are adrenaline (epinephrine),corticosteroids, glucagon, growth hormone andgrowth factors. These two opposing systemswork in harmony to maintain glucose home-ostasis. Insulin also enhances amino acid utiliza-tion and protein synthesis, the latter actionbeing shared with growth hormone.

Insulin action has three main components(Figure 9.1):

• Rapid: in certain tissues (e.g. muscle), insulinfacilitates the active transport of glucose andamino acids across cell membranes,enhancing uptake from the blood.

• Intermediate: within all cells, insulinpromotes the action of enzymes that convertglucose, fatty acids and amino acids intomore complex, more stable storage forms.

• Long-term: because of increased proteinsynthesis, growth is promoted.

One important consequence is the prompt(though not complete) clearance of glucose fromthe blood after meals. Glucose would otherwisebe lost in the urine because of the kidney’slimited capacity for reabsorbing glucose filteredat the glomerulus.

Glucose transport

Glucose uptake into cells across the cellmembrane is dependent on the concentrationgradient between the extracellular medium (e.g.blood plasma, gastrointestinal contents) and the cell interior. However, because glucose issuch an important metabolite, there exist anumber of membrane transport pumps or facili-tators in certain tissues. There are special insulin-

582 Chapter 9 • Endocrine system

absorptioninto plasma

PlasmaGLUCOSE

PlasmaAMINOACID

typical body cell

PROTEIN

AMINO ACID

GLUCOSE

Energy

GLYCOGEN

FAT

DIET

Figure 9.1 Simplified scheme showing the anabolicactions of insulin. Insulin aids the uptake of metabolites intobody cells and enhances the action of enzymes that utilizethem as precursors to synthesize more complex molecules.Note: not all actions shown occur in all body cells.

Diabetes mellitus

independent sodium-dependent transporters(SGLT) for uptake from the GIT into intestinalcells and a variety of insulin-dependent andinsulin-independent glucose transporters (GLUT)for most other tissues or organs (Table 9.1).

In muscle and adipose tissue the transporterdepends on an insulin-requiring active pump forglucose uptake, so insulin deficiency deprivesthem of glucose. Other cells, particularly in theliver, brain, kidney and GIT, do not absolutelyrequire insulin for glucose uptake, but diffusionis nevertheless facilitated by it. In the liver,enhanced phosphorylation of glucose drivesintracellular concentrations down, encouraginguptake. Insulin lack does not deprive tissues suchas these of glucose; on the contrary, the hyper-glycaemia associated with diabetes can produceintracellular glucose overload, and this may beresponsible for some diabetic complications (p.593). This is particularly relevant to tissues suchas nerves, which are freely permeable to glucose.

Insulin also facilitates the uptake of aminoacids into liver and muscle, and of potassiuminto most cells. This latter effect is exploitedtherapeutically for the rapid reduction ofhyperkalaemia (see Chapter 14).

Metabolic effects

By facilitating certain enzymes and inhibitingothers, insulin has wide-ranging effects on inter-mediary metabolism in most tissues (Table 9.2;Figure 9.1). The synthesis of the energy stores(glycogen in liver and skeletal muscle, fat in liverand adipose tissue) is facilitated, and their break-

down is inhibited. Tissue growth and cell divi-sion are also promoted by enhanced nucleic acid(DNA, RNA) synthesis, amino acid assimilationand protein synthesis.

Overall effect

Only a general appreciation of how insulin andthe catabolic hormones control everyday meta-bolic variations is given here (see also Referencesand further reading).

Anabolic actions of insulinFollowing a meal, glucose is absorbed from theGIT into the blood and rapidly transported intothe cells, to be converted into forms suitable forstorage and later use.

In the liver some glucose is converted intoglycogen and stored but most is converted intolipid (free fatty acid, FFA [or non-esterified fattyacid, NEFA], and triglyceride). Lipid is releasedinto the blood as very-low-density lipoprotein(VLDL), to be taken up and stored in adiposetissue. However, the release of glucose into theblood is inhibited. Hepatic regulation of glucoseoutput is an important mechanism for limit-ing the uptake of glucose into tissues wheretransport is independent of insulin.

In adipose tissue, fat breakdown is inhibitedand glucose uptake promoted. The glucoseprovides glycerol for esterification with FFAs,and the resulting fat is stored. Adipose tissue alsotakes up the fat-containing chylomicronsobtained by digestion (see Chapter 3). In muscle,fat metabolism is inhibited and glycogen is

Physiological principles of glucose and insulin metabolism 583

Table 9.1 Insulin requirement and transporters for glucose uptake into different tissues

Tissues not requiring insulin Transporter Tissues requiring insulin Transporter

Gastrointestinal – uptake SGLT AdiposeGastrointestinal – release to blood GLUT2 Muscle – skeletal, cardiac, smooth � GLUT4Liver GLUT7 Other tissuesNerves, brain GLUT1,3Kidney tubules GLUT2, SGLTEye – retinal vessels, lens SGLTLeucocytes GLUTBlood vessel endothelium GLUTPancreatic beta cells GLUT2

SGLT, sodium-dependent glucose transporter; GLUT, glucose transporter.

synthesized, which increases glucose availabilityfor immediate energy needs. Amino acid uptakeis promoted so that growth can be continued.

Catabolic actions of counter-regulatoryhormonesDuring stresses such as ‘fight or flight’, infectionor any major trauma, catabolic hormones reversethese processes. Blood glucose is rapidly raised tosupply energy for the muscles and if this is insuf-ficient fats can also be mobilized. Peripheraloxidation of FFAs produces large amounts ofenergy, but in the liver excess acetyl-CoA isproduced. This is condensed to produce high-energy ketoacids such as acetoacetate, whichmany tissues can utilize in small amounts. Ininsulin insufficiency these ‘ketone bodies’ mayaccumulate in the plasma, causing ketoacidosis.

Insulin deficiency

The consequences of insulin deficiency, and thusmany of the clinical features of diabetes, can bededuced from these considerations (Figure 9.2).

It will be explained below that obese type 2patients may not at first have an absolute defi-ciency of insulin; rather, there is a degree ofinsulin resistance. This may be described as arelative lack because the result is the same; more-over, eventually their insulin levels do fall. Thereare important differences between the physio-logical effects of partial (or relative) deficiencyand total insulin deficiency.

Partial deficiency (type 2)

Even small amounts of insulin will preventsevere metabolic disruption, especially acceler-ated fat metabolism, i.e. ketosis. Thus, althoughfasting blood glucose levels may be raised, themain problems only arise after meals; these arisefrom impaired glucose transport and cellularuptake resulting in impaired clearance from theblood. Adipose and muscle tissue cannot take upglucose efficiently, causing it to remain in theblood, and glucose deficiency in muscle maycause weakness. Becuse other tissues cannotcompensate sufficiently to assimilate the entire postprandial glucose load, the blood


Table 9.2 Metabolic effects of insulin

Metabolite Process TissueLiver Muscle Adipose

CarbohydrateIncreased • Glycogen synthesis (glycogenesis) ✓ ✓ ↔

• Glucose oxidation (glycolysis) ✓ ✓ ✓

Decreased • Glycogen breakdown (glycogenolysis) ✓ ✓ ↔• Glucose synthesis (gluconeogenesis) ✓ ↔ ↔

LipidIncreased • Fat synthesis (lipogenesis) ✓ ↔ ✓

• Utilization of dietary fat ✓ ↔ ✓

Decreased • Fat breakdown (lipolysis) ↔ ↔ ✓

• Fatty acid oxidation (ketogenesis) ✓ ✓ ✓

ProteinIncreased • Protein synthesis ✓ ✓ ✓

Decreased • Protein breakdown (proteolysis) ✓ ✓ ↔

Nucleic acidIncreased • DNA and RNA synthesis ↔ ✓ ↔

• Cell growth and division ↔ ✓ ↔

✓, insulin has important effect (increase or decrease) on process in this tissue; ↔, no effect.

glucose level rises causing hyperglycaemia(�11 mmol/L).

When the blood glucose level increases so thatthe concentration in the glomerular filtrateexceeds the renal threshold (see Chapter 14, p. 876), glucose is lost in the urine (glycosuria).Urinary glucose acts as an osmotic diureticcarrying with it large volumes of water (polyuriaand urinary frequency), resulting in excessivethirst and fluid intake (polydipsia). Because ofreduced fat uptake by adipose tissue, plasmalipid levels rise, especially triglycerides (dyslipi-daemia). LDL is relatively unaffected but HDL isreduced, increasing atherogenic risk (Chapter 4).Protein synthesis may be reduced but patientsare often still relatively obese. However, theyusually do lose weight in the weeks before firstdiagnosis, in part due to dehydration.

Total deficiency (type 1)

With no insulin at all there is severe hypergly-caemia at most times. This may raise the bloodosmotic pressure sufficiently to cause neurolog-ical complications including coma; this isdiscussed on pp. 594–596. Cellular metabolism isprofoundly disturbed. No glucose is available forenergy metabolism, and the first result is adepletion of liver and muscle glycogen stores.

Subsequently fat is mobilized, mainly fromadipose tissue, so that plasma triglyceride andFFA levels rise, as does lipoprotein. These supplyenergy needs for a little longer while the patient loses yet more weight. The brain cellsswitch to metabolizing the hepatically producedketo-acids. Fat stores are not replenished, andeventually may be exhausted. Finally, proteinmust be broken down into amino acids, whichcan be converted to glucose in the liver (gluco-neogenesis), at the expense of lean muscle mass.Other than in uncontrolled diabetes, this processnormally occurs only in times of prolonged star-vation; it is a desperate remedy that is akin toburning the house down to keep warm. Further,without insulin, any glucose so produced cannotbe utilized effectively anyway. This situation isinevitably fatal within months.

Thus many of the clinical problems in type2 diabetes are a direct consequence ofhyperglycaemia, while in type 1 diabetes there isalso disrupted intracellular metabolism. Inaddition, chronic complications occur in bothtypes, related to both hyperglycaemia anddyslipidaemia. These are discussed below.

Insulin physiology

Insulin (molecular weight about 5800 Da) iscomposed of 51 amino acids in two chains of21 (A chain) and 30 (B chain) amino acidsconnected by two disulphide bridges. It is syn-thesized in the pancreatic islet beta-cells. Othercells in the islets are the alpha-cells (producingglucagon) and the delta-cells (producing somato-statin). Islet cells altogether comprise less than3% of the pancreatic mass. Insulin is stored ingranules in combination with C-peptide asproinsulin (molecular weight 9000 Da), which issplit before release into the portal vein. Insulinhas a plasma half-life of only about 5 min.Approximately 50% of insulin is extracted by theliver, which is its main site of action, and afterutilization it is subsequently degraded. Eventu-ally, kidney peptidase also metabolizes someinsulin. C-peptide is less rapidly cleared and isthus a useful index of beta-cell function. Themain control of insulin level is plasma glucose: arise stimulates both the release and the synthesis

Physiological principles of glucose and insulin metabolism 585

absorptioninto plasma

PlasmaGLUCOSE

PlasmaAMINOACID

typical body cell

PROTEIN

AMINO ACID

GLUCOSE

Energy

GLYCOGEN

FATS

DIET

Urinaryloss

negativenitrogenbalance

Ketones Lipoprotein

Figure 9.2 Metabolic consequences of insulin lack.Cellular uptake of glucose is prevented so that afterexhausting their glycogen supplies, cells need to use fatsand even protein for their energy needs. Compare withFigure 9.1. Note: not all actions occur in all body cells.

of insulin. Amino acids and possibly fats alsopromote insulin release (Figure 9.3).

A wide variety of other neuronal, endocrine,pharmacological and local influences on insulinrelease have been identified (Figure 9.3), buttheir physiological or pathological significanceis not established. Adrenergic beta-receptorsmediate release, so beta-blockers can theoreti-cally inhibit this, though stimulation ofinhibitory adrenergic alpha-receptors, magnifiedduring the hyperglycaemic stress response,usually predominates.

Interestingly, glucose is a more powerfulstimulant orally than parenterally, and variousgut hormones have been implicated in this.Glucagon also promotes insulin release, possiblyto facilitate cellular uptake of the glucose that itcauses to be released into the plasma.

Pattern of secretion

It is important to note also that there is acontinuous basal level of insulin secretionthroughout the 24 h, independent of foodintake, which contributes to the regulation ofmetabolism and promotes glucose uptake intocells. This amounts to about 1 unit/h. Followinga meal there is an additional bolus secreted,which is biphasic. Within 1 min of bloodglucose levels rising, preformed insulin isreleased from granules in beta-cells into theblood. This release is stimulated by certainantidiabetic agents (insulin secretagogues) andis the first component to be compromised inearly diabetes. Should hyperglycaemia persist,further insulin synthesis is stimulated and there

is a delayed second phase of secretion afterabout 45 min. Appoximately 5–10 units aresecreted with each meal.

Thus the plasma insulin concentration curvenormally closely parallels the plasma glucoseconcentration curve throughout the day, reflect-ing every small change in nutrient supply ordemand (Figure 9.4). Considering these subtleand sometimes rapid adaptations, it can beappreciated how far current therapeutic methodsfall short of mimicking the physiological ideal.

In non-diabetics, the total daily secretion ofinsulin is probably rather less than the averagedaily requirement in type 1 diabetes of 50 unitsof exogenous insulin, mainly because of losses atthe injection site.

Amylin

The 37-amino acid peptide amylin is co-secretedwith insulin from beta-cells. It appears tocontribute to glucose regulation by a local(paracrine) action on islet cells, which moderatesintestinal glucose uptake, thereby reducing theload presented to the pancreas, or by suppressingglucagon secretion. In diabetes, amylin defi-ciency parallels that of insulin and it is believedthat patients whose postprandial hypergly-caemia is not adequately controlled by conven-tional therapy may benefit from amylinagonists, although none is yet in clinical use.

Insulin receptors

These are present on the cell surfaces of allinsulin-sensitive tissues and are normally down-


Amino acids (some)

INSULIN

Glucose

BETA-ISLETCELL

Hormones (e.g. glucagon, incretin, CCK)

Parasympathetic nervous system/cholinergic agents

Sympathetic nervous system –beta-adrenergic agents

Sulphonylureas/meglitinides

Somatostatin

Beta-blocking drugsThiazides

Sympathetic nervous system –alpha-adrenergic agents

Corticosteroids/oral contraceptives

Figure 9.3 Factors affecting the release or action of insulin. CCK, cholecystokinin. –––––● inhibition/antagonism;–––––●● stimulation/potentiation.

regulated by insulin, especially if it is present atcontinuously high levels, e.g. the hyperinsuli-naemia of over-eating, obesity or obesity-relatedtype 2 diabetes. This may account for thereduced insulin sensitivity (insulin resistance)found in some patients and the beneficial effectof weight reduction, especially of abdominal fat,on glucose tolerance: there is a vicious cyclewhereby hyperglycaemia and reduced insulinaction reinforce one another. Long-term insulintreatment also often gradually reduces theinsulin requirement, perhaps owing to reducedglucose levels. However, there is still much to belearned about the interactions between insulin,insulin receptors and carbohydrate metabolism.

Epidemiology and classification

The hallmark of diabetes is hyperglycaemia,owing to abnormalities of insulin secretion oraction. There are two primary forms of diabetesand a variety of minor secondary ones. In type 1diabetes there is usually gross destruction of theinsulin-secreting pancreatic beta-cells. In type

2 diabetes insulin is secreted but is eitherinadequate or insufficiently effective to meetmetabolic needs.

The current WHO definition of diabetes isbased on standardized measurements of plasmaglucose concentrations. It defines three classes,diabetes, impaired glucose tolerance andimpaired fasting blood glucose (Table 9.3).Patients in the second category are borderlineand about half will progress to frank diabeteseventually (up to 5% per year). However, theyneed not be treated immediately, depending onage and the presence of other risk factors: olderpatients or those with no cardiovascular riskfactors may just be monitored. More recently thecategory of impaired fasting glucose has beenintroduced in an attempt to identify at an evenearlier stage those with latent or ‘pre-diabetes’who should be monitored. It is a less reliablepredictor but has the advantge that it does notrequire a glucose tolerance test (see below).

Often, a single random plasma glucose of�11.1 mmol/L (blood glucose 10 mmol/L) issufficient for diagnosis in a patient with classicsymptoms, although this should be confirmedwith a fasting plasma glucose �7 mmol/L.

Epidemiology and classification 587

Bloo

d gl

ucos

e(m

mol

/L)

Time (hours)

Plas

ma

insu

lin (m

U/L

)

Eveningmeal

Mid-afternoonsnackLunch

Mid-morningsnackBreakfast

Fasting blood glucose level

Basal plasma insulin level

40

80

24001800120006002400

10

10

00

5

Figure 9.4 Schematic representation of normal diurnal variations in blood glucose and plasma insulin levels. As theblood glucose level rapidly rises after a meal, it is closely followed by an increase in insulin level to limit the rise. Theinsulin returns towards the basal level as blood glucose reaches the normal fasting level once more. Note how the twosubstances follow almost parallel curves, the insulin a little later than the glucose. The small but positive constant basalinsulin level emphasizes that insulin has functions other than just dealing with dietary glucose. Note: this diagram doesnot differentiate the two phases of insulin release.

Laboratories may report plasma glucose levels, asspecified by the American Diabetic Associationdiagnostic criteria, whereas finger prick testsmeasure blood levels; nevertheless, it iscustomary always to refer to blood glucose indiscussing diabetes. In borderline cases the oralglucose tolerance test (OGTT) can beperformed: the patient’s blood glucose ismeasured before and at 2 h after a standardized75-g glucose load, given orally following anovernight fast.

Epidemiology

Diabetes is known to affect more than 2% of theUK population, and probably as many again arelikely to have impaired glucose tolerance or evenfrank diabetes if screened. The prevalence variesconsiderably between populations. For example,Europeans are prone to type 1, especially innorthern Europe, whereas the incidence in Japanis less than 10% of that in Finland.

Type 2 seems to be related partly to the afflu-ence of a population, possibly through theprevalence of obesity, inactivity or both, whichare major risk factors. However, genetic factorsare also important. In some ethnic groups theprevalence is very high, e.g. in some PacificIslanders and the North American Pima Indiansit reaches 50%. Among South Asian immigrantsto the UK it is five times that in the host popu-lation, suggesting a possible genetic suscepti-bility to changed environmental factors, e.g. adiet richer in fats and sugar.

Classification

Primary diabetes – type 1 and type 2

In the vast majority of cases there is directdamage to the pancreatic islet cells. Differentattempts to classify diabetes comprehensivelyhave been confounded by the use of criteria thatare not mutually exclusive (e.g. age at onset,patient build or need for insulin). For example,some older (‘maturity onset’ or type 2) patientseventually require insulin, some older patientsneed it from the start (‘latent autoimmunediabetes in the adult’, LADA) and a few youngerpatients may not (‘maturity onset diabetes of theyoung’, MODY). Whether the patient needsinsulin may be the most practical distinction,but does not correspond consistently with otherimportant parameters.

A classification based on the pathogenesis ofthe pancreatic damage is now accepted as themost meaningful. This distinguishes two broadtypes (Table 9.4), which correspond roughly withinsulin dependency. The key criterion is the modeof pancreatic damage, but many other distinc-tions follow from this classification, includingnatural history, family history and patient type.These will be discussed in the following sections.

Secondary diabetes

A minority of cases with identifiable primarycauses (e.g. severe pancreatitis, steroid-induceddiabetes) do not fit readily into either of theconventional categories. They may or may notrequire insulin for treatment (p. 591).


Table 9.3 WHO definitions of diabetes mellitus (based on plasma glucose levels, as measured in laboratory)

Class Plasma glucose (mmol/L)

Fasting OGTT at 2 hDiabetes mellitus �7 and/or �11.1Impaired glucose tolerance (IGT) �7 and 7.8–11.1Impaired fasting glucose (IFG) 6.1–7Normal fasting glucose �6.1 and �7.8

If whole blood is used (as obtained by finger prick) all figures would be approx. 10% lower (e.g. 6.1 and 10 mmol/L for diabetes mellitus).

The apparently non-uniform thresholds derive from conversion from old mg/100 mL units, as still used in North America.

OGTT, oral glucose tolerance test.

Aetiology and pathogenesis

Primary diabetes

Despite having similar clinical pictures andcomplications, types 1 and 2 primary diabeteshave very different causes (Table 9.5).

Type 1 diabetes

In type 1 diabetes the islet beta-cells are almostcompletely destroyed by an autoimmuneprocess. Antibodies against all islet cells, andbeta-cells specifically, are found in 80% of

patients. However, interestingly, it is not theseanti-islet antibodies that mediate cell destructionbut T-cells; the islets are invaded by inflamma-tory cells causing insulitis. Insulin autoanti-bodies may also be found but their significanceis uncertain. As is usual with autoimmunedisease, there is rarely a strong family history:siblings or children of people with type 1diabetes have about a 5% chance of developingthe disease. However, there is a correlation withthe patient’s HLA tissue type (see Chapter 2) and in a minority of patients an association withother autoimmune diseases, especially ofendocrine tissues (e.g. thyroiditis, perniciousanaemia).

Aetiology and pathogenesis 589

Table 9.4 Comparison of the main types of primary diabetes mellitus

Type 1 Type 2

Endogenous insulin Absent PresentInsulin deficiency Absolute Relative or partial

Insulin receptor defect?Insulin resistance Usually absent May be presentPancreatic islet damage Severe (destruction) Slight/moderateImmunology Auto-immune; islet cell antibodies No antibodies demonstratedUsual age of onset �30 years �40 yearsBuild of patient Thin Obese (usually)Therapeutic class Insulin-dependent (IDDM) Non-insulin-dependent (NIDDM;

but may require insulin)Genetics Weak family history; HLA-linked Strong family historyKetoacidosis prone? Yes No

Table 9.5 Aetiology and pathology of primary diabetes

Type 1 Type 2

Risk factors HLA antigens (DR3, DR4) Family historyOver-eating; lack of exerciseToxin? Amyloid?Ethnic group

Trigger factors Viral infection ObesityMetabolic stress/excessive demand Metabolic stress/excessive demandEnvironmental toxin?

Pathogenesis Rapid autoimummune destruction of islet cells Gradual islet cell degeneration / depletionPeripheral insulin receptor defect?

Overt diabetes may follow many years ofsubclinical pancreatic damage, and when itoccurs there is usually less than 10% of func-tional islet cell mass remaining. Clinical onset isusually abrupt, over a few weeks, and often asso-ciated with, or precipitated by, a metabolic stresssuch as an infection, which acutely increasesinsulin demand beyond capacity. This mightaccount for the winter seasonal peak in inci-dence and also the brief temporary remissionthat frequently follows, as the infection remitsand the marginal insulin levels once again justcompensate. Subsequently, full-blown diseaseirreversibly takes hold. As with other autoim-mune diseases, viral infection may be causingthe expression of a normally suppressed HLAreceptor, which subsequently activates lympho-cytes (see Chapter 2). Other environmental trig-gers such as toxins or certain foods (includingmilk protein) may also be involved.

Autoantibodies may be found in somepatients up to 15 years before the onset of acutedisease. This could eventually provide a meansof early identification of prediabetes, so thatthey may be treated prophylactically, possiblyby immunotherapy. However, such markers arealso often found in close relatives who neverdevelop the disease, and the chance of theidentical twin of a diabetic patient subsequentlydeveloping diabetes is less than 50%. Theintroduction of the category of ‘impairedfasting glucose’ was another attempt at earlyidentification of potential sufferers.

Thus it seems that in type 1 diabetes there is agenetically determined HLA-dependent suscepti-bility that requires an environmental trigger forfull expression. Following contact with thistrigger, which may never be encountered, swiftdeterioration and complete insulin dependenceare inevitable. There is still considerable ignor-ance of the relative contributions of genes andenvironment and of specific environmentalfactors.

Type 2 diabetes

These patients have one or more of the followingfundamental abnormalities, and in establisheddisease all three commonly coexist:

• Absolute insulin deficiency, i.e. reducedinsulin secretion.

• Relative insulin deficiency: not enoughinsulin is secreted for metabolic increasedneeds (e.g. in obesity).

• Insulin resistance and hyperinsulinaemia: aperipheral insulin utilization defect.

In most cases type 2 diabetes is associated withobesity (particularly abdominal obesity) on first presentation, and in a quarter of all peoplewith diabetes simple weight reduction reversesthe hyperglycaemia. This is commonly associ-ated with peripheral insulin resistance owing to receptor-binding or post-receptor defects.Obesity and reduced exercise also contribute to insulin resistance and are modifiable risk factors for type 2 diabetes. The resultant hyper-glycaemia induces insulin hypersecretion,hyperinsulinaemia and insulin receptor down-regulation, i.e. further insulin resistance. Hyper-glycaemia itself is known to damage beta-cellsowing to the direct toxic effect of excessive intra-cellular glucose metabolism, which produces anexcess of oxidative by-products; these cannot bedestroyed by natural scavengers such as catalaseand superoxide dismutase. The vicious cycleeventually depletes (‘exhausts’) the beta-cells,intrinsic insulin levels fall and some patientsmay eventually come to require exogenousinsulin therapy. Thus, type 2 diabetes is usually aprogressive disease, although the late onsetusually means that some patients die beforerequiring insulin.

There is still debate as to the primary defect oftype 2 diabetes. It has also been proposed thatthe amyloid deposits (insoluble protein) longknown to be found in the pancreas of type 2patients are related to abnormalities in amylinsecretion (p. 586) and contribute to thepancreatic defect.

There is an association between abdominalobesity, hyperinsulinaemia, insulin resistance,hyperlipidaemia, type 2 diabetes and hyperten-sion, and this combination of risk factors istermed metabolic syndrome. However, despitemuch research, as yet it is not known which ofthese factors (if any) is the prime cause, or ifthere is another underlying reason.


GeneticsThe genetic component in type 2 diabetes ismuch greater than in type 1. A family history isvery common, often involving several relatives.Identical twins almost always both develop thedisease, and offspring with both parents havingdiabetes have a 50% chance of developing thedisease. The ‘thrifty gene’ hypothesis proposesthat the ability to store fat efficiently – andhence develop obesity – conferred a survivaladvantage in more primitive societies wherefamine was a regular phenomenon, hence itspersistence in the genome. This may explainwhy some pre-industrial groups (e.g. PacificIslanders) readily develop diabetes whenexposed to the industrialized lifestyle.

Secondary diabetes

Most diabetes results from primary defects of the pancreatic islet cells. However, there are occasionally other causes of ineffective insulin action, impaired glucose tolerance andhyperglycaemia (Table 9.6).

Natural history

Onset

About 80–90% of diabetic patients have type 2diabetes, which tends to occur late in life, hencethe obsolete description ‘maturity onset’. Onsetis usually insidious and gradual, patients toler-ating mild polyuric symptoms perhaps for manyyears.

The other 10–20% have type 1 diabetes andrequire insulin at the outset. Almost invariablythey become ill at an early age: the peak onset of type 1 is around puberty, starting mostcommonly in the winter months. Although thedisease may be present subclinically for someconsiderable time (months, or possibly years),clinical onset is invariably abrupt.

Presentation

Type 2 diabetes is usually first diagnosedfollowing one of three common presentations(Table 9.7):

Natural history 591

Table 9.6 Some causes of secondary diabetes

General mechanism Aetiology Example

Hepatic glucose metabolism defect Liver failure Viral hepatitis, drugsPancreatic destruction Cirrhosis Alcoholism

PancreatitisAnti-insulin hormones Corticosteroids Cushing’s disease

Steroid therapyPregnancy (‘gestational diabetes’)Major trauma/stress

Growth hormone AcromegalyAdrenaline (epinephrine), etc. PhaeochromocytomaGlucagon GlucagonomaThyroid hormones Hyperthyroidism

Major trauma/stressAdrenergic drugs

Hyperglycaemic/anti-insulin drugs Thiazide diuretics, diazoxideOral contraceptives

Insulin antibodies Autoimmune diseaseAbnormal insulin receptors Congenital lipodystrophy

• About half of patients first complain ofincreasing polyuria and/or polydipsia.

• In about a third it is a chance finding of glyco-suria or hyperglycaemia at a routine medicalexamination.

• In less than 20% of cases the patientcomplains of symptoms subsequently foundto result from a complication secondary todiabetes.

Type 2 patients may be asymptomatic or mayhave been only mildly symptomatic for severalyears. Commonly, they ignore these symptomsor attribute them to ageing, and only presentwhen classical symptoms such as polyuria, thirst,tiredness or recent weight loss (even though thepatient may still be relatively obese) becomeunacceptable. In many other cases their diabetesis only detected when they undergo a medicalexamination, e.g. for insurance purposes or anew job. Alternatively, the complaint may be ofan infective complication not obviously linkedto diabetes, at least not in the patient’s mind,such as recurrent candida infections or boils, anon-healing foot lesion or a persistent urinary-tract infection. Rarely, as the complicationsproceed insidiously even during this earlyperiod, the primary reason for consultation mayresult from vascular disease, nephropathy,neuropathy, retinopathy or impotence. In somecases IHD, even MI, is the first presentation.

A common manifestation of the complicationsis the ‘diabetic foot’. The patient presents with a possibly gangrenous foot lesion, probablyfollowing a recent injury and subsequentinfection.

Only very rarely will a type 2 patient firstpresent with metabolically decompensateddisease (ketoacidosis). These patients will prob-ably have had impaired glucose tolerance for

some time and then have undergone some majorstress such as MI or serious infection. Anotherpossible trigger factor could be starting a drugthat impairs glucose tolerance, e.g. a thiazidediuretic or an atypical antipsychotic. Suchstresses may also uncover latent disease in a lessdramatic manner.

Unfortunately, a severe acute presentation isfar more common at the onset of type 1 disease.This is usually associated with some metabolicstress (e.g. infection), and presents with rapidweight loss, weakness, extreme thirst, severepolyuria, urinary frequency and multiplenocturia. Some may even go on to acute meta-bolic decompensation (ketoacidosis) and evencoma, being practically moribund on hospitaladmission. Following recovery with insulintherapy there may follow some months ofapparent remission with a reduced or absentinsulin requirement, the so-called ‘honeymoonperiod’, but these patients then deterioraterapidly. Before the isolation and therapeutic useof insulin in the 1920s they inevitably diedshortly thereafter.

Progression

Insulin secretion in type 2 diabetes declinesrelatively slowly, but up to one-third of patientsmay eventually need exogenous insulin, i.e. theyare ‘insulin-requiring’ as opposed to insulin-dependent.

In most type 1 diabetes, pancreatic beta-celldestruction is already almost complete at diag-nosis, and routine insulin requirements do notgenerally increase. However, in both types themultisystem complications progress throughoutlife at rates that vary considerably betweenpatients and will very likely be the eventualcause of death. People with diabetes have areduced life expectancy, although the prognosishas greatly improved with advances in treat-ment. Younger patients have mortality rates ofup to five times that of the general population,while for older ones it is about twice normal. Theprecise prognosis for any given patient willdepend on many factors, but particularly theoverall consistency of control of blood glucose.


Table 9.7 Different presentations of type 2 diabetes

Typical diabetic symptoms (see text) 55%(a)

Chance finding 30%Complication – infective 15%

– other 2%

(a) Approximate figures; after Watkins (2003) (See References and

further reading).

Clinical features

Symptoms

The symptoms of diabetes as summarized inTable 9.8 are best understood in relation to theirpathogenesis.

Symptoms due to hyperglycaemia

The classic symptoms, which give diabetesmellitus its name (‘sweet fountain’), are easilyexplained by the osmotic effect of the elevatedblood glucose levels that occur when glucose isdenied entry to cells. They are more pronouncedwhen the blood glucose level rises rapidly, e.g.in decompensation or acute onset. The osmoticeffect of chronic hyperglycaemia will to someextent be compensated by compensatory hyponatraemia and an increased intracellularosmolarity (see Chapter 14).

When the blood glucose level exceeds therenal threshold (about 10 mmol/L), glucoseappears in the urine in large quantities. Thetraditional method of distinguishing diabetesmellitus from diabetes insipidus – almost theonly two idiopathic causes of chronic polyuria –was simply to taste the urine: in the former caseit is sweet, and in the latter literally insipid(tasteless). Glycosuria predisposes to urinary-tract infection, partly because of the favourablegrowth medium presented to perineal organismsand partly because diabetic patients are generallymore susceptible to infection (see below).

Diabetic urine dries to leave a white glucosedeposit, a clue that sometimes leads to diagnosis:there may be underwear stains or white speckson the shoes of elderly males (from carelessmicturition). Severe plasma hyperosmolaritymay reduce the intraocular pressure, causingeyeball and lens deformity, and glucose may alter lens refraction: both lead to blurredvision. This is sometimes a prodromal sign ofhyperglycaemic crisis in type 1 diabetes.

Impaired metabolism and complications

The metabolic consequences of insulin lack werediscussed in detail above. The pathophysiologyof hyperglycaemia and ketoacidosis is nowconsidered.

Complications

Most complications of diabetes are due to eitheracute metabolic disturbances or chronic tissuedamage.

Acute complications

The most common acute complications aredisturbances in glycaemic control. Optimalmanagement of diabetes aims for a delicatebalance, preventing excessive glucose levels butnot forcing glucose levels too low. A variety ofcircumstances can drive the glucose level outside

Complications 593

Table 9.8 Clinical features of diabetes

Direct consequences of high blood glucose levelsPolyuria, frequency, nocturia, polydipsia (osmotic diuresis)Visual disturbance (osmotic changes to intra-ocular pressure)Urethritis, pruritis vulvae, balanitis (urogenital infection)Metabolic consequences of impaired glucose utilizationLethargy, weakness, weight loss (intracellular glucose deficit)Ketoacidosis (increased fat metabolism)Long-term complications of hyperglycaemia and hyperlipidaemiaVascular disease, heart disease, renal disease, neuropathy, eye disease, infections, arthropathy

these narrow limits, and if treatment is notadjusted accordingly, the result is either excess orinsufficient glucose in the blood (Table 9.9).

Hyperglycaemia/ketoacidosis

Causes, pathogenesis and symptomsHyperglycaemia in treated diabetes usually arises because normal medication is somehowomitted or becomes insufficient to meet anincreased insulin requirement. Drugs that raiseblood glucose levels can also interfere withcontrol. When diabetic control is lost, bloodglucose rises and the symptoms develop gradu-ally over a number of hours. Above a blood

glucose level of approximately 15–20 mmol/L,both hyperosmolar and metabolic problemsdevelop (Figure 9.5; Table 9.10).

Blood glucose levels can exceed 50 mmol/Land this high osmotic load (which is also in theextracellular fluid) cannot be matched withinthose cells from which glucose is excluded owingto the absence of insulin. Thus, water is drawnfrom the intracellular compartment and thiscauses tissue dehydration. This particularlyaffects the brain where the resultant reducedintracranial pressure leads to CNS depression.The skin is also dehydrated, and loses its elas-ticity; this reduced skin turgor can be detected bypinching a fold of skin and noting its delay in


Table 9.9 Causes of acute disturbances in diabetic control

Hypoglycaemia Hyperglycaemia/ketoacidosis

Excess (mis-measured?) dose Missed antidiabetic dosePotentiation of oral hypoglycaemic (drug interaction) Hyperglycaemic drugs, e.g. thiazides, steroidsMissed meal; dieting Excess dietary intakeUnexpected physical activity Metabolic stress, e.g. infection, surgery, pregnancyExcessively tight blood glucose controlAlcohol

switch to fatmetabolism

REDUCEDINTRACELLULAR

GLUCOSEHYPERGLYCAEMIA

INSULIN DEFICIENCY

CNS depressionconfusion, coma

tachycardiahypotension

thirstpolydipsia

polyuria,glycosuria

Na/K depletionhyperkalaemia

hyperventilation

ketoacidosis

tissue dehydration

increased plasmaosmolarity

increased glucosein glomerular

filtrate

osmotic diuresis

hypovolaemia

Figure 9.5 Pathogenesis and clinical features of acute hyperglycaemia and ketoacidosis.

springing back, but this is less conclusive in theelderly, in whom skin elasticity is alreadyreduced.

In the kidney the high load of glucose in theglomerular filtrate, not all of which can be reab-sorbed, produces an osmotic diuresis. This resultsin a reduction in circulating fluid volume,leading to hypotension and reflex tachycardia.The high urine volumes also cause a loss of elec-trolytes, especially sodium and potassium.However, the plasma potassium level may beparadoxically high because acidosis inhibits theNa/K pump throughout the body, preventingintracellular potassium uptake (see below andChapter 14, p. 891). Osmoreceptors and baro-receptors detect the electrolyte and fluid losses,causing thirst, but as CNS depression and confu-sion develop the patient often cannot respondby drinking.

In the absence of glucose, many cells start tometabolize fat instead. Adipose tissue releases

fatty acids, and the liver converts some of theseto acid ketones that can be readily utilized as analternative energy source by many tissues. Theresulting metabolic acidosis (diabetic ketoaci-dosis) is misinterpreted by the respiratory centreas carbon dioxide retention, resulting in anincreased respiratory drive and hyperventilation.Acidosis impairs oxygen dissociation from Hb,exacerbating the gasping (overbreathing, ‘airhunger’), and also causes peripheral vasodilata-tion, exacerbating the hypotension. Both respi-ratory rate and blood oxygen level fall as comasupervenes. Ketoacidosis is more likely todevelop in type 1 patients, although fortunatelyit is uncommon.

People with type 2 diabetes usually secretesufficient insulin to prevent them developingketoacidosis (except during severe stress), butthey may still suffer hyperosmolar non-ketotichyperglycaemic states. This may result in comaand is associated with a higher mortality thanketoacidosis.

ManagementDiabetic ketoacidosis is a medical emergencywith about a 15% mortality rate. Close moni-toring and very careful attention to the patient’sfluid and electrolyte balance and bloodbiochemistry are essential (Table 9.11). Imme-diate attention is life-saving, but the patient maytake several days to stabilize.

IV soluble insulin is essential. An initialbolus of about 6 units is followed by contin-uous infusion (6 units/h). Fluid replacementneeds are estimated from measurements of theCVP and plasma sodium level. Hyponatraemia(‘appropriate hyponatraemia’, glucose having

Complications 595

Table 9.10 Clinical features of hyperglycaemia andketoacidosis

Glycosuria, ketonuriaPolyuria, nocturiaThirst, polydipsia

HypotensionRapid (bounding) pulse and respiration

Dry mouth, reduced skin turgorVisual disturbance

Hyperkalaemia, acidosis, ketonaemiaSweet smell of ketones on breath

Weakness, drowsiness, eventually coma

Table 9.11 Principles of the management of ketoacidosis

Problem Treatment

Underlying cause Discover and treatHyperglycaemia and hyperosmolarity Insulin (soluble): small bolus plus continuous infusionDehydration IV infusion: saline/dextran/plasmaAcidosis Bicarbonate?Hyperkalaemia/potassium deficiency Careful potassium repletion, after correction of acidosisHypoxaemia Oxygen, up to 60% initially

osmotically displaced sodium in the plasma)and/or sodium depletion require 0.9% salineadministration. However, if the dehydration hascaused hypernatraemia, especially in the non-ketotic patient, hypotonic saline (e.g. 0.45%)may be indicated. Severe hypotension or shockrequire plasma replacement (see Chapter 14p. 903). Potassium replacement is difficult tomanage because the initial hyperkalaemiamasks a total body potassium deficit. However,once insulin is started and potassium movesintracellularly, closely monitored IV potassiumreplacement is required. Acidosis will oftenresolve spontaneously with conservative therapyas ketone production falls and existing ketonesare metabolized. Many clinicians would not usebicarbonate unless blood pH was below 7.00 forfear of overcompensating.

Hypoglycaemia

Causes

In all forms of diabetes, hypoglycaemia (bloodglucose �3 mmol/L) is much more commonthan symptomatic hyperglycaemia, and itdevelops very rapidly, sometimes withinminutes. Usually, either an excessive insulindose is accidentally injected (many patients haveeyesight problems) or else the normal dose ofinsulin or antidiabetic agent is not matched byan adequate dietary intake (Table 9.9). Insulin-induced hypoglycaemia is usually associatedwith injections of short-acting insulin.Deliberate overdosing is not unknown.

Hypoglycaemia induced by sulphonylureaantidiabetic drugs is rarer but more prolonged,more severe and more difficult to treat thaninsulin-induced hypoglycaemia. The elderly areespecially prone, partly because the drugs arecleared more slowly and partly because ofimpaired homeostasis. Drug interactions thatmight potentiate oral antidiabetic drugs areconsidered on p. 615. Alcohol not only causeshypoglycaemia by inhibiting hepatic gluconeo-genesis but also impairs the patients’ perceptionof it, reducing their ability to respond.

Pathogenesis and symptoms

Hypoglycaemic symptoms fall into two maingroups (Table 9.12). At glucose levels belowabout 4 mmol/L insulin release is inhibited and the counter-regulatory hormones such asglucagon and adrenaline are released in an effortto raise blood glucose. At a glucose level below3.5 mmol/L the body responds by activating the sympathetic nervous system and adrenalmedulla (the ‘fight or flight’ response). Theconsequent sympathetic/adrenal symptoms(Table 9.12) should provide the patient with apreliminary warning (but see below).

As the glucose level falls below about2.5 mmol/L, neurological signs develop owing tothe deficiency of glucose in the brain. Theseneuroglycopenic features may be noticed moreby others than by patients themselves, althoughmany patients do report an awareness of subjec-tive prodromes. Sometimes the signs are subtlechanges in mood or visual disturbances, buteventually there is almost always erratic behav-iour resembling drunkenness. This has some-times led to police arrest and delayed treatment,occasionally with fatal results. Frequent hypo-glycaemic attacks may have a cumulative delete-rious effect on higher brain function (cognition),especially in the elderly. All people with diabetesshould carry, in addition to a readily availablesugar source such as dextrose tablets, a card or


Table 9.12 Clinical features of hypoglycaemia

Adrenergic (autonomic) – enhanced sympatheticactivity

• tremor, sweating• shivering, palpitations• anxiety, pallor

Neuroglycopenic – reduced CNS glucose delivery• drowsiness, disorientation, confusion• apparent drunkenness; aggression,

inappropriate behaviour• convulsions, coma, brain damage; death

Other effects – multiple or indirect pathogenesis• hunger, salivation, weakness, blurred vision

bracelet stating that they have diabetes andshould be given sugar if found acting strangely.

A patient’s ability to recognize ‘hypos’ (theirhypoglycaemic awareness) should be checkedregularly because it tends to diminish. Long-term diabetes patients become less sensitive tothe warning signs and thus more vulnerable.This may result partly from autonomicneuropathy and partly from reduced counter-regulatory hormone response. It is also possiblethat frequent attacks may reduce the patient’sability to recognize them. Awareness is pro-gressively reduced by frequent hypogly-caemic episodes but may be at least partiallyrestored by minimizing or eliminating episodesthrough relaxing control slightly, more careful monitoring and patient education.

Most of the adrenergic symptoms are medi-ated by beta-receptors, and so may be antago-nized by concurrent beta-blocker therapy.Although this rarely presents a serious problem,such drugs should be avoided in people withdiabetes if they already experience hypogly-caemic unawareness. Otherwise, there is nocontra-indication but a cardioselective beta-blocker is preferred. Theoretically, beta-blockersmight help by preventing beta-mediated insulinrelease (Figure 9.3), but this is swamped by thesymptom-masking effect.

ManagementAlthough both hypoglycaemia and hypergly-caemia can result in coma, there is rarely anyproblem distinguishing them, especially as rapidblood glucose test stick methods are readilyavailable. A test dose of glucose would clinchmatters because hypoglycaemia will be veryrapidly reversed, whereas glucose would have nosignificant effect, either helpful or harmful, inhyperglycaemia. In contrast, insulin givenblindly would severely exacerbate hypogly-caemia and should never be given where there isdoubt.

The conscious patient must take glucosetablets, or sugar, chocolate, sweet tea, etc. Semi-conscious or comatose patients require IVglucose 20% or IM glucagon (1 mg). The

response is usually satisfyingly prompt, occur-ring within minutes. Glucagon injection canusually be managed easily by patients’ relatives,who should be fully informed on how torecognize and deal with hypoglycaemicepisodes. Unless patients or their relatives aretaught to recognize the early signs, the patientmay become comatose before being able tocorrect it.

Persistent hypoglycaemic attacks requirereassessment of therapy. Dietary modificationmay be required (e.g. increased carbohydrate),although this might compromise weight reduc-tion efforts. Modern intensive insulin therapyregimens aimed at producing ‘tight’ glycaemiccontrol have increased the likelihood ofhypoglycaemia, and a judgement of risk andbenefit has to be made when such regimens areconsidered (p. 626).

Unstable diabetes

A small proportion of people with type 1diabetes prove exceptionally difficult to control,experiencing frequent episodes of hypogly-caemia, hyperglycaemia or both. They are vari-ously termed brittle, unstable or labile. It isunlikely that this condition is inherent to theirdisease, and specific causes are always sought.Poor compliance through error, ignorance ordisability, e.g. visual problems measuring insulindoses, unrecognized intercurrent illness anddrug interaction must first be eliminated. Inolder patients with recurrent hypoglycaemia thepossibility of reduced hypoglycaemic awarenessmust be investigated.

Recurrent hyperglycaemia/ketoacidosis ismore common in young patients and maysometimes be associated with psychological orpsychopathological factors such as teenage rebel-lion or illness denial, self-destructive impulsesor other emotional instability. A particularsubgroup has been identified of slightly obesefemales aged 15–25 years who may be covertlymanipulating their therapy adversely. SupervisedIV therapy in some of these patients seems toresolve the problem temporarily.

Complications 597

Chronic complications

In many patients, even before diagnosis, wide-spread damage occurs in the kidney, nerves, eyesor vascular tree (Figure 9.6). These long-termcomplications are to different degrees commonto both types of diabetes, and their prevention ortreatment are the real challenges for diabetesmanagement and research.

Pathogenesis

It is important to determine whether or notthese chronic problems are a direct consequenceof hyperglycaemia. If so, then optimal control toachieve normoglycaemia would be expected tominimize them. Evidence has accumulated thatthis is broadly true for the so-called microvas-cular complications (mainly kidney, eye, nerves).The fact that similar complications arise in mosttypes of diabetes, despite their different aetiolo-

gies, supports the hyperglycaemia hypothesis.The extensive Diabetes Control and Complica-tions Trial (DCCT; 1992) confirmed that bettercontrol is associated with less severe complica-tions in type 1 diabetes. The UK ProspectiveDiabetes Study (UKPDS; 1998) supported thesame hypothesis in type 2 patients.

Other hypotheses have been proposed. Itcould be that an as yet unidentified primarylesion in diabetes is responsible independentlyfor both the hyperglycaemia and the complica-tions. If so, correcting one would not neces-sarily improve the other. Some complicationscould be secondary to the abnormal pattern oramount of insulin secretion, which is notcompletely rectified by conventional treatment.For example, the hyperinsulinaemia seen inmany type 2 patients may contribute to bloodvessel disease (macrovascular complications) orhypertension. Alternatively, the abnormallyhigh levels of counter-regulatory hormones


HYPERGLYCAEMIA

PROTEINGLYCATION

POLYOLACCUMULATION

Small blood vesselsMICROANGIOPATHY

large blood vesselsMACROANGIOPATHY

(Atherosclerosis)

Hypertension

NEPHROPATHYRETINOPATHYDermopathyArthropathy

NEUROPATHY

Cataract

Infections

DyslipidaemiaPlatelet/clotting defect

StrokeIschaemic heart disease

Peripheral vasculardisease

Figure 9.6 Possible pathogenetic mechanisms of chronic diabetic complications. The central box lists the clinicalfeatures. Also shown is possible interlinking of pathogenetic mechanisms.

usually found in diabetes may be deleterious.The involvement of growth hormone andinsulin-like growth factor in angiopathy hasalso been investigated but no clear patterndetected. Finally, there seems to be a geneticvariation in the susceptibility to differentcomplications, regardless of the degree ofglycaemic control.

Thus there is unlikely to be a simple answer,but the general strategy of normalizing bloodglucose is well established as the best wecurrently have for minimizing complications.

Three general mechanisms are proposed for thepathological basis of the complications: proteinglycation (glycosylation), abnormal polyolmetabolism and accelerated atheromatousarterial changes.

GlycationNormally, almost all body protein is to someextent glycated, i.e. glucose molecules from bodyfluids are covalently bound to free amine groupson protein side chains. The degree of glycation isdirectly proportional to the average bloodglucose level. An accessible marker for this is Hb glycation, particularly the HbA1c fraction.Other proteins, and also lipids and nucleopro-tein, throughout the body are similarly affected.In excess, one result is the formation ofabnormal crosslinks between different regions ofprotein chains. Protein configuration is thuschanged, disrupting secondary and tertiarystructure and hence function. Basementmembrane proteins seem particularly susceptibleto glycation, the result being thickening andincreased permeability (i.e. reduced selectivebarrier function). As basement membranes arepresent in most tissues, and especially in bloodvessels, this could account for the widespread,multisystem distribution of diabetic complica-tions. Chronic hyperglycaemia also results inoxidative stress through increases in mito-chondrial superoxide formation, producingadvanced glycation end-products (AGPs) thatcan cause a variety of damaging effects.

Basement membrane damage in capillariesand smaller arterioles can cause microan-giopathy and subsequent ischaemia in almostany organ. Retinopathy is undoubtedly causedin part by this mechanism. Neuropathy may

result from a combination of this and directglycation of the sheaths of small nerve axons,e.g. sensory nerves. Similarly, glycation of theglomerular basement membrane probably causesthe characteristic glomerular sclerosis of diabeticnephropathy, although renal arterial diseaseprobably also contributes. Glycation of tendonsheaths and joint capsules may be responsiblefor the joint problems, particularly the stiffnessin hands and feet, that some patients suffer;glycation of collagen in skin sometimes gives it athickened, waxy appearance. The myocardiummay also be affected, as may immune cells suchas macrophages and leucocytes.

Polyol metabolismSome tissues do not require insulin for glucosetransport into their cells (Table 9.1), relyinginstead simply on diffusion down a concentra-tion gradient. Thus, while other tissues areglucose-depleted in diabetes, these will accumu-late excess glucose in the presence of hypergly-caemia. Being surplus to energy needs, some ofthe excess glucose is reduced to polyols such assorbitol by the enzyme aldose reductase via anotherwise little used pathway (Figure 9.7).

The resulting polyols are not readily elimi-nated from the cells, possibly because they aremore polar than glucose and of greater molec-ular weight. Furthermore, low dehydrogenaseactivity, particularly in the eye lens and nervesheaths, means that they are not metabolizedefficiently. The resultant accumulation ofosmotically active molecules draws water intothe cells, causing them to expand, severelydisrupting their function and possibly killingthem. Retinal blood vessels, the eye lens andthe glomeruli may be damaged in this way,contributing to retinopathy, cataract andnephropathy, respectively. It has long beenknown that an analogous intracellular accumu-lation of galactitol in the lens is linked to thehigh prevalence of cataracts in the inheritedmetabolic disorder galactosaemia.

A further abnormality may also contribute.Myoinositol, an important intermediate inenergy handling, may (although also a polyol)instead of accumulating become deficient. By apoorly understood series of steps this deficiencymay impair nerve conduction (Figure 9.7).

Complications 599

MacroangiopathyAlmost all people with diabetes suffer fromincreased obstructive vascular disease owing to agreatly increased predisposition to atheroscle-rosis. Several factors contribute to this. Because oftheir more active lipid metabolism, people withdiabetes have raised plasma levels of triglyceridesand lowered HDL, producing an unfavourable,atherogenic lipoprotein ratio (see Chapter 4,Figure 4.28). Furthermore, many type 2 patientsare initially hyperinsulinaemic and insulin mayitself be a growth factor for atheroma. Plateletaggregating ability is also usually raised, andhypertension is common. Thus major risk factorsfor atherosclerosis are intensified and cerebro-vascular disease, stroke, IHD and peripheral vas-cular disease are common. Macroangiopathy alsocontributes to kidney disease.

Other mechanismsAs illustrated in Figure 9.6, other complications ofdiabetes occur, the pathogenesis of which remainobscure. Moreover, different complicationsmay be inter-related or coexistent. Neuropathymay result partly from direct neuronal damageand partly from impaired blood supply to thenerve sheaths. Microangiopathy may resultpartly from glycation, partly from polyol accu-mulation and partly from hyperinsulinaemia.Once nephropathy is established, it promoteshypertension and vascular disease.

Diabetes and hypertension. There is anassociation between diabetes (especially type 2)and hypertension, as part of the metabolicsyndrome. The precise cause and effect relationships have not yet been elucidated.Many hypertensives have insulin resistance,hyperinsulinaemia and impaired glucose toler-ance, and insulin may have several hypertensiveactions including promoting renal sodium reten-tion, increasing sympathetic vasconstrictoractivity and directly increasing vascular reac-tivity, via an effect on sodium handling. Insome cases hypertension may be secondary todiabetic kidney disease, although the conversemay also be true (see Chapter 4, p. 213).Alternatively, it may be that a third, as yetunknown, independent factor first causes insulinresistance, which then leads to both type 2diabetes and hypertension. Hyperinsulinaemiacould then be a common link in the vascular complications of both diabetes andhypertension.

The UKPDS (1998) found that rigorous controlof blood pressure in diabetes reduced complica-tions. However, prolonged therapy with twocommon antihypertensive agents, thiazidediuretics and beta-blockers, while effectivelylowering blood pressure, can also lead to glucoseintolerance or even overt diabetes. For thisreason beta-blockers are not recommended asfirst-line treatment for hypertension in diabetes,and extra care is needed with both.


hyperglycaemia

GLYCOLYSISGlucose-6-phosphate

reduced Na pump activity

SORBITOLand other polyols

MYOINOSITOLdepletion

reducedMYOINOSITOL

uptake

aldose reductase

hexokinaseNormal pathway

OSMOTICDAMAGE

IMPAIRED NERVECONDUCTION

excess glucose

??

GLUCOSE

Figure 9.7 Polyol pathway and effects of polyol accumulation.

Clinical consequences

Almost any system in the body may be affectedby diabetic complications, which is why diabetesis regarded as a multisystem disease (Table 9.13).

Eyes. Diabetes is the most common cause ofacquired blindness in developed countries. After30 years of diabetes, about 50% of patients havesome degree of retinopathy, and up to 10%become blind. The blindness is due to small-vessel damage in the retina, with dilatation,haemorrhage, infarction and ultimately exces-sive proliferation of new vessels that project into the vitreous humour (neovascularization).Retinopathy is frequently associated withnephropathy. People with diabetes also have anincreased incidence of glaucoma and cataract.

Nervous system. Diabetic neuropathy mayaffect any part of the peripheral nervous system,but most commonly starts with the peripheralsensory nerves, causing tingling and numbness(paraesthesias), loss of vibration sense or thesense of balance and limb position. It may inter-fere with the ability of blind people withdiabetes in reading Braille. Autonomicneuropathy is potentially devastating becauseit can seriously disturb cardiovascular, gastro-intestinal or genitourinary function, causingnumerous symptoms; postural hypotension andimpotence are common. Voluntary motor nervesare less commonly affected.

Renal. Diabetic nephropathy is the cause ofdeath in about 25% of type 1 diabetes. Predomi-nantly a form of sclerosis of the glomerular base-ment membrane, it develops very slowly and somost commonly occurs in type 1 patients, up to40% of whom may be affected. The increasedglomerular filtration rate (‘hyperfiltration’) inearly diabetes, which is due to hypertension andto the osmotic loading of hyperglycaemia, mayoverload renal capillaries. Nephropathy isheralded by microalbuminuria, with increasingproteinuria frequently progressing to end-stagerenal failure, associated with worsening hyper-tension. Diabetic nephropathy is one of the mostcommon causes of chronic renal failure, withpeople with diabetes comprising about 15% ofthe caseload of UK renal replacement therapyunits. Renal decline is hastened by inadequate ortardy treatment of associated hypertension.

Cardiovascular. About half of diabetic deathsare from the consequences of macroangiopathy.People with diabetes have a twofold greater risk ofstroke and a fivefold greater risk of MI comparedwith matched non-diabetic subjects. Peripheralvascular disease is also common, with a 50-foldhigher risk of peripheral gangrene. Some patientsundergo progressively extensive amputation;usually the lower limbs (especially the feet; seebelow) are affected, but fingers are also at risk.

Hypertension is often associated with diabetes.Up to 50% of type 1 patients have it, and it isprobably secondary to nephropathy. About a

Complications 601

Table 9.13 Clinical consequences of diabetic complications

System Clinical features

Eyes Retinopathy, glaucoma, cataract; blindnessNerves Sensory, autonomic and motor defectsRenal Glomerulosclerosis; chronic renal failureCardiovascular Ischaemic heart disease (angina, MI), peripheral vascular

disease, stroke; cardiomyopathy; congestive heart failure Locomotor Slow-healing peripheral lesions; ‘the diabetic foot’; amputations;

joint stiffnessImmune Increased susceptibility to infection

MI, myocardial infarction.

fifth of type 2 patients are hypertensive; the aeti-ology is uncertain but related to the metabolicsyndrome, with obesity and hyperinsulinaemiacontributing.

A rare complication is diffuse cardiac fibrosis(cardiomyopathy), which may lead to heartfailure.

Locomotor. The ‘diabetic foot’ is a commonproblem. In normal people minor foot injuries,such as a blister or a lesion from ill-fitting foot-ware, usually heal before being noticed. In peoplewith diabetes, however, these often develop intonon-healing painless ulcers that become infectedand irreversible damage sometimes occurs beforemedical attention is sought. In some cases thisresults in osteomyelitis or gangrene, both ofwhich can lead to amputation. This results froma combination of poor peripheral sensation(neuropathy, so that the wound is not felt), poorperipheral circulation (angiopathy, so thathealing is impaired) and reduced resistance toinfection. All people with diabetes should see achiropodist regularly. Correctly fitting footwear isessential. No pharmacist should attempt to treatany foot problem in a diabetic, or sell them ‘cornplasters’ or similar products. Any foot problem,however minor, should be referred to theirchiropodist or doctor urgently.

Diabetes can also cause soft tissue damageresulting in limited joint mobility (stiffness), anda characteristic arthropathy, usually in the feet,where angiopathy and sensory neuropathy alsocontribute (Charcot joints; see Chapter 12).

Systemic. People with diabetes are very proneto infections owing to an impaired immuneresponse caused by defects in immune andinflammatory cells. Recurrent bladder infectionis common, which can ascend to causepyelonephritis: urinary retention and stasis dueto autonomic neuropathy exacerbate this. Skininfections are also frequent, and contribute tofoot problems.

Management of complications

General strategyThe overall approach to preventing diabeticcomplications, minimizing them or delaying

their onset combines control of blood glucose,risk factor reduction and regular monitoring.

Optimal glycaemic control. Although theaetiology and pathogenesis of the complicationsare still uncertain and likely to be multiple, themain clinical approach has been to aim forscrupulous control of blood glucose levels,keeping them within the normal range, in anattempt to mimic physiological normality. Thisis based on the assumption that complicationsare due to hyperglycaemia. This seems to beparticularly likely for the microvascular, possiblypolyol-related, complications in nerves, eyes andkidney. Evidence derives from clinical trials,including those using the more ‘physiological’treatments such as continuous SC insulin infu-sion (p. 624) or other methods of achieving‘tight’ glycaemic control. This means keepingfasting blood glucose levels below 7 mmol/L andnot exceeding 11 mmol/L after meals, and maynecessitate conversion to insulin therapy inpoorly controlled type 2 patients.

Good control has been shown to reduce the incidence of complications. The mostconvincing evidence in type 1 diabetes was theDCCT trial, which reported significant slowingof deterioration in retinopathy, microalbumin-uria and, to a lesser extent, neuropathy. TheUKPDS trial found broadly similar benefits intype 2 patients and also strongly demonstratedthe synergistic role of hypertension in exacer-bating complications and the importance ofachieving normotension as well as normogly-caemia. Unfortunately, this study failed to iden-tify clearly the treatment mode that offered thebest protection, although this had been one ofits aims.

An unwanted side-effect of tight control is thatby keeping the average blood glucose low theincidence of hypoglycaemia is increased, espe-cially among elderly and unstable diabetics. Inthe DCCT trial there was a threefold increase inthe incidence of hypoglycaemia when undertight control. This means that in some circum-stances a compromise is necessary because of theacute and the long-term complications offrequent hypoglycaemic attacks. Thus, olderpatients in whom the diabetes onset occurredquite late, i.e. type 2, are usually allowed to run


higher average levels. The long delay in onset ofcomplications will mean that life expectancymay be little reduced, whereas quality of lifewould be markedly reduced by frequenthypoglycaemia.

For the macrovascular complications (cardio-vascular, cerebrovascular and peripheral athero-sclerosis) this approach is less successful, perhapsbecause insulin and related endocrine abnormal-ities and hypertension may contribute directly,independently of glycaemia. It is still unknownwhether the generally higher insulin levels asso-ciated with tight control regimens can actuallyexacerbate some macrovascular problems.

Minimize risk factors. It is important tocontrol any additional risk factors that couldexacerbate organ damage, especially via athero-sclerosis. These include smoking, hypertension,obesity, hyperlipidaemia and hyperuricaemia.

Monitoring. This essential component inminimizing complications is discussed below(see also Table 9.22).

Reduce polyol accumulation. According tothe polyol hypothesis for certain of the compli-cations, it should be possible to impede this process by interfering with the metabolismof polyols. Unfortunately, aldose reductaseinhibitors (e.g. sorbinil), although they do minimize sorbitol accumulation and preventmyoinositol depletion, have not proven clini-cally successful in reversing or even retardingneuropathy, cataract, nephropathy or retino-pathy. Dietary myoinositol supplementation hasalso been unsuccessful.

Specific complicationsNephropathy. There are currently fourmethods that have been shown to slow the rateof deterioration in renal function:

• Careful glycaemic control.• Control of hypertension.• Use of ACEIs or ARAs.• Moderate protein restriction (in more

advanced nephropathy).

It is essential that people with diabetes are moni-tored annually for the onset of hypertension and

microalbuminuria. In treating hypertension,ACEIs (and ARAs) seem to have an additionaldirect beneficial effect in diabetes, dilatingintrarenal (efferent glomerular) vessels and thusminimizing glomerular hypertension. ACEIs areincreasingly used early unless contra-indicatede.g. by bilateral renal artery stenosis, which isalways a possibility in someone with diabetes.ACEIs are indicated when there is hypertensionwith proteinuria or microalbuminuria; in type 1diabetes their use is recommended if there ismicroalbuminuria, even with normotensivepatients. However, at present there is noevidence that ACEIs benefit normotensivediabetes with no evidence of nephropathy.Other antihypertensives may not offer similarextra benefits but another antihypertensiveshould be used if ACEIs are contra-indicated orinadequate at reducing pressure.

Once established, renal failure is managed asusual (see Chapter 14), although haemodialysisis more difficult because of vascular and throm-botic complications. Continuous ambulatoryperitoneal dialysis is particularly suitable indiabetes because insulin may be administeredintraperitoneally (thus directly entering theportal circulation, which is more physiological).However, there may be a problem with theglucose, which is usually added to dialysis fluidto promote water removal. People with diabetesare nowadays unlikely to be given low priorityfor renal transplantation, as they tended to be inthe past, and this is sometimes combined withpancreatic transplantation (p. 605). There arehowever some problems: the poor general healthof these patients and multiple organ damageincrease the operative risk, and there is anincreased likelihood of post-transplant infectionowing to the immunosuppression required.Nevertheless, graft survival is only about 10–15%poorer than the average for renal transplants.

Macroangiopathy. The usual dietary constraintson saturated fat and cholesterol are important.Monounsaturated fats, especialy olive oil, arerecommended. The HPS study supported theuse of statins for all people with diabetes ofeither type at cardiovascular risk, whatevertheir lipid level, and this is now accepted. TheCARDS study extended the recommendation in

Complications 603

type 2 diabetes to those patients with evennormal or low lipids, regardless of CVS risk.However, such routine use is not yet officiallyrecommended. In the PROactive trial type 2patients with pre-existing macrovascular diseaseused pioglitazone in addition to their usualtreatment. A small but significant reduction inall-cause mortality, MI and stroke was achievedbut at the expense of weight gain and anincrease in heart failure.

Other conventional atheroma risk factors suchas smoking and hypertension must also bescrupulously addressed (see Chapter 4).

Neuropathy and neuropathic pain. Little canbe done for diffuse neuropathy, but neuropathicpain can be partially relieved and fortunatelysevere attacks, although prolonged, tend toremit. Drug therapy may be of help in the some-times excruciating pain. Conventional analgesicor anti-inflammatory drugs are generally ineffec-tive. A variety of other drugs have been triedand the first-generation tricyclic antidepressants(e.g. amitriptyline) are standard first-line therapy.Second-line agents include anticonvulsants suchas carbamazepine, gabapentin or topiramate (seealso Chapter 6).

Retinopathy. Retinal disease is conventionallytreated by laser photocoagulation.

Management

Aims and strategy

Preventative methods for diabetes are as yetpoorly developed. More progress has been madewith potentially curative surgery. However, atpresent the vast majority of people with diabetesrequire long-term management of establisheddisease.

The cardinal aim of management in diabetes isto keep blood glucose levels within the normalrange; this should produce patterns of glucoseand insulin levels in the blood similar to thosethat follow normal changes in diet and activity(see Figure 9.4). Blood glucose levels shouldremain below the maxima in the WHO defini-

tion for impaired glucose tolerance (Table 9.1).Ideally, this would require a continuous basallevel of insulin to maintain metabolism, supple-mented by rapid pulses following meals and areduced level during exercise.

Optimal management should attain threeimportant interlinked aims:

• Prevent symptoms.• Maintain biochemical stability.• Prevent long-term complications.

At present, this ideal is not achievable. Even ifpancreatic transplantation were to be perfected,insulin receptor defects might still remain.Current therapy is limited to artificially manipu-lating diet and insulin (endogenous or exoge-nous) in order to mimic normal patterns asclosely as is practicable.

The older directive, paternalistic medicalmodel for such manipulation is no longeracceptable, clinics preferring to negotiate a ‘ther-apeutic contract’ with the patient. The aim is toagree a desired level of control – optimal,prophylactic or perhaps merely symptomatic –based on the severity of the disease and thepatient’s age, understanding, likely complianceand normal way of life.

Sometimes it is inadvisable to strive too zeal-ously to approach the ideal. For the elderly,where long-term complications are of lessconcern, keeping symptoms at a tolerable levelwithout excessive disruptions to normal lifepatterns may be adequate. For this, the targetneed only be to achieve random blood glucoselevels below 12 mmol/L. In some patients theincidence of hypoglycaemic attacks is unaccept-ably high if control is too tight. The advent ofthe insulin pen has enabled the flexibility toachieve these differing aims.

Prevention

Because type 1 disease involves immune destruc-tion of the pancreas, immunotherapy has beenattempted experimentally, as early as pos-sible after initial diagnosis or even in the pre-symptomatic stage in at-risk individuals, e.g.where there is a strong family history or impairedglucose tolerance. In animal models anti-Tcell antibodies, bone marrow transplantation,


thymectomy, azathioprine and ciclosporin havebeen tried. In the Diabetes Prevention Trial-1early introduction of insulin therapy, to ‘spare’the beta cells and perhaps to reduce their expres-sion of autoantigens, was unsuccessful. Anothertrial using nicotinamide to inhibit macrophageshas also failed to reduce progression.

However, considerable pancreatic damage hasusually occurred by the time symptoms arenoticed. Only about 10% of functional islet cellsthen remain, so no great improvement can beexpected. Research is now concentrating ondiscovering reliable early prognostic markers,such as islet cell antibodies. Patients at risk couldthen be identified by screening.

No specific aetiological agents have been iden-tified for type 2 diabetes, but risk factors are wellknown. These correspond with many of the well-established cardiovascular risk factors associatedwith the lifestyle of industrialized countries, i.e.diets high in sugar and fats and low in fibre andslowly absorbable complex carbohydrates, lackof exercise and obesity. Weight loss in particularhas been shown to delay development of thedisease in high-risk individuals and achieveremission in severely overweight people withdiabetes. In the Diabetes Prevention Programmeboth intensive lifestyle intervention andmetformin significantly reduced the risk of devel-oping diabetes in people with impaired glucosetolerance. Another trial showed benefit withacarbose. In the Finnish Diabetes PreventionStudy dietary modification and exercise wassimilarly beneficial. More recently the DREAMtrial with rosglitazone over 3 years showed signif-icant reduction in progression from impairedglucose tolerance/impaired fasting glycaemia toovert type 2 diabetes.

Cure: organ replacement

Pancreatic transplants are now a realisticoption. Dual renal plus pancreatic transplanta-tion is especially considered for people withdiabetes with advanced nephropathy, becausesuch patients are going to have to undergoimmunosuppression anyway. One-year patientsurvival exceeds 90% and 5-year graft survivalexceeds 50%. Transplantation substantiallyincreases the quality of life, although of course

is still limited by the risks of surgery andthe penalty of lifelong immunosuppression(Chapter 14).

The implantation of donated beta-islet cells isstill experimental but looks promising. Stemcells may offer even more fundamental a solu-tion for the future. A number of artificialpancreas devices have been devised, althoughnone is yet available for routine use (p. 624).

Therapeutic strategy

Using conventional methods, the only way for adiabetic to enjoy relatively normal eating andactivity (i.e. unpredictable, unplanned anduncontrolled) would be to have frequent,precisely calculated injections of soluble insulin(or appropriate doses of a rapidly acting oralhypoglycaemic [insulin secretagogue] drug). Thedose would be based on blood glucose measure-ment or guided by experience and recent dietand activity level: thus insulin is supplied ondemand in a manner emulating normal physi-ology (see Figure 9.4). With the introduction ofinsulin pens, such an ‘insulin demand-driven’strategy is becoming practicable, althoughdosage adjustment is still imprecise. The artificialpancreas, if perfected, may prove a better option.

‘Insulin supply drive’The alternative (and original) approach, stillused for many older patients, is to turn physi-ology on its head and to accept a model drivenby insulin supply. Instead of matching insulinsupply to instantaneous changes in demand,demand in the form of diet and activity isadjusted and controlled to conform to availableinsulin (whether endogenous or administeredexogenously). Because both drugs and insulinmust be given prospectively this is in effect‘feeding the insulin’, as opposed to the normalsituation where insulin follows feeding. Mealsand activity must be regular and of predictablecomposition: explicit adjustments in drug orinsulin dose must be made to allow fordeviations (Figure 9.8).

This places considerable constraints onpatients, particularly children. Education andcounselling are extremely important andDiabetes UK performs a valuable role here.

Management 605

People with type 1 diabetes are inevitablyreasonable compliers in the strictest sense, inthat the severe metabolic upset precipitated bydrug defaulting is a powerful motivator. Never-theless, excellent compliance with diet, and thevery tight control of blood glucose demandedfor avoidance of long-term complications, is lesscommon, especially in type 2.

Treatment modes

Dietary management is the bedrock of treat-ment. All people with diabetes, irrespective ofother treatments, require some control of theireating and exercise patterns, both in terms oftotal calorific intake, types of nutrients andeating schedule. Indeed, about half of patientswill need no more than this, especially thosewho lose weight. A further 25% will need toaugment their natural insulin with drugs. Theremainder will need insulin.

The initial choice is usually related to how thepatient first presents (Figure 9.9). Youngerpatients, who are frequently non-obese, usually

present unambiguously with type 1 insulin-dependent diabetes, although a variable insulin-independent (‘honeymoon’) period may occurfollowing diagnosis.

Older patients, who are often obese, willalmost always be type 2 and must be tried firston diet alone. Should this fail, drug therapywill be added. All drugs used in diabetes areclassed in the BNF as antidiabetic, and thisterm will be used generically here (althoughNICE refers to these drugs as ‘glucose-loweringdrugs’). The older term ‘oral hypoglycaemic’ isobsolescent, owing to the development ofclasses of drugs that do not directly lowerblood glucose. Those that do, i.e. sulphony-lureas and meglitinides, are more accuratelydescribed as insulin secretagogues.

Type 2 patients are usually to some extentoverweight on presentation, and a biguanide isthe first choice. Otherwise a sulphonylurea isselected. Sometimes a synergistic combination ofthe two types will be required. For those forwhom these measures are ineffective a glitazonemay be added. For some patients even this is


Bloo

d gl

ucos

e (m

mol

/L)

Eveningmeal

0

40

80

24001800120006002400Time (hours)

Blood glucose level

Basal plasma insulin level

Plas

ma

insu

lin (m

U/L

)

Lunch

Mid-morningsnackBreakfast

0

5

10

10

Figure 9.8 Matching food intake to available insulin – schematic representation. In the insulin supply-driven model,insulin levels are maintained artificially either by direct injection or by augmentation using oral hypoglycaemic agents. Toprevent hypoglycaemia, sufficient glucose must be provided by the diet at regular intervals. Note what would be the effectof missing the mid-afternoon snack: blood glucose would start to fall dangerously low just before the evening meal.Unusual activity, by causing increased glucose demand, would complicate this picture.

unsatisfactory and, especially if ketoacidosisoccurs, insulin treatment is needed, as it will beeventually in those whose disease progressesfaster. Type 2 patients may also need insulintemporarily during periods of increased require-ment such as major infection, surgery or preg-nancy. Combining antidiabetic drugs withinsulin therapy is being used increasingly (seebelow).

At any point in this sequence, an adjunctivedrug that reduces intestinal glucose absorptionor reduces insulin resistance may be added.

Initiation of treatment

On first diagnosis, all patients will be fully exam-ined and investigated to establish baselinemeasures for monitoring development andprogression of any complications. This willinclude ophthalmological, renal, cardiovascular,neurological, lipid and foot assessment.

Some patients will need to be treated first inhospital, especially type 1 patients firstpresenting with ketoacidosis. Blood glucoselevels will be measured 3-hourly during thisperiod, to establish the diet and possibly thedrug or insulin dosage necessary to achieve theagreed level of control. After discharge some willcontinue to attend as outpatients. Others will bemanaged by general practice clinics, which often

include specialist diabetic nurse practitioners.However, regular diabetic clinic visits are desirable if they have developed complicationsor management becomes difficult. Some type 1and most type 2 patients without acute compli-cations may be treated by their GP from theoutset.

Diet

Most type 2 patients must first be encouraged totry to control their disease on diet alone, and nopatient taking antidiabetic drugs or insulinshould believe that these obviate the necessity tocontrol their diet. Recommendations about diethave evolved in several important ways. Fats arenow discouraged, while complex carbohydrateand fibre are encouraged, and the overallapproach is now far less restrictive. The recom-mended diabetic diet, save in a few respects, nowclosely resembles the normal healthy diet thateveryone should eat: regular meals low in fats,simple sugars and sodium and high in complexcarbohydrate (starch) and fibre.

Formerly, inflexible, unrealistic or impracticalprescriptions and restrictions (diet sheets,‘exchanges’) took little or no account of thepsychological importance of individual dietaryhabits, dietary preferences and ethnic variations.

Management 607

YOUNG/LEAN(Ketoacidosis;recent weight loss)

OLDER/OBESE(Asymptomatic;polyuric;complication)

DIET ONLY

DIET �insulin

DIET �biguanide

DIET �sulphonylurea

DIET �biguanide �

sulphonylurea �glitazone

DIET �insulin �

insulin sensitizer(a)

DIET �biguanide �

sulphonylurea

Non-obese(uncommon)

Obese(common)

Figure 9.9 Treatment strategy for diabetes related to presenting symptoms. Meglitinides may be substituted for sulpho-nylureas. A glitazone or glucosidase inhibitor can be added or substituted at any stage to attempt to improve controlbefore moving to next stage (but avoid meglitinide + sulphonylurea). See text. (a)Insulin sensitizer = metformin orglitazone.

The result was poor compliance complicated byguilt and anxiety. The modern approach recog-nizes that:

• Dietary records or recall are an imprecise basisfor future modification.

• Nutrient uptake varies even from preciselyregulated and measured portions, owing tothe interactions between foodstuffs, varia-tions in temperature, physical form anddegree of chewing, etc.

• Compromise is needed to devise a regimenwith which the patient can be concordant.

Thus a perfect diabetic diet is difficult to achievein practice, and although the pursuit of it isworthwhile, this could be counter-productive insome patients. Rather, efforts are made to ensurethat patients understand, in their own fashion,what the aims are. Counselling and educationare then used to maximize motivation. Advicefrom a dietician with experience in modifyingdiabetic diets to suit individual lifestyles canhelp achieve good compliance.

Four aspects of diet need to be considered(Figure 9.10):

• Total energy intake.• Constituents.

• Timing.• Variation.

Energy intake

All patients need to adjust their calorific intaketo achieve and maintain the desired bodyweightfor their size, aiming for a body mass index ofabout 22 kg/m2. For most people with type 2diabetes, who are frequently obese, this implies aweight-reducing diet. Reliable tables are nowavailable to predict the required energy intakeaccording to age, gender, activity level andlifestyle.

Constituents

MacronutrientsThe unselective restriction on carbohydrate thatused to characterize diabetic diets is now consid-ered misconceived. Carbohydrate is not harmfulif taken mainly as slowly absorbed complexpolysaccharides, e.g. starch. Such carbohydratesallow people with type 2 diabetes to make bestuse of their limited endogenous insulin secretorycapacity by not raising postprandial bloodglucose too rapidly. Foods can be classified


Reduced total energy content

Main nutrients: energy proportions

Other nutrients: amount, type

Other nutrients: maxima

Timing

Variability

ALL PATIENTS

IF OBESE

Carbohydrate, fat, protein

Fibre

Sugar, alcohol

Small frequent regular meals

(1) Varied content(2) Vary amounts for variations in exercise/stress

Figure 9.10 Dietary considerations in diabetes (see also Table 9.14).

according to their glycaemic index, whichrepresents the ratio of the total glucose absorp-tion they produce compared with that from astandard test meal of wholemeal bread andcottage cheese. The lower the index the better,and representative values are rice 80%, potatoes77%, pasta 60% and lentils 45%. Foodsacceptable to various ethnic minorities, such aschappatis, kidney beans, chickpeas, etc. are alsonow encouraged where appropriate.

The relatively high fat content of early dia-betic diets, which was needed in a carbohydrate-reduced diet to provide calories more cheaplythan with protein, is now seen to be danger-ously atherogenic. A reduced fat intake, low insaturated fats and comprising about one-thirdpolyunsaturated and one-third monounsatu-rated fat (e.g. nuts, fish, olive oil) is nowencouraged. Cholesterol itself is usually reducedinherently along with saturated fats. There areno particular constraints on protein except forpatients with suspected nephropathy, whenrestriction is indicated.

Other nutrientsA small amount of simple sugar (sucrose) is nowconsidered acceptable, if the calorific content isaccounted for. This is usually consumed as aconstituent, e.g. of baked products. Artificialnon-nutritive sweeteners are still preferred andpatients must be advised to monitor their intakeof ‘hidden’ sugar in processed foods. So-called‘diabetic foods’ often contain sorbitol or fructoseand, while they may not raise blood glucose asmuch as sucrose, have a high energy content andcause diarrhoea in excess. They are also expen-sive, offer nothing that a well-balanced diabeticdiet cannot offer, and are not recommended byDiabetes UK.

Alcohol is not prohibited if its high calorificcontent is accounted for and its hypoglycaemiceffect is appreciated, i.e. it should be taken withsome carbohydrate. Recent evidence of its protec-tive effect against heart disease suggests thatonce again similar recommendations shouldapply to the diabetic population as to the popu-lation as a whole. There should be little addedsalt, to minimize rises in blood pressure.

Fibre is extremely important. Although fibreis primarily carbohydrate, the terminology is

somewhat inconsistent; however, the distinc-tions are relevant (Figure 9.11). Starch, in staplefoods like bread, potatoes and rice, is the maindigestible carbohydrate energy source. Olderclassifications grouped all other indigestiblematter together as ‘dietary fibre’, but there areimportant and distinct components. The non-starch polysaccharides (NSP) are now knownto be particularly important in diabetes. Theyprovide no energy but further delay absorptionof glucose from starch digestion (see above), andby forming intestinal bulk promote a feeling ofsatiety that may reduce appetite and thereforehelp weight control.

The (semi)soluble or viscous fibres and gumsfound in fruit, vegetables and pulses (Figure9.11) produce in addition a modest reduction inblood cholesterol, possibly by binding bile saltsand thereby preventing their enterohepatic recir-culation. The insoluble NSP fibres, as in bran andunmilled cereals and grains, have little effect oncholesterol, but contribute to stool bulk alongwith other fibrous residues, e.g. lignin. Althoughundigested in the ileum, some of this material ishydrolysed by colonic flora to release absorbableand metabolizable carboxylic acids.

ProportionsThe recommended proportions of macronu-trient energy intake are approximately 60:30:10(carbohydrate:fat:protein; Table 9.14); tradi-tional diabetic diets used to be nearer 25:65:10.Within the fats, only a third should be saturatedfats. How the patient implements this has alsochanged. Clinics no longer issue rigid menus,kitchen scales and detailed tables of what can beexchanged for what. More generalized recom-mendations with much wider variability arefound to be more successful.

One such approach simply visualizes a mealplate divided into segments (Figure 9.12). Abouttwo-thirds contains polysaccharide: equal partsstaple carbohydrate sources such as rice, pasta orpotatoes starch and fibre such as fruit or vegeta-bles. The remainder is mostly composed of fatsand protein sources such as meat, fish and dairyproducts. A small amount of sugar is allowed.The patient is advised to construct each meal inthese proportions. This roughly conforms to therecommended proportions, allowing for some

Management 609

fat and protein being included along with the carbohydrate.

TimingSmall, regular, frequent meals are important.This means similar calorific intake at all mainmeals and regular snacks in between. For type 2patients this minimizes the load put on thepancreas at any one time. For both types ithelps to keep blood glucose levels within closerlimits, minimizing the risk of hypoglycaemiabetween drug or insulin doses and the risk ofpostprandial hyperglycaemia. There is someevidence that this too is a pattern that mightbenefit the general population. Nibbling or‘grazing’ appears to produce lower averageplasma lipid and blood glucose levels and lessobesity compared with a similar calorific intakeobtained from intermittent, larger meals.

VariationPeople with diabetes need to understand thatthese constraints do not prevent them having avaried, appetizing and nutritious diet. Theyshould also understand how to augment theirdiet to match any unplanned or unusual exerciseor stress so as to avoid hypoglycaemia. Tempo-rary changes in a patient’s metabolic require-ments (as in serious illness) or oral absorptivecapacity (e.g. gastroenteritis) require appropriateadjustment, which may involve temporaryinsulin therapy in a type 2 patient, and regularblood glucose monitoring is then essential.

Type 1 patients using the ‘insulin pen’ willgenerally be even more flexible (see below). Inmildly diabetic elderly patients the diet will alsobe far less rigid, for reasons already discussed. Onthe other hand, the diets of growing childrenneed constant reassessment. The availability of


CARBOHYDRATE

COMPLEXPOLYSACCHARIDE

Simple sugars(a)

Sucrose, dextrins

Digestible polysaccharide

white, brown, chapatirice, pastapotato, cassava

Breads:Cereals:

Vegetables:

Starch

Provides energy

Non-starch polysaccharide

Delays glucose absorption,induces satiety and weight reduction

‘NSP’, ‘dietary fibre’Undigestible or unavailable

Soluble (viscous)

Pectin, gum, alginateGalactomannan

Pulses, oats, barleyGuar gum

Reduced cholesterol

Insoluble

Bulk-forming

(Hemi)cellulosesBran, wheat, rice

(unmilled)Vegetables

Figure 9.11 Different forms of dietary carbohydrate, including fibre; with functions and examples of foodstuffs. (a)Avoidas far as possible.

nutrients and the habits and constraints ofdifferent ethnic groups also need to be takeninto account. Dieticians are an essential part ofthe diabetic team.

Diet as sole management may fail in up totwo-thirds of type 2 patients. Primary failure isusually due to poor compliance, poor motiva-tion or inadequate counselling. Secondaryfailure usually results from disease progression,with falling insulin production. The next stage isto introduce oral antidiabetic drugs.

Oral antidiabetic drugs

Aim and role

Oral antidiabetic drugs (OADs) are used as thenext step for type 2 patients in whom diet hasfailed to control their condition adequately. Themajority may then be controlled by a combina-tion of diet and oral drugs for a number of years,but some type 2 patients may eventually requireinsulin treatment.

Management 611

Fruit and veg(non-starch polysaccharide)

33%

Foods with fat and sugar6%

Milk, dairy (fat)12%

Meat, fish (protein)16%

Bread, cereal, potatoes(starch)

33%

Figure 9.12 The ‘plate model’ of meal planning recommended by Diabetes UK. Each meal should be constructed roughlyof the types of foods and in the proportions shown, visualizing them as making up the complete plate of food. (Adaptedfrom www.diabetes.org.uk/eatwell/food_diabetes/index.html).

Table 9.14 Nutrients in diets recommended for diabetic and general population(a)

Nutrition Subcommittee of the National recommendations Diabetes Care Advisory Committee for optimal UK diet(c)

of Diabetes UK (b)

Carbohydrate 50–60% 50–60%Total fat 30–35% [saturated fat 10%] 30–35%Protein 1 g/kg body weight 10%Simple sugars �10% 60 gCholesterol(d) Not specified Not specifiedSoluble fibre(e) Not specified 18 gAlcohol 2 units (female) 2 units (female)

3 units (male) 3 units (male)

(a) Percentages are rounded and given as maximum proportion of total energy intake. Amounts are per day.(b) Diabetes UK, Diabetic Medicine 2003; 20: 786–807.(c) Based on reports from National Advisory Committee on Nutrition Education (NACNE, UK DoH) and Committee on Medical Aspects Of Food

Policy (COMA, UK DoH).(d) Cholesterol intake usually automatically reduced sufficiently if saturated fat intake less than 10%.(e) 15 g soluble fibre equivalent to 18 g non-starch polysaccharides (NSP) or 30 g total dietary fibre.

There are four main therapeutic targets forOADs (Table 9.15). Doubts over the safety ofsome of these drugs have now been resolved.The results of the University Group DiabetesProgramme (UGDP) trial in the 1970s, whichsuggested significant toxicity in the sulphonyl-ureas, are now discredited. Phenformin, an earlybiguanide, caused numerous deaths from lacticacidosis and was withdrawn. Newer biguanidesare much safer: only metformin is currentlyavailable in the UK; elsewhere buformin is used.

Novel incretin analogues are undergoing trials.Incretin is a newly discovered peptide hormone,secreted in the small intestines following foodintake, which enhances insulin secretion andsuppresses glucagon, slows gastric emptying and reduces food intake. It was isolated from alizard that eats only four times a year. Exenatidehas been shown to lower glycated Hb levels andweight. Sitagliptin inhibits incretin inactivation.

All OAD strategies depend on endogenousinsulin secretion and are therefore effective onlyin patients with type 2 disease who retain appre-ciable beta-cell function. Ketosis-prone patients,patients with brittle disease or those whosefasting blood glucose exceeds 15–20 mmol/L,almost invariably need exogenous insulin, inboth type 1 and type 2 patients.

Action

These drugs have different, albeit comple-mentary and sometimes overlapping, actionson the underlying abnormalities in type 2diabetes, so combination therapy is indicated ifmonotherapy fails.

Alpha-glucosidase inhibitors (acarbose)inhibit the final stage of the digestion of starchwithin the intestine by blocking the enzymedisaccharidase. This reduces the rate of glucoseabsorption and thus the postprandial glucoseload presented to the islet cells. Thus, a pancreaswith a limited insulin secretory rate might bebetter able to handle this load with less hyperglycaemia. It can be regarded as anti-hyperglycaemic rather than a hypoglycaemicagent. It has a relatively small effect onglycaemia and is used only as an adjunct to othertherapy, but may be added at any stage toimprove control.

Sulphonylureas enhance the release ofpreformed insulin in response to circulatingglucose, partly by increasing beta-cell sensi-tivity to blood glucose. This mimics the acutephase of the normal response to hyperglycaemia.However, sulphonylureas do not directly stim-ulate subsequent insulin synthesis. Inhibition of


Table 9.15 Oral antidiabetic drugs

Therapeutic target Site of action Group Examples

Reduce or retard Intestine Alpha-glycosidase Acarboseglucose uptake inhibitors

Enhance insulin Pancreas Sulphonylureas Tolbutamide, glibenclamide,secretion(a) � glipizide, glimepiride, gliquidone

Meglitinides Repaglinide, nateglinide

Enhance insulin action Peripheral Biguanides Metformin, buformin(b)

receptors �Reduce gluconeogenesis Liver

Reduce insulin resistance Peripheral receptors Thiazolidinediones Rosiglitazone, pioglitazone(esp. adipose tissue) (Glitazones)

(a) Insulin secretagogues.(b) Buformin not licensed in UK.

glucagon has also been suggested. Pharmacody-namically, they differ only in relative potency butthere are important pharmacokinetic differencesbetween them. Sulphonylureas can be combinedwith most other OADs except the meglitinides.Althoughsomedoubtwascastoverthesafetyofthelong-established combination with metformin bythe UKPDS, this has not been confirmed and thecombination is stillwidelyused.

Meglitinides (prandial glucose regulators) alsostimulate insulin release but not at the sulpho-nylurea receptor. They are claimed to do so morespecifically in response to the blood glucose leveland thus to mealtime glucose load, making themmore glucose-sensitive. They have two mainadvantages over sulphonylureas. A more rapidonset means they can be given 15 min or lessbefore a meal, giving patients more flexibilityand control; and their shorter duration of actionreduces the likelihood of postprandial hyper-insulinaemia and between-meals hypoglycaemia.In addition, if a meal is missed they can easilybe omitted. Nateglinide has a prompter andshorter action than repaglinide. Currentlynateglinide is only licensed for use withmetformin, whereas repaglinide can be substitutedfor sulphonylureas at any stage. The combina-tion of a meglitinide with a sulphonylurea isirrational.

Biguanides do not stimulate or mimic insulinbut are insulin sensitizers. They have two mainactions: they increase peripheral glucose uptakeand utilization and they inhibit hepatic gluco-neogenesis and release of glucose from the liverinto the blood. The underlying effect is probablyvia a general inhibitory action on membranetransport. Intracellularly, this would preventglucose entering mitochondria, thus promotinganaerobic glycolysis in the cytosol. Because thisis less efficient than aerobic glycolysis, cellularglucose uptake and utilization are increased. Thismay also account for a tendency to cause lacticacidosis. In the intestine, reduced membranetransport may be useful in slowing and reducingglucose absorption. There may also be intestinallactate production. They may also have an anti-obesity action. Only metformin is licensed inthe UK.

Biguanides can be combined with most otherOADs.

Glitazones (thiazolidinediones: rosiglitazoneand pioglitazone) are also insulin sensitizers.They activate a nuclear transcription regulatorfor an insulin-responsive gene (peroxisomeproliferators-activated receptor-gamma, PPAR�),which has numerous complex effects on lipidand glucose metabolism. An important compo-nent is to promote triglyceride uptake andperipheral adipose growth. The effect of this is toreduce triglyceride availability, increase glucoseutilization, reduce insulin resistance and thusreduce insulin levels. They also shift fat fromvisceral, muscle and hepatic sites to peripheraladipose tissue, which although resulting in anincrease in weight, produces a more favour-able cardiovascular risk. This is partly becausethey alter blood lipids favourably, loweringtriglyceride and raising HDL levels.

The PROactive study suggested this group mayreduce complications, both macrovascular (byreducing insulin and lipid levels) and microvas-cular (by reducing hyperglycaemia) complica-tions, but this has not yet been confirmed. Theprototype, troglitazone, was withdrawn soon afterrelease owing to liver toxicity but rosiglitazoneand pioglitazone are safe and effective eitheralone or in combination if other OADs fail toachieve control, although their precise role hasnot yet been determined. Currently NICErecommends that they should not be added assecond-line drugs to either metformin or asulphonylurea, except when these latter twodrugs cannot be used in combination owing tocontra-indications or intolerance.

Biopharmacy and pharmacokinetics

Sulphonylureas are generally well absorbedalthough potential bioavailability differencesmean that patients should avoid changingformulation or brand. Most sulphonylureas aremore than 90% protein-bound (except tolaza-mide, 75%), and so are liable to competitivedisplacement interactions.

There are important differences in clearance,half-life and duration of action, which deter-mine frequency of administration, precautionsand contra-indications. Clearance is usuallyhepatic with subsequent excretion of inactive orless active metabolites (Table 9.16; Figure 9.13),

Management 613

usually renally. The older chlorpropamide ispartially cleared renally and also has activemetabolite, which accounts for its long half-life.Those with inactive metabolites (e.g. tolbu-tamide) generally have the shortest half-lives.Some sulphonylureas have metabolites that arechiefly excreted in the bile, which makes themmore reliant on hepatic function.

The duration of action, or biological half-life,is related to the plasma half-life but is oftenlonger, owing partly to the activity of metabo-lites. Chlorpropamide has too long a duration ofaction and frequently produces between-mealshypoglycaemia; it has little if any role now andis contra-indicated in the elderly. The otherpopular first-generation sulphonylurea, tolbu-tamide, fell from favour because its action wasfelt to be too short, requiring frequent dosing.However, for this reason it may be useful in theelderly, to minimize hypoglycaemia. Most newer

second-generation drugs avoid these problems,but there are wide interpatient variations in thehandling of all sulphonylureas and dose regi-mens must be individualized. Glibenclamide is aspecial case because it is concentrated withinbeta-cells so its biological half-life is considerablylonger than its plasma half-life. For this reason,it too is avoided in the elderly.

Biguanides differ substantially from thesulphonylureas, being poorly absorbed, littleprotein-bound and cleared predominantly byrenal excretion (with about 30% cleared byhepatic metabolism). Metformin has a short half-life and may require thrice daily dosing at higherdoses. However, modified-release preparationsare available for dosages up to 1 g twice daily;higher doses need standard-release therapy.Buformin is longer-acting.

Renal clearance of biguanides may exceedglomerular filtration rate, implying some tubular


Table 9.16 Relative duration of action of sulphonylureas

Relative duration of action(a) Very short Short Medium Long Very long

Daily dose frequency 2–3 2–3 1(–2) 1 1Examples Gliquidone Tolbutamide Glimepride, Glibenclamide Chlorpropamide(b)

glipizide, gliclazide

(a) Approximate descriptive indication of relative durations: precise time will vary from patient to patient.(b) No longer recommended.

CLEARANCE

mainlyHEPATIC

BILIARY(metabolites)

Excretion

RENAL

partiallyRENAL

inactive or lessactive metabolites

active metabolitesor partly unchanged

glibenclamide (50%),glimepiride (50%)

gliquidone

tolbutamidetolazamideglipizide

glimepride (50%)gliclazide

chlorpropamide (80%) chlorpropamide (20%)

Figure 9.13 Clearance and excretion of the sulphonylureas.

secretion. Thus minor renal impairment, un-noticed because of a normal serum creatininelevel, might still permit significant accumula-tion, and renal function monitoring is essentialwith their use.

Meglitinides are rapidly absorbed, reaching apeak within 1 h and have a very short half-life,being cleared and eliminated hepatically. Thismeans they may be useful in controlling bloodglucose in close association with meals.

Glitazones (thiazolidinediones) are rapidlyabsorbed and hepatically metabolized.Although the half-life is less than 24 h, andonce or twice daily dosing is adequate, fulleffect takes at least a week, owing to the slowspeed of fat redistribution.

Alpha-glucosidase inhibitors are notabsorbed, acting slowly within the gut.

Adverse reactions

Sulphonylureas are well tolerated and free fromserious long-term adverse effects. The principalproblem is hypoglycaemia, which may beprotracted and even fatal. A related drawback isthe tendency to produce or maintain obesity.Both effects can be linked to increased insulinlevels, which also are giving concern over apossible exacerbation of macrovascular compli-cations, insulin being a possible growth factor inarterial walls.

Hypoglycaemia may be caused by an overdose,an interaction, a missed meal or unexpectedactivity and occurs more commonly with thelonger-acting drugs (glibenclamide and chlor-propamide), especially in the elderly, who mustavoid them. (The possible compliance advantageis far outweighed by the likelihood that a mealwill be forgotten while plasma drug levels arestill significant.) With the newer, shorter-actingdrugs any hypoglycaemia that does occur is briefand more easily rectified.

Chlorpropamide can occasionally cause a milddisulfiram-like flush with alcohol (due toacetaldehyde dehydrogenase inhibition), andoccasionally hyponatraemia and a syndromeof inappropriate secretion of ADH. Theseeffects, as well as minor idiosyncratic reac-tions, are uncommon with second-generationsulphonylureas.

Meglitinides do not present such risks of hypo-glycaemia and weight gain as the sulphonyl-ureas. No serious class effects have becomeapparent so far.

Biguanides (with the exception of phenformin)cause minor adverse effects, being somewhat lesswell tolerated than sulphonylureas. The nausea,diarrhoea, muscle discomfort and occasionalmalabsorption experienced may be due to themembrane effects inherent in their mode ofaction. Malabsorption of vitamin B12 can occur.Biguanides are best taken with food, the dosebeing increased gradually to improve tolerance.Iatrogenic lactic acidosis, which has a highmortality, occurs rarely with metformin and therisk can be further reduced by careful monitoringof renal and hepatic function and ensuring that it is avoided in patients with renal impairment and hypoxic/hypoxaemic condi-tions such as cardiopulmonary insufficiency.Because biguanides do not release insulin, theycannot cause hypoglycaemia and they do notcause weight gain.

Alpha-glucosidase inhibitors frequentlycause uncomfortable and sometimes unaccept-able or intolerable gastrointestinal problemsowing to the increased carbohydrate load deliv-ered to the large bowel. Subsequent bacterialfermentation causes distension, pain, flatulenceand diarrhoea.

Glitazones can cause a number of problems.Fluid retention results in oedema, and heartfailure in up to 3% of patients: this is potentiatedin combination with insulin. There may also bea mild dilutional anaemia. Hypoglycaemia is rarebut weight gain is common. In view of thehepatotoxicity of the withdrawn troglitazone,monitoring of hepatic function and avoidancein hepatic impairment is needed, but they aresafe in renal impairment if allowance is made forthe fluid retention.

Interactions

Interactions with OADs are potentially seriousbecause the patient’s delicate biochemicalbalance is maintained by a specific dose. Potenti-ation can rapidly cause hypoglycaemia, whereasantagonism could lead, more slowly, to a lossof glycaemic control and a return of polyuric

Management 615

symptoms. Pharmacokinetic interference withabsorption, binding or clearance occurs almostexclusively with the sulphonylureas, when thetemporary introduction of an interacting drugcan alter the free OAD plasma level, with poten-tially dangerous consequences. A number ofdrugs cause a pharmacodynamic interaction by adirect effect on glucose tolerance (Table 9.17).Fortunately, clinically significant problems arerelatively rare, and certainly far fewer than thetheoretical possibilities. Moreover, differentdrugs, especially among the sulphonylureas, havedifferent tendencies to show a given interaction.

Pharmacokinetic potentiationDrugs that increase gastric pH may enhanceabsorption of sulphonylureas. Highly plasmaprotein-bound drugs can theoretically displacesulphonylureas. However, following redistribu-tion and alterations in clearance there may belittle overall change in free drug levels. More-over, the newer sulphonylureas bind to differentplasma protein sites and are less prone to thiseffect. The hepatic clearance of sulphonylureascan be reduced by severe liver disease andby enzyme-inhibiting drugs and enhanced byenzyme inducers; similar considerations apply to


Table 9.17 Important interactions and precautions with antidiabetic therapy

Potentiation → hypoglycaemia Antagonism → hyperglycaemia

Interference with antidiabetic therapy generally(a)

Beta-blockers – mask/may cause hypoglycaemia (Beta-blockers, calcium-channel blockers)(b)

ACEIs – increase glucose uptake CorticosteroidsAlcohol – potentiates hypoglycaemia Thiazide (and loop) diureticsFibrates (Antipsychotics)MAOIs

Interactions with sulphonylureas(c)(d)

Absorption(Antacids, H2-RAs)Binding displacementSalicylates (high doses)

Hepatic clearanceSulphonamides ChloramphenicolWarfarin � enzyme inhibition Rifampicin � enzyme induction

Antifungals: imidazoles AnticonvulsantsLiver failure Excess alcohol

Renal clearanceNSAIDsSulfinpyrazoneRenal impairment

Interactions with biguanidesRenal/hepatic impairmentAlcohol (potentiates lactic acidosis)

(a) Problems possible with either oral or insulin therapy.(b) Entries in parentheses are known to be rare or minor.(c) There are wide variations in the significance of specific interactions with individual oral antidiabetic drugs, and not all possible interactions are

indicated. This table is merely to show possible effects and mechanisms. A detailed text is recommended to ascertain clinical significance of an interaction.(d) Meglitinides have similar pharmacokinetic properties to the sulphonylureas.

ACEIs, angiotensin-converting enzyme inhibitors; H2-RAs, H2-receptor antagonists; MAOI, monoamine oxidase inhibitor; NSAIDs, non-steroidal anti-

inflammatory drugs.

the meglitinides. The glitazones have not beenreported to cause any hepatic enzyme interac-tions. The renal clearance of unchanged drug oractive metabolites of any of these drugs can bereduced by renal impairment and by certaindrugs that cause fluid retention (e.g. NSAIDs).

Pharmacodynamic potentiationAlcohol is directly hypoglycaemic in fastingconditions, and it may also potentiatebiguanide-induced lactic acidosis. Both MOAIsand beta-blockers tend to cause hypoglycaemia;the former may inhibit glucagon secretion and the latter inhibit hepatic glycogenolysis.Beta-blockers can ‘mask’ the effects of hypo-glycaemia as perceived by the patient. Beta-blocker interactions are seen mainly withnon-cardioselective agents if at all, but asidefrom those with propranolol they are rare andusually insignificant. ACEIs enhance glucoseuptake and utilization by cells, although theeffect may diminish with continued therapy andis of uncertain significance.

AntagonismDrugs that induce liver enzymes can increase theclearance of hepatically metabolized sulphonyl-ureas. Various drugs tend to raise blood glucose,either directly or via the suppression of insulinrelease. Paradoxically, given the masking effectreferred to above, beta-blockers can block insulinrelease.

As a consequence of the inhibited disaccharidedigestion, oral treatment of hypoglycaemia inpatients taking glucosidase inhibitors shouldpreferably be with glucose/dextrose rather thansucrose preparations.

Contra-indications and cautions

The main precautions may be summarized thus:

• People with diabetes need to take particularcare when changing dose, brand or type ofantidiabetic medication.

• Medication records should be monitored toidentify the introduction of potentiallyinteracting drugs.

• The elderly are particularly prone to hypo-glycaemia with the longer-acting OADs;

these patients may be forgetful about meals,less able to recognize hypoglycaemia, andless tolerant of it homeostatically andneurologically.

• Alcohol use must be carefully controlled:although initially it may cause hypergly-caemia (owing to its caloric content), itenhances hypoglycaemia and may impair theability to respond to it.

• Alcohol also dangerously enhances thepossibility of lactic acidosis with biguanidesand it causes unwelcome flushing withsulphonylureas, particularly chlorpropamide.

• Some clinicians manage all patients withsignificant renal impairment (common inpeople with diabetes) or hepatic impairment(less common) with insulin.

Selection

CombinationsMost type 2 patients are overweight and abiguanide is the preferred first choice. It is alsosatisfactory for others but a sulphonylurea mightbe started in the non-obese. Patients who fail toachieve blood glucose control on either regimenuse a biguanide in combination with a sulpho-nylurea. Meglitinides, with their faster, shorteraction may be substituted for the sulphonylureaat any stage if the patient prefers it, especially ifthey are tending to suffer hypoglycaemia orweight gain. A glitazone can be added as a thirdagent when dual therapy fails, especially if thepatient has persistent postprandial or between-meals hyperglycaemia, both of which implyinsulin resistance. However, if metformin plus asulphonylurea combined fail to control thepatient, it is likely that they have very little beta-cell capacity left and the introduction of insulinshould be considered rather than adding a thirddrug (see below).

Acarbose could be added at any of these stagesto improve control but has a limited benefit andis often poorly tolerated. Sitagliptin and exenatide(p. 612) are available as third-line agents.

ConstraintsIn addition to these pharmacodynamic consider-ations, the choice of any OAD must take accountof:

Management 617

• duration of action;• mode of clearance;• age;• renal and hepatic function;• tolerance of adverse effects;• patient preference for number of daily doses.

The elderly must avoid the longer-acting drugs,while other patients may have particular reasonsfor preferring more or less frequent dosing. Byanalogy with insulin regimens, a combination ofa single daily dose of a long-acting drug,combined with regular top-up doses of a short-acting one, has been recommended, but is littleused. In general there is little to choose betweenthe sulphonylureas, but patients with renalimpairment might do better with gliquidone(Figure 9.13).

Some patients cannot be controlled onmaximal tolerated doses of combined OADs.This may occur after many years of therapy asthe beta-cell function inexorably declines (i.e.secondary failure), occurring in up to one-thirdof type 2 patients within 5 years of diagnosis.Alternatively some patients present late, whenthere has already been considerable degenera-tion (primary failure). In either case the situationsignifies that there remains insufficient residualbeta-cell function, and exogenous insulinsupplement becomes mandatory.

At that stage small doses of insulin may beadded to OAD therapy to provide a basal level.This may delay the onset of full insulin therapy,and may be preferred by patients anxious aboutfull insulin dependence. When type 2 patientseventually need to be controlled with insulinthey do not of course become type 1, and theymay more accurately be referred to as havinginsulin-requiring type 2 diabetes. Insulin-augmented OAD therapy will be consideredbelow after insulin has been discussed (see p. 627).

Insulin

About two-thirds of people with diabetes aretreated with insulin, about half of whom aretruly insulin-dependent type 1 and othersare type 2 in secondary failure of OAD therapy.

Patients using insulin require much finercontrol of all aspects of management,including diet, activity and dose measurement,than other people with diabetes. There is lessmargin for error because patients rely totallyon the injected dose. In contrast to type 2patients, they lack the small basal insulin secre-tion that, although insufficient to preventhyperglycaemia, keeps the type 2 patient freefrom metabolic complications like weight lossand ketosis.

Aims and constraints

In theory, it should be possible to attainglycaemic control with insulin that closelymimics the natural physiological variations inresponse to food intake, exercise and metabolicrequirement. However, until recently it was notpossible even to approach that.

Recall that natural insulin secretion from thepancreas into the portal vein is finely andcontinuously tuned to variations in bloodglucose level (p. 586; see Figure 9.3): this is verydifferent from the usual exogenous insulintherapy. An approximation might be attainedwith a continuous basal injection plus regular IVboluses of a rapidly acting insulin preparation tocoincide with meals and, ideally, continuouslytitrated against the blood glucose level. Thiswould resemble the natural pattern except forthe portal delivery to the liver. However, such aregimen is impractical for most patients.

Absorption from the usual SC injection sites,whether as depot injections or by continuousdelivery, can vary in any one patient from timeto time and from site to site, particularly withthe otherwise more convenient longer-actingpreparations. Moreover, whereas exerciseinhibits normal insulin secretion, it tends tospeed absorption from an injection site bypromoting peripheral circulation; thus when lessinsulin is required, more is delivered exoge-nously. It is also likely that SC injections admin-istered by some patients are effectively IM nowthat perpendicular injection is recommended,changing absorption characteristics. Alterna-tively, some patients retard absorption byinjecting into fat, which is less painful. Further-more, the clearance of most forms of injected


insulin is generally slower than endogenousinsulin, the half-life of soluble insulin after SCinjection being about 1 h.

Until recently the most common compromisewas to give a mixture of a fast-acting and amoderately long-acting preparation beforebreakfast (e.g. soluble plus lente), perhaps with abooster dose of soluble in the evening. Withappropriate ‘feeding the insulin’ throughout theday (p. 605) acceptable control can be achieved.However, it results in relative hyperinsulinaemia,a tendency to hypoglycaemia during the day andafter midnight (especially if a meal or snack ismissed or there is unanticipated exertion), andhyperglycaemia before breakfast (Figure 9.8).

Three recent advances have brought treat-ment closer to the ideal for many patients.Ultra-short-acting analogues such as lispro allowcloser matching to meals; long-acting analoguessuch as glargine provide more consistent basallevels; and ‘insulin pen’ systems permit easierand more accurate injection.

Insulin types

Developments in insulin technology haveproduced a range of chemically pure, immuno-logically neutral preparations of standardstrength (100 units/mL in the UK and NorthAmerica) with a wide range of pharmacokineticparameters.

Pharmacokinetic differencesFormulations of insulin can be divided into fourbroad groups depending on their duration ofaction; their times of onset and periods of peakactivity also vary considerably (Table 9.18). Thefastest action is provided by solutions of insulin.In solution, insulin molecules normally associatenon-covalently into hexamers, which areprogressively dissociated by dilution in bodyfluids to the active monomer. This process,which delays onset and prolongs duration, canbe accelerated by small rearrangements ofmolecular structure that affect association char-acteristics but not pharmacodynamic activity.Increased duration may also be provided by forming stable suspensions of carefullycontrolled particle size that gradually dissolve ina uniform manner. Alternatively, solubility char-

acteristics can be manipulated. Other chemicalmanipulation produces ultra-long-acting (basal)formulations. A number of premixed formula-tions provide combinations of these properties.

Ultra-short (rapid) action. By substitutingdifferent amino acids at key positions, insulinanalogues have been produced that exist inmonomeric form with little tendency to asso-ciate but retain full activity at insulin receptors.In insulin lispro lysine and proline are placed atpositions B28 and B29 near the end of the Bchain; insulin aspart has aspartic acid at B28.These agents have an onset of less than 15 min,reach a higher peak within about half the time ofconventional soluble insulin (1 h as opposed to1.5–2.5 h) and a duration of action little greaterthan 5 h (as opposed to 6–10 h). Thus, they canbe injected less than 15 min before a plannedmeal, or even just after one has been started; theoptimal time will need to be determined for eachpatient. The advantages include:

• Less imposed delay between injection andfood intake (especially breakfast), and/orreduction of postprandial hyperglycaemia ifdelay not observed.

• Convenient pre-meal bolus doses, as part ofbasal-bolus regimen (below).

• Easier adjustment for unexpected food intakeor missed meal.

• Reduction of between-meals hypoglycaemia,caused in some patients by excessive durationof action on regular short-acting preparations.

• Less reliance on foods with a low glycaemicindex.

However, there is no advantage in using theseintravenously instead of soluble insulin in emer-gencies or as part of a ‘sliding scale’ regimen (seebelow). Patients who switch need careful re-education about the relative timing of injectionand food intake.

Short action. Clear solutions of soluble(neutral) insulin act less rapidly than ultra-shortanalogues and are cleared within 6–10 h. Theyare useful:

• when IV use is required, e.g. for ketoacidosis;• when titrating a newly diagnosed patient’s

requirement;

Management 619

Tabl

e 9.

18A

ppro

xim

ate

phar

mac

okin

etic

par

amet

ers

of in

sulin

pre

para

tions

(a)

Insu

lin p

repa

ratio

nD

urat

ion

of a

ctiv

ity(h

)(b)

Reta

rdin

g ag

ent

12

34

56

78

1012

1416

1820

2224

...

... 3

6

Ultr

a-sh

ort

Lispr

o/as

part

Shor

tN

eutra

l/so

lubl

e 1(c

)

Neu

tral/

solu

ble

2

Inte

rmed

iate

/IZ

S am

orph

ous

Zinc

prol

onge

dBi

phas

ic is

opha

ne 1

(c)

Prot

amin

eBi

phas

ic is

opha

ne 2

Biph

asic

isop

hane

3

Isoph

ane

(NPH

)Pr

otam

ine

IZS

mix

edZi

nc

IZS

crys

talli

neZi

nc

Prot

amin

e zi

nc s

uspe

nsio

nPr

otam

ine

�zi

nc

Basa

lIn

sulin

gla

rgin

e/de

tem

ir

(a)

Dat

a gi

ven

are

only

app

roxi

mat

e co

mpa

rativ

e in

dica

tions

. Act

ivity

in p

atie

nts

varie

s w

ith m

anuf

actu

rer,

dose

, site

and

tech

niqu

e of

inje

ctio

n, e

tc. (

see

Tabl

e 9.

20).

(b)

Dur

atio

n of

bio

logi

cal a

ctiv

ity

Peak

act

ivity

(c)

Num

bers

indi

cate

rep

rese

ntat

ive

diffe

rent

pre

para

tions

ava

ilabl

e in

UK.

IZS,

insu

lin z

inc

susp

ensi

on; N

PH, n

eutra

l pro

tam

ine

Hag

edor

n; B

ipha

sic,

gen

eric

nam

e fo

r ra

nge

of m

ixtu

res

of s

hort-

and

med

ium

-act

ing

prep

arat

ions

.

• in a continuous SC infusion system;• for the temporary insulin therapy of type 2

patients during pregnancy, surgery or severeillness.

Soluble insulin is being replaced by ultra-short-acting preparations when a booster dose isneeded rapidly, or when frequent injections areneeded for patients with brittle diabetes.

Soluble insulin to cover a particular mealshould be injected 15–30 min, or occasionally45 min, beforehand. When newly diagnosedtype 1 patients are being assessed they areusually put on an insulin sliding scale regimen, with 4-hourly soluble insulin doses adjustedaccording to the current blood glucose level.

Intermediate and prolonged action. Manypatients still receive part of their daily insulindose as a depot injection. This is intended toprovide a continuous basal level of insulin formetabolic activity, with little effect on postpran-dial glucose disposal. The particular regimen isdictated partly by life pattern and patient prefer-ence, but ultimately by trial and error. Depotpreparations are formulated by complexinginsulin with either zinc or protamine, a non-allergenic fish protein. This produces a finesuspension that is assimilated at a rate that isdependent on particle size and injection siteperfusion. Being a suspension, it cannot be givenintravenously. Available products span a widespectrum of times of onset, peak activity andduration, allowing flexibility in tailoringregimens (Table 9.18).

The insulin zinc suspension (IZS) range containsan insulin–zinc complex in either crystalline oramorphous form, the latter being more readilyabsorbed. Insulin zinc suspension mixed is 30%crystalline and 70% amorphous, and insulin zincsuspension crystalline is 100% crystalline, withproportionate increases in duration of activity.(Insulin zinc suspension amorphous, which is nolonger available, was purely amorphous andcombined prompt onset with quite prolonged,but rather variable, action). Isophane insulincontaining protamine as the retarding agent alsohas an intermediate activity.

Protamine zinc insulin and insulin zinc suspen-sion crystalline are the longest-acting prepara-

tions available. If an excessive dose of this typeis injected, the hypoglycaemic effect is corre-spondingly prolonged and glucose or glucagoninjection may be needed to reverse it. Becausethe variability in response between differentpreparations increases with the duration ofaction (even in the same patient), these verylong-acting forms are little used unless patientsof long standing are stabilized on them.

A variety of premixed biphasic preparations(compatible combinations, usually of solubleand isophane forms) are available to providefurther flexibility. Some patients mix specificcombinations immediately before injection.

Basal. A more physiological approach toinsulin provision has recently evolved. Thebasal-bolus regimen is designed to provide acontinuous backgound level of insulin supple-mented by bolus doses at mealtimes. Existingprolonged action formulations, while lasting24 h or more, did not provide the requiredconsistency of release: they all tended to give apeak at 6–12 h (Table 9.18). Two different strate-gies have been devised to solve this problem. Ininsulin glargine, amino acid substitutions havechanged the isoelectric point of the moleculefrom below pH 7 to neutral. As a result it issoluble when administered in a slightly acidsolution but precipitates out as microcrystals atbody pH after injection. Subsequent dissolutionand absorption from the depot provides apredictable, consistently sustained action withan essentially flat activity profile for up to 24 h(Table 9.18). In insulin detemir, attaching a C14fatty acid chain to the insulin molecule substan-tially increases reversible binding to albumin inbody tissue, with a similar result.

Purity and antigenicityThere are two significant factors here: chemical,and therefore immunological, similarity tohuman insulin; and contamination with extra-neous antigenic material. Originally, all insulinwas extracted from ox or pig pancreases suppliedby slaughterhouses. (Approval for insulin treat-ment from these sources has been obtained frommost major religions, but strict vegans maypresent a problem.) Beef insulin differs from thehuman insulin polypeptide sequence by three

Management 621

amino acids, and porcine by just one. Thesedifferences affect antigenicity but not hypogly-caemic potency. As may be expected, porcine isthe better tolerated, but neither causes greatproblems.

Contaminants derived from the extractionprocess (e.g. pro-insulin), insulin breakdownproducts and other unrelated proteins, can stim-ulate the production of insulin antibodies, andallergic reactions used to be quite common.Consequently, chromatographic purification isnow used giving highly purified or mono-component animal insulins that cause far fewerproblems.

Human insulin is made either semi-synthetically, by chemically modifying thesingle variant amino acid in purified porcineinsulin (emp, enzyme-modified porcine), orbiosynthetically (crb, chain recombinant-DNAbacterial; prb, proinsulin recombinant-DNAbacterial; pyr, precursor yeast recombinant).Biosynthesis is becoming the preferred processand human insulin now costs less than animalforms.

Unfortunately, the expectation that humaninsulin would be vastly superior has not beenrealized. Anti-insulin antibodies are not signifi-cantly less common with human insulin thanwith the highly purified porcine form, andallergic phenomena still occur, probably due tobreakdown products occurring during manufac-ture, storage, etc. Nevertheless, almost all newpatients are started on human insulin, and use ofanimal-derived insulin is now rare.

Human insulin is slightly more hydrophilicthan animal forms. Thus, although it has anidentical biological action to pork insulin whengiven intravenously, it is assimilated morerapidly from SC sites and acts more quickly inotherwise identical formulations. It is alsocleared more rapidly, possibly by binding moreavidly to those hepatic and renal enzymes thatdestroy it. These differences are slight and onlyrelevant to patients transferring from one formto the other.

Adverse reactions

The chief adverse effects of insulin are hypogly-caemia, injection site problems, immunological

phenomena and resistance. These may bepartially inter-related.

HypoglycaemiaThis is the most common complication ofinsulin therapy and potentially the mostharmful; the clinical aspects were discussed onpp. 596–598. Insulin can cause hypoglycaemiaeither through an excessive (e.g. mis-measured)dose or through an unexpectedly reducedinsulin requirement (most commonly, a missedmeal).

Human insulin has been associated with anapparent increase in the incidence of hypogly-caemic attacks, including some deaths. This wasinitially attributed to a reduced hypoglycaemicawareness, i.e. hypoglycaemia is not morecommon but is permitted to progress morefrequently. The autonomic warning symptomsof hypoglycaemia (see Table 9.12) seemed to beexperienced less intensely or at a later stagewhen using human insulin, perhaps owing toautonomic (sympathetic) neuropathy.

There is no pharmacodynamic rationale forthis phenomenon and it has been suggested thatit is only incidentally related to human insulinuse. The change to human insulin came at a timewhen the need for tighter control becameapparent and aids to this, e.g. injector pens andhome blood glucose monitoring, were devel-oped. Improved control produces lower meanglucose levels and therefore an increased risk ofhypoglycaemia. Thus it is not now regarded as aserious problem of human insulin, although it isstressed that great care is necessary in transfer-ring a patient to human insulin. Close moni-toring is essential and the daily dose may need tobe reduced, particularly when changing frombeef insulin or for patients with a higher thanaverage daily insulin requirement.

Injection site lipodystrophySome patients develop unsightly lumps (lipo-hypertrophy) or hollows (lipoatrophy) atfrequently used injection sites if they fail torotate the sites regularly. These are not due toscar tissue but are caused by local disturbances oflipid metabolism. Lipoatrophy seems to be animmunological phenomenon; immune complexdeposition may possibly stimulate lipolysis in SC


adipose tissue. It responds to changing to a purerform of insulin, initially injected around thedepression. Lipohypertrophy is more commonwith the newer insulins and may result fromenhanced local lipogenesis, a known insulinaction. It is reversed when the site is no longerused. Although patients may prefer to inject atthese easily penetrated, relatively painless sites,such an approach results in delayed and erraticabsorption.

Insulin antibodies and insulin resistanceInsulin antibodies (insulin-binding globulins)occur in up to 50% of insulin-treated patients. Itmight be expected that they would speed theclearance of insulin by forming immunecomplexes that would be eliminated in the usualway by the monocyte–macrophage system.However, on the contrary, insulin antibodiesdelay assimilation and prolong the action and soare potentially beneficial. They are otherwiseusually harmless, although they may sometimesbe responsible for insulin resistance (see below).

Insulin allergy ranges from minor local irrita-tion to, very rarely, full-blown anaphylaxis. Theless serious reactions commonly remit onprolonged use and are minimized by using thehighly purified modern insulin formulations asfirst choice. The size of the insulin molecule isborderline for antigenicity. Hyposensitizationhas been used to treat insulin allergy, byinjecting extremely dilute insulin solutions at

progressively higher concentrations to inducetolerance. Very occasionally, local steroidinjections need to be given with the insulin.

The term insulin resistance tends to be used inan ambiguous manner (Table 9.19). In patho-genetic terms, it refers to one of the commonunderlying problems of type 2 diabetes, namelyreduced receptor sensitivity. As an adverse effectof insulin treatment, it refers to the requirementin some insulin-dependent diabetes for doses ofinsulin far above the physiological norm.

In the latter sense, insulin resistance occursonly rarely and may be defined as an insulinrequirement �1.5 units/kg/day (about 100 unitsdaily in an average patient). There are manypossible causes; probably the most common issimply obesity, but poor injection techniquemay be an unsuspected problem. Insulin resis-tance is less common now with the use ofthe monocomponent and human formula-tions. Treatment involves eliminating anyobvious cause and then gradually switching tohighly purified or human insulin. As a finalresort, systemic steroids, which are themselvesdiabetogenic, may be needed.

Administration

Delivery systemsPen injectors. Multidose insulin reservoirinjector pens are now the most popular deliverysystem. Each pen has a replaceable cartridge

Management 623

Table 9.19 Possible causes of insulin resistance

Metabolic ObesityIncreased catabolic hormonesInteracting diabetogenic drugs (e.g. steroids)

Immunological Anti-insulin antibodiesAnti-insulin-receptor antibodies

Pharmacokinetic or biopharmaceutic Poor injection techniqueIncreased insulinase activityReduced assimilation from injection site

• local enzymic degradation• scar tissue• lipohypertrophy

Genetic Receptor defect

loaded with up to 300 units (3 mL), representingup to 1 week’s supply for some patients. Thereare various forms of metered-dose injectors. Oneautomatically delivers a 2-unit dose for eachdepression of a trigger, i.e. 2 units per ‘click’, asituation that is particularly beneficial to visuallyimpaired people with diabetes; another formpermits full doses to be preset visually on adigital scale, which may be palpable or audible.Most have a maximum deliverable single dose tominimize the risk of overdose. Each type of penshould only be used with the appropriatecartridge. The main advantages are correct dosemeasurement, and hence less error, and thefacilitation of multiple daily dosing as part of abasal-bolus regimen.

Standard syringe. The use of disposable plasticsyringes with fixed needles is no longer the normin the UK. If stored in a fridge, these syringes maybe re-used for up to 1 week, without significantcontamination of the vial contents (whichcontain a bacteriostat) and no increase in skinreactions. Patients change the syringe when theneedle is blunted or the barrel graduationsbecome unclear. Injection through clothing, longpractised by some people with diabetes, has alsobeen reported to not cause significant problems.Patients must use a safe method of contaminatedwaste and ‘sharps’ disposal.

Artificial pancreas. The ideal replacementpancreas has not yet been constructed. Oneexperimental approach involves a feedback-controlled, blood-glucose driven ‘closed loop’system. A sensor in an IV catheter monitorsblood glucose continuously and the results arefed to a microprocessor that calculates theinstantaneous insulin requirement. This drives aportable pump, strapped to or implanted in thepatient, delivering the appropriate dose. Themain problem is in designing a suitably sensitiveindwelling blood glucose sensor. In anotherexperimental system, an implanted insulinreservoir enclosed in a glucose-sensitive gelmembrane permits insulin diffusion in propor-tion to external glucose concentration. Thereservoir is replenished percutaneously.

Continuous SC insulin infusion is a morepracticable but still relatively expensive ‘open

loop’ option, without the automatic dosagecontrol. An external reservoir/pump strapped tothe body delivers a continuous basal level ofinsulin via an indwelling catheter, with meal-time boosts being activated manually. Modifica-tions include an implanted pump, contolled byradio, and the use of an intraperitoneal catheter,which has the theoretical advantage of moreclosely mimicking the natural insulin secretion.Clearly this method would only suit patientswho are able to manage the technology andunderstand the relationship between bloodglucose, diet, activity and insulin dose. However,current prototypes are as yet too bulky, expen-sive and demanding of patients’ motivation forgeneral use.

Other forms. Simple oral administration ofinsulin is impossible owing to intragastricenzymic destruction. Systems are being devel-oped that avoid this but do not require thecomplications of injection. One approach is toincorporate insulin into liposomes that would betaken orally. The lipid coat would act as anenteric coating and the liposomes would beabsorbed unchanged from the gut, as arechylomicrons. Percutaneous jet injection hasalso been tried. Intranasal administration isbeing explored, using a liposomal or polymervehicle. People with diabetes with advancednephropathy on peritoneal dialysis find itconvenient to add insulin to their dialysis fluid.

A metered dose inhaler (inhaled human insulin,Exubera) for pulmonary absorption is now avail-able in the UK. It seems to offer an activityprofile similar to injection with the rapidlyacting insulin analogues but may be moreacceptable to some patients in combination witha single basal insulin dose by injection. Concernsremain over cost and possible lung damage,especially in smokers, who should not use it.Moreover, bioequivalence is an issue for patientsswitching to inhaler, not least because the dose isexpressed in milligrams, 1 mg being equivalentto 3 units.

StorageInsulin should always be kept cool, but isstable at room temperatures for up to 28days. Formulations incorporating polyethylene-


polypropylene glycol, specially developed forprolonged reservoir use, are stable for evenlonger. Thus, insulin may safely be used inpens and continuous SC infusion, etc. andwhile travelling. Pharmacy stocks and patients’reserve supplies are refrigerated (but notfrozen). Before withdrawing a dose, the vialshould be warmed to body temperature andgently mixed by inversion or rotation (but notshaken).

MixingIf a combination of two preparations of differentdurations is required, specially formulatedproprietary mixtures should be used wheneverpossible, and extemporaneous mixing avoided.The insulin zinc suspension formulations areintended to be stable after intermixing butothers are not, and mixtures of these must beinjected within 5 min. One problem is theadsorption of the soluble form onto the retar-dant from the longer-acting one, which mayseriously interfere with the expected rapid actionof the former. The order of mixing is important:the soluble form is drawn up first, then the depotform. This avoids contamination of the wholevial of soluble insulin with zinc or protamine.

InjectionNow shorter needles have become available,deep SC injection perpendicular to the skin is

universally recommended. Most patients copewell, but instruction and counselling when treat-ment is started are clearly important, especiallywith children. Diabetes UK and a number ofinterested manufacturers produce helpfulliterature on this and all other aspects of diabetescare.

Equally important is the need to rotate the siteof injection regularly so that any one site is onlyused once in 10–20 injections. Seven generalareas are recommended by Diabetes UK (upperarms, thighs, buttocks, abdomen), but withinthese areas the precise injection site used on oneoccasion can be avoided on the next; theyprovide a template to assist such variation. Thisminimizes skin reactions, especially lipohyper-trophy. Patients can also use the slower assimila-tion sites, e.g. thighs, for the overnight dose.Sites usually covered by clothing are preferred.Factors that may alter absorption from theinjection site, possibly upsetting control, aresummarized in Table 9.20.

Dosage regimens

An initial dose titration period on first startinginsulin will indicate the total daily doserequired, but decisions on how this is to bedistributed throughout the day require discus-sion with the patient. With fewer, medium-acting injections overall control is poorer and

Management 625

Table 9.20 Factors affecting insulin absorption from injection site

Factor Example Effect on absorption

Slower/reduced Faster/increased

Pharmaceutical Incorrect mixing of delayed and rapid forms �

(especially if injection delayed)Depth of injection � �

Local inactivation Proteolytic enzymes �

Insulin antibodies �

Local perfusion Regional differences (abdomen � arm � thigh) � �

Exercise (in limb sites); massage �

Skin temperature � �

Angiopathy �

Scar tissue �

Lipohypertrophy �

there is the added risk of hypoglycaemia, withthe threat of coma if a meal is missed. However,when multiple injections are linked with closeblood glucose monitoring, aimed at achievinglower glucose levels (so-called ‘tight glycaemiccontrol’) there is the risk of more frequentepisodes of hypoglycaemia. On the other hand,use of multiple short-acting regimens can lead tohyperglycaemia between injections and poorercontrol. For each patient a balance must bestruck that imposes no more restriction on theirlife than they are prepared to tolerate, which asclosely as possible meets their treatment objec-tives. Achieving this is not easy. Factors toconsider are:

• the patient’s pattern of glycaemia (e.g.nocturnal hypoglycaemia, morning hyper-glycaemia);

• age;• severity of complications;• occupation, social habits and routine;• compliance;• physical disabilities;• comprehension of disease, prescribed regimen

and associated equipment;• patient preference;• ethnic and religious constraints.

The blood glucose targets are usually:

• never below 4 mmol/L;• fasting (preprandial) 4–7 mmol/L;• postprandial and bedtime �9 mmol/L.

Specialized units can organize test periods of 24-h blood glucose monitoring via temporaryindwelling sensors to plot the patient’s patternof glycaemia. However, the interpretation ofthese data is complex and it is an uncommontechnique. Somewhat easier is for the patient toperform a short period of self-monitoring andrecording, the results of which can be discussedwith their diabetologist.

There are also more general considerations,especially when first starting insulin. Manypeople have a distaste for injections or fear them,and the psychological stress of accepting relianceon injections for life can be substantial. This ismore of an issue with type 2 patients as theyapproach secondary failure on OADs, which isconsidered below.

The choice ranges from multiple daily injec-tions of short-acting insulin closely co-ordinatedwith eating and activity pattern, to a convenientbut very unphysiological single daily dose of alonger-acting preparation (Table 9.21).

Whatever the regimen, the total daily doserequired is usually 0.5–1.0 unit/kg (about50 units). This is usually divided as 2⁄3 during theday and 1⁄3 at night for minimum frequencyregimens and 50/50 for basal-bolus regimens.

Minimum dose frequency regimenBecause of the potential compliance benefits offewer daily injections, this method used to befavoured. However it is no longer preferredbecause it imposes inflexibility on activitypatterns and mealtimes, and risks both poorcontrol and episodes of hypoglycaemia. Theregimen usually consists of morning andevening doses of a combination of short- andmedium-acting preparations, the relative dosesbeing determined by trial and error.

There are numerous possible variations. Forexample, the morning dose could be a mix ofabout one-third soluble and two-thirds interme-diate-acting forms, which covers breakfast andprovides a sustained level throughout the day.This may be repeated in the evening, or later inthose patients who get serious pre-breakfasthyperglycaemia. Alternatively there may simplybe a booster dose of soluble before the eveningmeal. If one of the commercially availablepremixed combinations can be used it iscertainly convenient, especially with a pen.More flexibility is provided by individuallydetermined combinations, but a pen cannotthen be used.

Some patients can be controlled satisfactorilywith just a single dose of a long-acting form. Thisincludes type 2 patients with significant residualendogenous insulin production in whom OADshave failed, and some elderly patients requiringonly symptomatic relief and for whom the threatof long-term complications is less critical.

Multiple injectionsThese are now preferred for all patients who canmanage to self-inject frequent doses of short-acting insulin throughout the day, before eachfood intake. In addition, an evening dose of


long-acting insulin is given for basal needs. Themost recent variation of this basal-bolus regimenutilizes ultra-short-acting and long-actinganalogues, e.g. Table 9.21, regimen 5.

A multiple injection regimen is especiallyuseful for brittle patients requiring close control,or for temporary transfer of patients to insulin,e.g. type 2 patients during pregnancy or withserious infections. However, many clinics arestarting most new patients on such a regimen,for which injector pens are ideal. Existingpatients are also being converted. Many patientscan, with experience, finely judge the doserequired according to their food intake and exer-cise. Others, more committed, will measure theirblood glucose level immediately before the nextscheduled dose and adjust the insulin doseaccordingly.

The improved, more physiological controlprovided by this type of regimen reduces thedevelopment or progression of complications; insome trials they have even remitted. Such regi-mens can also, if used properly, minimize therisk of hypoglycaemia between meals and ofovernight hyperglycaemia.

Insulin for type 2 patientsMost type 2 patients will eventually need insulinas their beta-cell capacity become exhausted(secondary failure). Owing to the insulin resis-tance common in type 2, their insulin require-ment when they become completely insulindependent will often exceed that of a type 1patient. However, it may be preferable not towait until insulin is absolutely necessary toinitiate treatment. It may be psychologicallypreferable to start patients on a combination oforal agents and small insulin doses. They willinvariably note an improvement in their well-being and can adjust to insulin injections beforebecoming completely reliant on them. Thecombination can also be helpful in difficult tocontrol type 2 patients with high insulin resis-tance, or those with persistent morning hyper-glycaemia. Another situation where thecombination is useful is with patients who arealready using insulin but are showing resistance:adding metformin may reduce their insuinrequirement.

The most logical combination is insulin plusan insulin sensitizer. Metformin is the usual

Management 627

Table 9.21 Examples of insulin regimens

Regimen Before Before Before Bedtime Examples of patient groups suited breakfast lunch evening meal to the regimen

1a Long(a) � � � �

short � Insulin-requiring type 2 (with metformin)

1b � � � Intermediate Some elderly patients

2 Intermediate � Intermediate � Some type 1� short � short

3 Intermediate � Short Intermediate Some type 1� short

4 Short Short Short Intermediate Well-motivated type 1UnstableMorning hyperglycaemia and/or nocturnal hypoglycaemia

5 Ultra-short-acting Ultra-short Ultra-short Long-acting Well-motivated type 1 (Basal-bolus) analogue(b) analogue(c) Many newly diagnosed type 1

(a) Duration of action.(b) E.g. Insulin aspart or lispro.(c) E.g. Insulin glargine or detemir.

choice. Combining insulin with a secretagoguesuch as a sulphonylurea or a meglitinide is irra-tional. One regimen is to add a small dose ofabout 15 units of medium-acting insulin eachevening. When switching to this regimen, OADdoses are reduced.

SummaryDiabetes therapy must be individualizedfollowing regular close consultation betweenpatients and their clinicians. To a certain extentthe optimal result is found by trial and error, butthis must be supported by diligent monitoring ofblood glucose and reporting of all hypogly-caemic episodes and other disturbances ofcontrol.

Monitoring

People with diabetes require self-monitoring oftheir biochemical control, and regular assess-ment by a clinician of the development orprogress of long-term complications. The formerhas recently been much simplified andimproved. Type 1 patients need much closermonitoring than type 2.

Biochemical control

While even moderate control relieves symp-toms and prevents serious biochemical abnor-malities, tight control is believed to be essentialif complications are to be minimized. In general,diligent monitoring is more important for type 1diabetes, but all patients should record all testresults.

Glucose

Urine glucoseThis has been the traditional way of assessingcontrol. A few elderly patients still use the colourreaction based on Benedict’s test for reducingsubstances. It is imprecise, non-specific andcumbersome, even with the ingeniouslyformulated Clinitest reagent tablets.

Urine glucose estimations can never provideprecise information about current blood glucoselevels, particularly low, potentially hypogly-caemic ones. Urinary concentrations will varyaccording to urine volume independently ofblood glucose. Furthermore, aglycosuria doesnot necessarily guarantee normoglycaemia,owing to differences in renal threshold betweenpatients and in the same patient at differenttimes.

Nevertheless, urine testing remains useful as asimple initial screen and for type 2 patients notprone to hypoglycaemia when tight control isnot essential, e.g. the elderly and patients averseto repeated skin puncturing. A few patients maybe monitored adequately by regular urinalysisand occasional blood glucose measurements,once the relationship between the two has beenestablished.

Urinary glucose measurement also has theadvantage that timing is less important thanwith blood testing because urine concentrationreflects control over the previous several hours.Thus, newer glucose oxidase-based urinedipsticks have been developed that are morespecific for glucose and far more convenientbecause they can simply be passed through theurine stream.

Blood glucoseThere are three main uses for blood glucosemonitoring: to detect hyperglycaemia or incip-ient hyoglycaemia; to monitor closely in timesof changing glucose/insulin need (e.g. intercur-rent illness); and to determine a new patient’sdiurnal glucose profile so as to construct anappropriate insulin regimen.

Most patients, especially with type 1 disease,measure their blood glucose directly using a dropof blood from a finger prick on a glucose oxidasestick. This provides an immediate measure ofglycaemia that is reasonably accurate and reli-able, not overly prone to error from poor tech-nique, and easy to read. Sticks for unaided visualreading are being replaced by ones to be insertedinto automated meters that display the resultdigitally and may give audible warnings. Somemeters can store the most recent results, forreporting at clinics.


Various spring-operated skin puncture devicesmay be used to help obtain the blood drop easilyand safely, and percutaneous techniques ofmeasurement are being developed.

A few type 1 patients regularly test four timesdaily, including at the lowest points, before mealsand in the morning, and at the high point aftermeals. This is necessary only in the more erratic,brittle patients, in intensive multiple-dose regi-mens in younger patients, or when previouslywell-controlled patients start to experience prob-lems. Others, once stabilized, will test randomly afew times weekly and some may perhaps useurine dipsticks daily. The main guideline is toidentify a patient’s risk times (e.g. between-mealhypoglycaemia or postprandial hyperglycaemia)and subject those to special scrutiny.

Once type 2 patients have become stabilized,weekly or even monthly fasting blood glucosemeasurement is usually sufficient.

Dose modification falls into two basic strate-gies. Well-motivated patients on the basal-bolusregimen who are suitably trained by the diabeticteam will be able to modify their next insulindose, based on the results of their preprandialglucose level and the glycaemic content of theirnext meal. This is termed ‘dosage adjustment fornormal eating (DAFNE)’.

For patients who prefer regular dosing,frequent changes following this apparentlylogical DAFNE strategy is both inconvenient andinappropriate. More systematic is to note pre-meal blood glucose over several days. If it isconsistently unsatisfactory, they must alter theprevious scheduled dose on future days, becausethe current pre-meal level is a reflection of theprevious dose.

Glycated (glycosylated) haemoglobin

The abnormal, quantitative glycation ofsystemic protein as a consequence of excessblood glucose (p. 599) applies also to bloodproteins, including Hb and albumin, as well as toplasma fructosamine. Because these substancesremain in the blood for long periods (120 daysfor Hb, 7–14 days for the others), their glycationgives a long-term, integrated picture of bloodglucose levels over those periods. This can be

measured at diabetic clinics and is useful intracing any problems with control that mightnot be revealed by patients’ tendency to be extrameticulous on the few days before each clinicvisit.

Care must be taken to ensure adequate timebetween successive measurements, especiallyafter a treatment change. This allows the level torestabilize, bearing in mind the normal 120-dayred cell lifespan. A reading taken too soon willgive a falsely high reading because the glycosy-lated red cells originally measured will not havedied. One the other hand, when there is areduced red cell number or increased cellturnover, e.g. in haemolysis or blood loss, a falsely low reading may be given.

The glycated Hb level gives the best index ofthe control needed to minimize complicationsand is now regarded as the ‘gold standard’. Non-diabetics have about 5% of glycated Hb (HbA1c)and the target level for optimal diabetes controlis currently �7.5%, or �6.5% in patients atincreased arterial risk, e.g with hypertension.

Ketones

Regular ketonuria monitoring is unnecessary fortype 2 and most type 1 patients, but is essentialin brittle ketosis-prone people with diabetes, andin all patients during periods of metabolic stresssuch as infection, surgery or pregnancy. Greataccuracy is not required and urine dipsticks areadequate because any ketonuria at all in the pres-ence of glycosuria indicates a dangerous loss ofcontrol. Combined glucose/ketone sticks arepreferred, especially as heavy ketonuria mayinterfere with some standard glucose sticks.

Clinical monitoring

In addition to biochemical monitoring, regularmedical examination is important in the long-term care of people with diabetes. This will iden-tify as early as possible the development of anyof the many possible systemic complications.Table 9.22 lists the factors that need to bemonitored at intervals that will vary frompatient to patient.

Monitoring 629

Thyroxine is a simple catechol-based hormonebut it has multiple crucial subcellular actionsessential to life. It is involved with oxygenutilization within all cells, and thyroid abnor-malities have profound effects on most organsystems in the body. Thyroid disease is one ofthe most common endocrine disorders, yetfortunately it is relatively straightforward to treatand to monitor. This is partly because thethyroid axis is largely independent of otherendocrine systems and partly because the longhalf-life of thyroid hormone means that dosingis not as critical as for insulin replacement. Thusfrequent variations in thyroid hormone levels donot normally occur and the acute disturbancesof control seen with abnormal insulin andglucose levels are rare. Furthermore, long-termcomplications are few and uncommon.

Physiological principles

Synthesis

Thyroid hormone is synthesized from thearomatic amino acid tyrosine (closely related tophenylalanine and cathecholamine) in thethyroid gland, which sits across the trachea inthe front of the neck. Iodine is an essential ingre-dient, and the conversion from dietary inorganiciodide to iodinated thyroid hormone is termed

the organification of iodine. Ionic iodide in theblood is taken up by the thyroid gland by anactive sodium/iodide symporter (Figure 9.14)then, catalysed by thyroid peroxidase, oxidizedto give I2, which is then bound to the aromaticring of the tyrosine residues.

Mono- and di-iodotyrosine are covalentlyattached to thyroglobulin within the colloid-filled thyroid follicles, dimerized with anothertyrosine ring, then further iodinated to eithertri-iodothyronine (T3) or thyroid hormone(T4). T3 is five times more potent than T4, but75% of thyroid hormone is synthesized andtransported as T4. This is largely converted intarget tissues to T3. Several weeks’ supply isstored in the gland in the bound form, but it isreleased into the blood as free hormone.

In this chapter the term thyroid hormone(s)will be used when referring to the natural physi-ological hormone. Thyroxine (T4) when used asa drug is officially termed levothyroxine, and tri-iodothyronine (T3) when used as a drug as liothyronine.

Control and release

Control of thyroid function is an example of aclassic endocrine negative feedback loop, whichenables fine control of many body systemsaccording to need. A relatively simple peripher-ally active hormone (thyroid hormone) is


Table 9.22 Regular assessment in diabetic clinic

System Test or examination

Biochemical Glycated haemoglobinBlood lipids, body weight

Feet Chiropody; pulsesEye Fundoscopy, acuity, cataractRenal Proteinuria, creatinine (clearance)Neurological Detailed sensory, motor and autonomic neurological examinationCardiovascular Blood pressure, ECG, peripheral perfusion (pulses, etc.)

Symptoms of ischaemia

Thyroid disease

released from an endocrine gland that is its siteof synthesis, stimulated by a peptide trophichormone (thyroid-stimulating hormone,thyrotropin, TSH) from the pituitary and alsounder CNS influence via a releasing hormone(thyrotropin-releasing hormone, TRH) fromthe hypothalamus (Figure 9.15). Both thesynthesis and the release of trophic and releasinghormones are inhibited by the active hormone.

TSH is a 221-amino acid glycoprotein withreceptors on the thyroid that mediate both thesynthesis and the release of thyroid hormones. Itis the main physiological control on thyroidfunction, an important clinical indicator ofthyroid malfunction and a component in the aetiology of some thyroid diseases. Hypo-thalamic control via the tripeptide TRH is lessimportant because low thyroid hormone levelscan stimulate TSH release directly, but it enablesthe CNS to exert an influence on thyroid function; it is particularly concerned withtemperature control. Disease of this arm of thethyroid axis is rare.

Distribution and metabolism

Approximately 80 lg of thyroid hormones arereleased daily, peaking overnight when TSHlevels are highest. It has a biological half-life ofabout 7 days, being cleared by de-iodination inthe liver and kidneys. It is carried in the bloodalmost entirely bound, mostly to thyroid-binding globulin; only 0.02% is carried as freeT3 and free T4 (FT3, FT4). However, only the freehormones are biologically active.

Actions of thyroid hormone

Thyroid hormone enters target cells and afterconversion to T3 interacts with nuclear receptorsto influence the expression of genes coding forproteins invoved in energy metabolim, oxygenconsumption and general tissue growth; thus ithas far-reaching effects on metabolism, growthand development (Table 9.23). In some ways theaction resembles that of catecholamines (e.g

Physiological principles 631

Thyroid gland

Na/l symporter

I�

bloodI� I2

Free T4, T3

Free T4, T3

(0.02%)

MIT

Tyrosine

DIT

Thyroid peroxidase Thyroid peroxidase

Serum

Thyroglobulin

T4 T3

Thyroglobulin

T4 T3

Thyroxine-binding globulin

Figure 9.14 Thyroxine synthesis. I, iodine; MIT, mono-iodotyrosine; DIT, di-iodotyrosine; T3 tri-iodothyronine; T4, thyroxine;Na/I symporter, see text.

adrenaline), to which it bears a structural resem-blance, but the effect is far more prolonged andmore fundamental, whereas the catecholamineshave a very brief action.

Metabolism and growth. Thyroid hormonehas a generally catabolic effect, stimulatingmetabolism and increasing oxygen consump-tion, basal metabolic rate and body temperature.However, in children there are anabolic effectsleading to protein synthesis and growth. Carbo-hydrate absorption is increased and plasma lipidlevels fall.

Cardiovascular and renal. There are inotropicand chronotropic effects mediated via up-regulation of numerous systems, includingbeta-receptors. In addition the calorigenesis pro-motes peripheral vasodilatation and secondary

fluid retention to maintain cardiac output andblood pressure.

CNS. The action on the CNS is knownmostly through the consequences of thyroidmalfunction, considered below. There areimportant effects on mentation and CNS devel-opment. However, little is known of the precisemechanisms.

Investigation

The three key parameters of thyroid function areserum levels of FT3, FT4 and TSH. Older testsmeasured protein-bound iodine and totalthyroid hormone, but now the precise radio-immunoassay of free hormones and TSH gives afar better correlation with physiological and


Hypothalmus

Anteriorpituitary

TSH

T3 (T4)

Negativefeedback

loop

T3 (T4)

Thyroid

serumthyroxine[T4, T3]

TRH

Clearance viahepatic and

renalmetabolism

Metabolicaction

Figure 9.15 Control and release of thyroxine (hypothalamic-pituitary-thyroid axis). –––––● inhibit; –––––●● stimulate; T3,tri-iodothyronine; T4, thyroxine TSH, thyroid-stimulating hormone; TRH, thyrotropin-releasing hormone.

clinical status. It is possible to measure thebinding proteins, and several factors canchange binding, but the feedback control issensitive and precise, so FT4 levels tend to bevery stable. It is not usually necessary tomeasure TRH. The measurement of FT4, FT3and TSH together is know as a thyroid func-tion test (TFT). It has become clear that TSHlevels are the most important index of thyroidstatus, and management now stresses normalTSH levels.

In the initial investigation of thyroid diseasean autoantibody screen should be done for anti-thyroid antibodies, and also for anti-intrinsicfactor and anti-gastric parietal cell antibodies,because there is an association with otherautoimmune diseases including perniciousanaemia. Liver function, lipid profile, bloodglucose and full blood count are also necessary.

The possibility of primary hypothalamic orpituitary disease should always be borne in mindwhen thyroid dysfunction is detected, particu-larly hypothyroidism. In this case both thyroidhormone and TSH levels will be low.

Thyroid disease

Normal thyroid function is described as euthy-roidism. Hypothyroidism (underactivity) andhyperthyroidism (overactivity) are aboutequally common and together constitute the

most prevalent endocrine abnormalities. Usuallythe cause is idiopathic, often involving autoim-muity, although iatrogenic causes occur. Detec-tion and diagnosis are usually straightforward,and management of hypothyroidism is alsosimple. Hyperthyroidism is more complex tomanage and may develop complications.

There are several potentially confusing aspectsto thyroid disease. Firstly, certain aetiologicalfactors, such as autoantibodies, amiodarone andiodine, are common to both hypo- and hyper-thyroidism; similarly, an enlarged thyroid gland(goitre) can occur in both. The action ofiodine/iodide can seem paradoxical, causingeither stimulation or inhibition in differentcircumstances. Long-term hyperthyroidism caneventually evolve into hypothyroidism, andsome forms of hypothyroidism can have ahyperthyroid phase.

Hypothyroidism

Aetiology and epidemiology

Hypothyroidism is far more common in womenthan in men (prevalence 1.5% vs 0.1%) andmore common in the elderly, although it canaffect the very young and is then far moreserious. It is usually due to intrinsic thyroidgland disease although rarely it may occur

Hypothyroidism 633

Table 9.23 Actions of thyroid hormone

System Effect

MetabolismCalorigenesis oxygen consumption and oxygen dissociation

basal metabolic rate, temperatureLipid ↓ total cholesterol ( hepatic LDL receptors)Carbohydrate absorption from GIT

Cardiovascular Inotropic: beta-receptors up-regulatedheart rate, cardiac output

Peripheral vasodilatationRenal Fluid retentionNeurological Essential for nervous system development

(No calorigenic effect in brain)Growth Essential for normal growth, including bone

↓↓

↓↓

↓↓↓

secondary to hypothalamic or pituitary disease,or to drugs (Table 9.24).

Simple atrophy is the commonest cause,mainly affecting elderly women. There may bean autoimmune component as it is sometimesassociated with other autoimmune disease, butno antibodies are found. Autoimmune destruc-tion is the main cause of Hashimoto’sthyroiditis, which can affect the middle-agedand elderly. Also common is hypothyroidismsecondary to the treatment of hyperthyroidism(see below).

In the developed world dietary iodine defi-ciency is now almost unknown, partly owing toiodination of salt, but it is far more common indeveloping countries. Congenital hypothy-roidism secondary to maternal iodine deficiencyaffects the developing nervous system of thefetus to produce cretinism.

The term myxoedema is sometimes used as asynonym for hypothyroidism but more preciselydescribes one characteristic dermatological sign.

Pathology

Low levels of thyroid hormone compromisemany crucial metabolic processes, as can beinferred from Table 9.23. There is a generalslowing of basal metabolic rate, a fall in temper-

ature, and slowing of physical and mentalprocesses. More detail is given on p. 635.

Investigation and diagnosis

The standard thyroid function test is definitive.When thyroid hormone levels fall there isalmost invariably a compensatory rise in TSH.However, a small rise in TSH may precede bothclinical signs and a fall in thyroid hormone levelby many months; this is known as subclinicalhypothyroidism (see below).

In rare hypothalamic-pituitary causes thecombination of low FT4 and low TSH levels isdiagnostic.

Screening for autoantibodies to thyroid perox-idase or thyroglobulin is not necessary for diag-nosis but can indicate a possible cause, and canact as an alert for possible autoimmune compli-cations in Hashimoto’s thyroiditis. Occasionallythere may be anti-TSH receptor antibodies with ablocking effect, although such antibodies areusually stimulant, causing hyperthyroidism(Graves’ disease, see p. 637).

Hypothyroidism is often diagnosed followingvague generalized complaints of tiredness andlack of energy. However, these common symp-toms can of course have many other causes,which sometimes makes diagnosis of mild


Table 9.24 Causes of hypothyroidism

Aetiology Example

Common (90% of cases in developed countries)Atrophy (idiopathic) Commonest form; no goitreAutoimmune destruction Hashimoto’s thyroiditisIatrogenic Secondary to treatment of hyperthyroidism (antithyroid

drugs, surgery, radiotherapy)

Less commonDietary Iodine deficiency where natural level low; goitre presentCongenital Cause of cretinism (1/4000 in UK)Iatrogenic Lithium, amiodaroneSecondary Hypothalamic or pituitary disease

disease problematic. Thyroid disease shouldalways be borne in mind as a differentialdiagnosis of depression in the elderly.

Clinical features

Most of the features of hypothyroidism can beunderstood from a knowledge of the physiolog-ical action of thyroid hormone (Table 9.25). Theoverall clinical impression is of slowness and

dullnes of intellect combined with an unprepos-sessing appearance. Therefore a history from arelative might be helpful, to identify recent orspecific changes, which may be less apparent tothe patient because onset is usually insidious.The two most common erroneous diagnoses inmild disease would be simple ageing, owing tothe slowness, stiffness and general aches andpains, or depression.

The most characteristic symptoms are thegeneral physical and mental sluggishness,

Hypothyroidism 635

Table 9.25 Principal clinical features of hypothyroidism

Common feature Less common

Systemic/metabolic TirednessWeight gainCold intolerance Hypothermia, cold peripheryGoitreHyperlipidaemia

Haematological Anaemia – iron deficiency– pernicious/macrocytic– normochromic normocytic

Cardiovascular Bradycardia Pericardial/pleural effusion HypertensionHeart failure

Dermatological Dry, thick skin Dry, thin hair; alopeciaMyxoedema (non-pitting oedema)Periorbital oedemaVitiligo

Musculoskeletal Delayed relaxing reflexes Peripheral myopathySlow movement Carpal tunnel syndromeMyalgia, arthralgia, stiffness Ataxia (unsteadiness)

DeafnessHoarseness

Neuropsychiatric Depression DementiaSlow thought PsychosisPoor memory

Gastrointestinal Constipation Anorexia

Reproductive Menorrhagia/oligomenorrhoea InfertilityDelayed puberty

Developmental Growth retardationMental retardation; cretinism

lethargy, intolerance of cold, weight gain andcoarsening of the skin. The voice is hoarse andhair is dry, brittle and falling. There may be acharacteristic swollen thyroid, visible in the neckas a goitre.

The classic dermatological feature ismyxoedema, which is an accumulation ofmucopolysaccharide in the dermis that causeswidespread skin thickening and puffiness. Thisform of oedema is non-pitting because it is notcaused by excess fluid accumulation (contrastwith the pitting oedema of heart failure; seeChapter 4, p. 184).

The heart rate is slowed and this may causeheart failure. The periphery is cold.

Thought processes and memory are impairedand mild depression is common. There is usuallyweight gain and constipation, despite anorexia.

Biochemically, in addition to abnormalthyroid functions tests (low FT3 and FT4, raisedTSH), there is usually hyperlipidaemia andpossibly abnormal liver enzymes. Haematology(see Chapter 11) may show a mixed picture ofiron deficiency (hypochromic, microcyticanaemia), folate and/or B12 deficiency (macro-cytic anaemia) or simply a normochromic,normocytic pattern.

Subclinical hypothyroidism

In some patients there are few if any symptomsand FT4/FT3 levels are within normal limits butTSH is elevated; this might be identified as achance finding. The pathogenesis is probably anearly stage of thyroid insufficiency being compen-sated by slightly elevated TSH level, initallykeeping thyroid hormone levels adequate. Even-tually the slowly progressive nature of idiopathichypothyroidism will lead to frank insufficiencythat does not respond to increasing levels of TSH:thyroid hormone levels then fall and symptomsdevelop. Regular monitoring is all that is requiredduring the asymptomatic phase.

Complications

If thyroid hormone levels are corrected there isno reduction in life expectancy and there are nolong-term problems.

Heart failure or ‘myxoedemic’ coma can beprecipitated by severe metabolic stress, such astrauma, infection or hypothermia, which mayacutely increase thyroid hormone require-ment. Psychosis can also occur (‘myxoedemicmadness’). If the fetus is exposed to inadequatethyroid hormone in utero, irreversible neurolog-ical damage leads to cretinism. Hypothyroidismin children results in retardation of mental andphysical development that is partially reversibleon thyroid hormone treatment. Newborn areroutinely screened.

Management

The management of hypothyroidism is relativelystraightforward, simply requiring oral thyroxinefor life. The general aim is to restore T4, T3 andTSH levels to within the normal ranges. TSHshould not be suppressed too much in anattempt to maintain T4/T3 at high–normal levels:this represents overtreatment and can lead tolong-term cardiovascular complications. Thus amid-range TSH level is usually regarded as theprimary objective, ensuring of course that T4/T3are also within range. However, low-end TSHlevels are regarded by some as preferable.

Levothyroxine

This is the synthetic replacement drug used formaintenance therapy, which is identical tonatural thyroxine (T4). (This has completelyreplaced the original dried thyroid gland, anatural product derived from animal sources,with all the quality control risks these entail.)Levothyroxine is well absorbed on an emptystomach, but absorption is delayed and possiblyreduced by food. Dosing is not nearly as criticalfor levothyroxine in hypothyroidism as it is for


insulin in diabetes, because levothyroxine has ahalf-life of about 7 days and a gentledose–response curve. Moreover, day-to-dayrequirements do not change even with intercur-rent illness, nor do they tend to alter over thelong term. Owing to the natural diurnal varia-tion of TSH secretion, which peaks overnight, asingle dose is usually taken each morning beforebreakfast.

Levothyroxine is initialized at 50 lg daily andincreased by 50 lg daily every 2–4 weeks depend-ing on response. Clinical improvement is usuallyevident within the first month of therapy.Thyroid function testing is required 6 weeks aftereach dose change. Most patients are stabilized on100–200 lg daily; subsequently only annual TFTswill be needed.

More care is needed when initializing treat-ment in the elderly or those with known IHD,using a lower starting dose, e.g. 25 lg on alternatedays, and smaller increments, because the cardiacover-stimulation could precipitate ischaemicsymptoms or even an MI. Sometimes liothyronine(T3) is used for its shorter half-life, permitting amore rapid correction of overdosing. RegularECGs are advisable and beta-blocker cover may beneeded to limit the heart rate.

Side-effectsThe adverse effects of excess levothyroxine (thyro-toxicosis) are exactly what would be predictedfrom the physiological action of excessthyroxine and are described below (p. 640). Withoverdosage, as with untreated hyperthyroidism,there is the possibility of osteopenia or osteo-porosis in women, which should be monitored.

Cautions and interactionsThe dose may require increasing in pregnancy.Hepatic enzyme inducers (e.g. rifampicin, pheny-toin) increase clearance. Some drugs reduceabsorption, so levothyroxine should be taken at adifferent time from sucralfate, aluminiumhydroxide and iron salts (Table 9.26). Otherfactors that affect the control of hypothyroidismare also shown in this table.

Liothyronine

Liothyronine (tri-iodothyronine, T3) has a swifteronset and shorter half-life than levothyroxine andit is about five times more potent. It is mainlyused for emergency treatment of severelyhypothyroid states such as coma, or for initi-ating treatment in those with CVD. It is availablein injectable and oral forms.

Hyperthyroidism

For several reasons, hyperthyroidism is notsimply the opposite of hypothyroidism. Thecauses are more diverse, there are more potentialcomplications and there are more treatmentoptions with worse side-effects. Note that theterm thyrotoxicosis is used to describe thesyndrome resulting from excess thyroidhormone levels, but hyperthyroidism refersspecifically to when the syndrome is due toexcessive secretion from the thyroid gland.

Aetiology and epidemiology

Hyperthyroidism is about 10 times morecommon in women, in whom the point preva-lence is about 1%. However, the lifetime inci-dence in women is over 2%, some forms beingacute or reversible.

Graves’ disease, caused by IgG auto-antibodies that stimulate the TSH receptor, is thecommonest form, representing some 75% of allcases (Table 9.27). It typically follows a fluctatingbut progressive course, eventually leading tohypothyroidism, either naturally or as a result ottreatment.

Autonomous growth of multiple, hyper-secreting ‘toxic’ nodules in the thryoid gland isthe second most common form and this is moreoften seen in elderly females, but isolated ‘toxic’adenomas (benign tumours) can also occur. Theseare usually associated with goitre. Occasionally

Hyperthyroidism 637

Tabl

e 9.

26C

autio

ns a

nd in

tera

ctio

ns o

f lev

othy

roxi

ne th

erap

y

(a) F

acto

rs a

ffect

ing

thyr

oid

func

tion

Incr

ease

d th

yroi

d ho

rmon

e ac

tion

or

Redu

ced

thyr

oid

horm

one

actio

n or

thyr

oid

func

tion

or le

voth

yrox

ine

treat

men

tth

yroi

d fu

nctio

n

Inte

rfere

with

thyr

oid

–En

zym

e in

duce

rs (e

.g. r

ifam

pici

n)ho

rmon

e ac

tion

abso

rptio

n (e

.g. a

lum

iniu

m h

ydro

xide

, iro

n, e

tc.)

TSH

sec

retio

n (c

ortic

oste

roid

s, d

opam

ine)

Inte

rfere

with

thyr

oid

statu

sA

mio

daro

ne (i

nhib

it pe

roxi

dase

)A

mio

daro

ne (e

xces

s io

dine

)Lit

hium

(unk

now

n ef

fect

)Lit

hium

(blo

cks

iodi

ne u

ptak

e an

d th

yroi

d ho

rmon

e re

leas

e)Io

dide

/iod

ine

exce

ss, e

.g. o

lder

‘exp

ecto

rant

s’Io

dine

def

icie

ncy

Mon

oval

ent a

nion

s e.

g. p

erte

chne

tate

(TcO

4�),

perc

hlor

ate

(ClO

4�),

thio

cyan

ate

(SC

N�):

com

pete

for

iodi

ne u

ptak

e.Pr

egna

ncy

(th

yroi

d ho

rmon

e re

quire

men

t)

(b) D

rugs

affe

cted

by

Dru

gs w

ith a

ctio

n en

hanc

edD

rugs

with

act

ion

dim

inis

hed

levo

thyr

oxin

e tre

atm

ent

Dru

g in

tera

ctio

nSy

mpa

thom

imet

ic (m

imic

act

ion)

Prop

rano

lol,

digo

xin

(se

rum

leve

l)W

arfa

rin (p

oten

tially

ac

tion

– m

onito

r)In

sulin

/ora

l hyp

ogly

caem

ic (

gluc

ose

tole

ranc

e)↓

↓

↓

↓

↓↓

general thyroid inflammation (thyroiditis)occurs following radiation, childbirth or viralillness; there may be an underlying auto-immune aetiology to this. It usually remitswithout recurrence. Thyroid cancer is one ofthe most common radiation-induced tumours,via ingestion of radioiodine (131I), e.g. afterradiological accidents such as at Chernobyl.

Amiodarone, which has a high iodine con-tent, frequently causes mild hyperthyroidism,possibly leading to thyrotoxicosis on prolongedtherapy. It can also cause hypothyroidism (Table9.24). Very rarely hyperthyroidism can besecondary to pituitary hyperactivity (Table 9.27).

Pathology

High levels of thyroid hormone cause a generalacceleration of metabolic processes withincreased metabolic rate and energy utilization,hyperthermia and increased cardiovascularactivity (see below). There is a compensatory fallin TSH, often to undetectable levels.

Investigation and diagnosis

Owing to the several possible aetiologies, moreextensive investigation is required than forhypothyroidism. Typical clinical features willinvariably be borne out by a TFT, which will

usually show raised FT4 and FT3 and barelydetectable TSH.

Further investigation will depend upon thedegree of suspicion of different aetiologies, butcould include:

• Autoantibody scan; thyroid peroxidase andthyroglobulin antibodies are usually found,but there is a 10–20% false-negative ratebecause they may also occur in unaffectedindividuals. TSH-stimulating receptor anti-bodies are difficult to assay and are notroutinely sought.

• Imaging is best done with radiolabelledsodium pertechnetate (99mTc), which is prefer-entially taken up into the thyroid by thesymporter but not organified. This will showthe overall size of the organ, with concentra-tion in any nodules, showing their numberand size. It is a prerequisite if ablation therapyis planned. Ultrasound is less invasive. MRI orCT scanning is used if ophthalmopathy (seebelow) is suspected.

• Biopsy: if a tumour is suspected.

Clinical features

The clinical features of hyperthyroidism (Table9.28) should be contrasted with those ofhypothyroidism (Table 9.25): the picture isstrikingly different. The range of features variesslightly according to aetiology but is broadly

Hyperthyroidism 639

Table 9.27 Causes of hyperthyroidism

Aetiology Examples

Most common (75%)Autoimmune stimulation Graves’ disease (stimulatory anti-TSH receptor antibody)

Less commonMultinodal goitreAdenomaThyroiditis Post partum, viral, autoimmuneIatrogenic Amiodarone, excessive levothyroxine doseDietary Excess iodine

RareSecondary Pituitary – excessive TSH secretion

Other endocrine abnormalities

consistent. Typically the patient is thin,nervous, agitated, hyperactive, hot, thirsty andsweaty.

Examination will show a raised heart rate,possibly even atrial fibrillation; in severe casesthere may be signs of heart failure. The neck willusually be swollen and auscultation of the goitrewill reveal bruits (the sound of rapid, excessiveblood flow). There are also usually diarrhoea andanxiety.

In Graves’ disease the common complicationof ophthalmopathy (see below) will causebulging eyes and an unblinking stare, known asexophthalmos – the classic sign of thyrotoxi-cosis. Another characteristic Graves’ feature is

pretibial myxoedema, where the deposition offibrous material causes painless dermal noduleson the shin.

Course

Graves’ disease may follow a relapsing and remit-ting course, with remissions facilitated bytherapy. However, remissions become decreas-ingly likely following each successive relapse.Paradoxically, the end-stage for some patientsmay be autoimmume hypothyroidism. There isan increased risk of osteoporosis and heartdisease in untreated disease.


Table 9.28 Principal clinical features of hyperthyroidism

Common feature Less common

Systemic/metabolic Fatigue, weaknessHyperactivity, restlessnessWeight lossHeat intolerance, sweating Hyperthermia, warm peripheryPolydipsia PolyuriaGoitre Bruit over gland (excess perfusion)

Cardiovascular Tachycardia, atrial fibrillation Heart failure (high output)Palpitations Hypertension

Musculoskeletal Hyper-reflexia MyopathyTremor Lid retraction, lid lag

Neuropsychiatric Irritability, anxiety, dysphoria DepressionInsomnia Psychosis

Gastrointestinal Increased appetite Diarrhoea

Reproductive OligomenorrhoeaLoss of libido

Dermatological PruritisPretibial myxoedema(a)

Ophthalmopathy(a) Grittiness DiplopiaPeriorbital, conjunctival oedema Impaired acuityScleral injection (‘red eye’)Proptosis (exophthalmos)

Family history Autoimmune disease e.g. Graves’ disease, pernicious anaemia, vitiligo, type 1 diabetes mellitus

(a) Only in Graves’ disease.

Complications

Ophthalmopathy (thyroid eye disease)

A characteristic eye disease affects about half ofGraves’ disease patients. It is potentially seriousand for unkown reasons it is associated withsmoking. The cause is autoimmune inflamma-tion of the oculomotor muscles, with fibrousovergrowth. This pushes the eyes forward andimpairs eye movement. The overexposed corneascan become dry and painful, and there may bediplopia (double vision). In the most severe form(�10% cases) the retro-orbital swelling cancompress the optic nerve and threaten sight.

It can be detected by examination of eyemovement and testing for double vision at theextremes of lateral eye rotation, but MRI scan-ning is needed for precise assessment. Its severityis not related to thyroid hormone levels nor is itrelieved if euthyroidism is achieved by medicalor surgical means, probably because it is due toantithyroid antibodies rather than excessivethyroid hormone itself. For most patients it is anunsightly inconvenience rather than a threat tosight.

Thyroid crisis (‘storm’)

This rare condition, which occurs when there arevery high levels of thyroid hormone, is poten-tially fatal. There is excessive cardiovascularstimulation, high fever and extreme agitation. Itcan be triggered in hyperthyroid patients byextra metabolic stress, such as infection, bymental stress, or by radioiodine therapy.

Autoimmune disorder

Other autoimune diseases including perniciousanaemia, myasthenia gravis, type 1 diabetes andvitiligo are more common among Graves’disease sufferers.

Management

The aims of management are symptom controland reduction of thyroid hormone output. Forthe latter, three modes are available:

• Pharmacological suppression.• Radio-isotopic thyroid gland reduction/

ablation.• Surgical thyroid gland reduction/ablation.

Beta-blockers are used for symptom controlwhile other therapy is initialized. This is effectivebecause many of the effects of thyroid hormoneare sympathomimetic and resemble those ofadrenaline (epinephrine), including cardiacstimulation, tremor and anxiety (Table 9.23).Propranolol is preferred, probably because it isnon-selective and crosses the blood–brainbarrier, helping the anxiety. Agents withintrinsic sympathomimetic activity (e.g. pindolol;see Chapter 4) should not be used.

Patients may use different modes at differentstages in their illness. Typical paths are shown inFigure 9.16. After initial stabilization withantithyroid drugs patients may go into remissionafter a year or so and drugs may be withdrawn.However, relapse is common and remission isthen less likely. Thyroid gland reduction aims ata graded reduction in thyroid mass, hoping toleave enough remaining to produce normalamounts of thyroid hormone. However, judgingthis is difficult and it is always preferable to erron the side of greater destruction, obviating theneed for further invasive therapy at a later date.Consequently, eventual iatrogenic hypothy-roidism is common. Alternatively, a full ablationmay be decided on at the outset, removing doubtand easing management by starting the patienton thyroxine replacement immediately. Thusthe choice of options depends on the cause,severity, patient age and patient preference.

Pharmacotherapy

Antithyroid drugs are usually first-line treat-ment. They block thyroid peroxidase rapidly, butsymptom control takes 2–4 weeks owing tostores of thyroid hormone and its long half-life.The most common agents are the thionamides.Carbimazole is preferred in the UK but propylth-iouracil is used in the USA. The latter also blocksT4–T3 conversion but this may not be clinicallysignificant. Most antithyroid drugs also haveimmunosuppressant activity, reducing TSH-receptor antibodies, which may account for the

Hyperthyroidism 641

sustained remission seen in about half ofpatients after withdrawal of drug therapy It mayalso be related to the most serious side-effect,agranulocytosis.

A high initial dose (e.g. 40–60 mg carbimazole,depending on initial TFTs) is tapered after4–6 weeks, with advice to the patient to be alertfor overtreatment (sluggishness, constipation,slow pulse, etc). Repeated T3/T4 level estimationsguide dose reduction at 4- to 8-week intervals.TSH takes longer to rise than thyroid hormonelevels do to fall. A maintenance dose of 5–10 mgdaily is continued for 18 months, after which atrial withdrawal can be attempted. About half ofpatients remain in remission and are monitoredannually. Some eventually relapse; othersdevelop autoimmune hypothyroidism. Thosewho relapse have less chance of a further remis-sion and either long-term pharmacotherapy oran alternative mode of therapy is then indicated.

Block and replaceAn alternative strategy is to continue antithyroiddrugs at a high dose for the same period of2 years, effectively producing a chemical abla-tion. Standard replacement doses of levothyroxineare given, eventually withdrawing all drugs ifeuthyroidism is achieved. This strategy issimpler, requiring less monitoring and titration,and it allows for a more sustained immunosup-pressant action from higher doses of antithyroiddrug. However, there is little evidence that it ismore effective, and there is an increased risk ofside-effects.

Side-effects include minor dermatologicalproblems, avoided by changing to anotherantithyroid agent, and other minor non-specificdrug side-effects. Most important, however, isbone marrow suppression and agranulocytosis,

which can affect 0.1% of patients. This is usuallyrapid in onset, occurring during the first3 months of treatment, so not easily detectedfrom blood counts. All patients must be warnedto watch for swollen glands, throat infectionsand bruising. If these occur they should stoptheir drug and consult their GP urgently. Theproblem is reversed on withdrawal but antimi-crobial cover (for neutropenia) and filgastrim (tostimulate leucocyte recovery) may initially berequired. A change of drug may subsequently betried: the effect may not recur.

Iodide/iodine have an antithyroid effect andare sometimes used as an adjuvant in thyroidstorm or before thryoidectomy, to reduce glandsize, but they are no longer first-line therapy.

Radiotherapy

Selective thyroid reduction using sodiumradioiodide (131I) exploits the concentration ofiodide in the thyroid, which minimizes exposureof other organs and allows a low total body dose.In the USA it it often the first line treatment forthose over 50, owing to the potential cardiovas-cular risks of hyperthyroidism. In Europe it ispreferred to surgery for medical failure to controlhyperthyroidism or following relapse. Althoughthe aim is to spare enough gland to permitnormal thyroid hormone output, there is a10–20% chance of hypothyroidism in the firstyear following treatment and subsequently up toa 5% annual incidence. Sodium radioiodide istaken as an oral solution. Little special contactavoidance is necessary afterwards, except foravoiding public transport and sustained closecontact with children for about 4 weeks. It iscontra-indicated in pregnancy.


Hyperthyroidism Antithyroiddrugs

remission monitor

Euthyroid Euthyroid HypothyroidismStop therapy Relapse Surgery orradiotherapy

Antithyroiddrugs

Levothyroxine

Figure 9.16 Treatment pathways for hyperthyroidism.

ComplicationsThe effect takes several months to develop,during which thyroid hormone levels may risetemporarily, and antithyroid drug or beta-blocker cover may be needed. Ophthalmopathyis a relative contraindication because it may beexacerbated. There is a small increase in the riskot thyroid cancer.

Thyroidectomy

Surgery has a similar aim to radiotherapy, i.e.subtotal thyroid gland reduction, but has thesame imprecision and is more invasive. It isparticularly indicated if there is a large goitre. Itis important that patients are rendered euthyroidbefore surgery, to avoid thyroid storm. Some aregiven oral iodine (Lugol’s iodine) or potassiumiodide for a few weeks before surgery, to inhibitthyroid hormone synthesis and reduce glandvascularity. As with radioiodine, many patientseventually become hypothyroid. Potentialsurgical complications include laryngeal orparathyroid damage.

Ophthalmopathy

Milder cases need symptomatic treatment,including artificial tears and eye protection. Ifsight is threatened, high-dose corticosteroids,surgery or radiation therapy may be indcated.

Thyroid storm

Urgent antithyroid therapy with thionamidesand iodine are required to reduce thyroidhormone output. Symptomatic cover with beta-blockers, corticosteroids and possibly IV fluidswill usually be necessary. The precipitating causemust be discovered and treated.

References and further reading

Alberti K G M M, Defronzo R A, Zimmet P, eds (1997).International Textbook of Diabetes Mellitus, 2nd edn.Chichester: John Wiley.

Bloomgarten Z T (2004). Diabetes complications.Diabetes Care 27: 1504–1512.

Davies M, Srinivasan B (2005). Glycaemic managementof type 2 diabetes. Medicine 34(2): 69–75.

Devendra D, Liu E, Eisenbarth G S (2004). Type 1diabetes: recent developments. BMJ 328: 750–754.

Diabetes Control and Complications Trial ResearchGroup (1993). The effect of intensive treatment ofdiabetes on the development and progression oflong-term complications in insulin dependentdiabetes mellitus. N Engl J Med 329: 977–986.

Dinneen S F (2006). Management of type 1 diabetes.Medicine 34(2): 63–7.

Franklyn J A (2005). Hypothyroidism. Medicine 33(11):27–29.

Marshall S M, Flyvbjerg A. Prevention and detection ofvascular complications of diabetes. BMJ 333:475–480.

Nathan D M (1998). Some answers, more controversy,from UKPDS. Lancet 352: 832–833.

NICE (September 2002). Type 2 diabetes – bloodglucose: Management of type 2 diabetes –Managing blood glucose levels (Clinical Guideline).Available from http://guidance.nice.org.uk/CGG/?c�91523 (accessed 16 August 2007).

NICE (July 2004). Type 1 diabetes: Diagnosis andmanagement of type 1 diabetes in children, youngpeople and adults (CG15). Available from http://guidance.nice.org.uk/CG15/?c�91523 (accessed 16August 2007).

Nutrition Subcommittee of Diabetes UK (2003). Theimplementation of nutritional advice for peoplewith diabetes. Diabetic Med 20: 786–807.

Phillips P (2002). Insulins in 2002. Aust Prescr 25:29–31.

Shepphard C S (2005). Goitre and thyroid cancer.Medicine 33(11): 35–37.

Stumvoll M, Goldstein B J, van Haeften T W (2005).Type 2 diabetes: principles of pathogenesis andtherapy. Lancet 365: 1333–1346.

Watkins P J (2003). ABC of Diabetes, 5th edn. London:BMJ Publishing Group.

Weetman A P (2005). Thyrotoxicosis. Medicine 33(11):30–34.

Internet resources

http://www.diabetes.org.uk (website of Diabetes UK,the charity for people with diabetes).

References and further reading 643

Health & Medicine

Endocrine system