WHAT IS TYPE 2 DIABETES MELLITUS?

Published on 03/04/2015 by admin

Filed under Endocrinology, Diabetes and Metabolism

Last modified 22/04/2025

Print this page

This article have been viewed 3978 times

CHAPTER 1 WHAT IS TYPE 2 DIABETES MELLITUS?

OVERVIEW

DEFINITION OF DIABETES MELLITUS

Diabetes mellitus is a group of metabolic diseases characterised and diagnosed by a chronic elevation of blood glucose (hyperglycaemia) that results from defects in insulin secretion, insulin action or both. This may be accompanied by various disturbances of carbohydrate, protein and fat metabolism. The effect of different clinical manifestations may depend upon the underlying cause(s) of the diabetes, the degree of deficit of insulin action, coexisting conditions and the extent of diabetic tissue damage. Much of the morbidity and mortality in diabetics results from vascular damage.

CLASSIFICATION OF DIABETES MELLITUS

The World Health Organization’s classification of diabetes (WHO 1985) has been adopted internationally. The American Diabetes Association re-examined the diagnostic criteria and classification and recommended modifications in 1997, subsequently agreed by WHO (The Expert Committee 1997, Alberti et al 1998). The terms type 1 and type 2 diabetes (Table 1.1) replaced the old categories of insulin-dependent diabetes mellitus (IDDM) and non-insulin-dependent diabetes mellitus (NIDDM). The older classification was based upon treatment (many NIDDM patients are on insulin), but did not indicate the nature of the underlying cause (Wroe 1997).

TABLE 1.1 1997–1998 ADA and WHO classification of diabetes mellitus

Type 1

(beta-cell destruction, usually leading to absolute insulin deficiency)

Autoimmune

Idiopathic

Type 2

(may range from predominately insulin resistance with relative insulin deficiency to a predominately secretory defect with or without insulin resistance)

Other specific types

Genetic defects of beta-cell function

Genetic defects in insulin action

Diseases of the exocrine pancreas

Endocrinopathies

Drug- or chemical-induced

Infections

Uncommon forms of immune-mediated diabetes

Other genetic syndromes sometimes associated with diabetes

Gestational diabetes mellitus (includes gestational impaired glucose tolerance)

Types 1 and 2 diabetes are compared in Table 1.2.

TABLE 1.2 Comparison of the characteristics of types 1 and 2 diabetes mellitus (Beers 1999)

Characteristic	Type 1 diabetes	Type 2 diabetes
Commonest age at onset	Usually <30 years	Most often >30 years, but note recent trends
Associated obesity	No	Yes
Propensity to develop ketoacidosis (requiring insulin to prevent/control)	Yes	No
Presence of classic symptoms of hyperglycaemia at diagnosis	Yes, often severe	May be absent
Presence of classic symptoms of hyperglycaemia at diagnosis	Yes, often severe	If present, often moderate
Endogenous insulin secretion	Very low to undetectable	Variable, but low relative to plasma glucose levels
Insulin resistance	Not present	Yes, but variable
Twin concurrence	<50%	>90%
Associated with specific HLA-D antigens	Yes	No
Islet cell antibodies at diagnosis	Yes	No
Islet pathology	Insulitis, selective loss of most beta cells	Smaller, normal-looking islets
Islet pathology	Insulitis, selective loss of most beta cells	Amyloid deposits common
Associated increased risks for micro-and macrovascular disease	Yes	Yes
Hyperglycaemia responds to oral agents	No	Yes, initially in most patients

CRITERIA AND METHODS FOR THE DIAGNOSIS OF DIABETES MELLITUS

NSF

CONCEPTS INVOLVED IN MAKING A DIAGNOSIS

Establishing a diagnosis of diabetes mellitus “should” be straightforward: the blood glucose level is either above or below the diagnostic threshold. However, consensus must be reached as to what level should be set as the diagnostic threshold (criteria) and which diagnostic test(s) to use (methods). Making the correct diagnosis reliably in an asymptomatic patient (half of type 2 patients at diagnosis) may pose a challenge.

CURRENT CRITERIA AND METHODS

The current recommendations are based on a 1998 WHO consultative document (Alberti et al 1998). The following criteria are for diagnosis only, and are not criteria for initiating treatment or therapeutic goals:

• Fasting (no caloric intake for at least 8 hours) plasma glucose level of 7.0 mmol/l (126 mg/dl) or greater

• Random plasma glucose level of 11.1 mmol/l (200 mg/dl) or greater

• 2 hour plasma glucose level (post 75 g glucose load) of 11.1 mmol/l or greater, from an oral glucose tolerance test (OGTT).

A single reading above the diagnostic threshold is sufficient to make the diagnosis if the patient has classic symptoms (e.g. thirst, polyuria, lethargy, blurred vision and weight loss); otherwise, if the patient is asymptomatic, diagnosis requires two abnormal results on separate days.

Although the OGTT is the “gold standard”, it may be impractical as a first-line investigation in general practice, being undertaken when the fasting or random plasma glucose results fail to resolve diagnostic uncertainty, especially in patients whose fasting levels fall into the “impaired fasting glucose” range (see below), and who are thought to be at high risk of developing either diabetes or vascular disease.

Glycosuria, abnormal finger-prick blood glucose or elevated glycated haemoglobin suggest, but do not satisfy, the diagnostic criteria for diabetes.

RATIONALE FOR DIAGNOSTIC CRITERIA AND METHODS

The distribution of plasma glucose concentrations is a continuum; so there needs to be a threshold that separates those who are at a substantially increased risk of developing adverse outcomes caused by diabetes from those who are not (The Expert Committee 2003b). The medical, social and economic costs of making a diagnosis in those not at increased risk must be balanced against the costs of failing to diagnose those at increased risk.

Historically, this threshold was determined by the relationship between blood glucose levels and microvascular complications in type 1 diabetics: however, it may be more important that this threshold is determined by the relationship between blood glucose levels and macrovascular risk in a wider population (i.e. those with either impaired glucose tolerance or type 2 diabetes).

The WHO and other bodies have adopted the ADA’s diagnostic criteria (The Expert Committee 1997), but there have been different recommended optimal methods of diagnosis. The WHO prefers the OGTT, supported by evidence that 2 hour post-load plasma glucose levels were more accurate than fasting plasma glucose levels in identifying those at increased risk of death associated with hyperglycaemia (DECODE 1999). The drawback of ADA’s preference for fasting plasma glucose levels is that “normal” results carry the risk of missing some diabetics, especially among the elderly and in some ethnic groups: the earliest defect in the natural history of beta cell dysfunction is the reduction of first-phase insulin release, associated with 2 hour post-load hyperglycaemia.

Although diabetes is “arbitrarily” and solely diagnosed on the basis of blood glucose levels, it should be regarded as a syndrome that includes other metabolic and haemodynamic features.

ORAL GLUCOSE TOLERANCE TEST (OGTT)

Standard protocol

After 3 or more days with a daily carbohydrate intake of at least 150 g, the OGTT should be performed in the morning after an overnight fast of 8–14 hours (during which plain water may be drunk). A venous blood sample is taken then a drink containing the equivalent of 75 g of glucose (e.g. Lucozade 388 ml) is consumed within 5 minutes. The subject should be seated, not smoke and take no unusual exercise during the test period. A second venous blood sample is taken exactly 2 hours after the start of the glucose drink. Both samples should be sent to an accredited laboratory for estimation of plasma glucose. Interpretation of the results of the OGTT test is provided in Table 1.3.

TABLE 1.3 Interpretation of the oral glucose tolerance test (The Expert Committee 2003a, b)

Based on plasma venous glucose	Fasting (no caloric intake for 8 hours)	2 hours post 75 g glucose load
Diabetes mellitus	7.0 mmol/l or greater	11.1 mmol/l or greater
Impaired glucose tolerance	–	7.8 to 11.0 mmol/l
Impaired fasting glucose	6.1 to 6.9 mmol/l	–
Normal glucose homeostasis	6.0 mmol/l or less	7.7 mmol/l or less

DISEASE PROCESSES OF TYPE 2 DIABETES MELLITUS

DEFECTS RESPONSIBLE FOR TYPE 2 DIABETES

The chronic hyperglycaemia of type 2 diabetes results from diverse and progressive disease processes that cause:

• defects in the ability of pancreatic beta (β) cells to secrete insulin (β-cell dysfunction)

• defects in the ability of insulin to inhibit hepatic glucose production and to promote glucose utilisation (insulin resistance).

In type 2 diabetes, both of these defects coexist and both can be caused by a plethora of genetic or environmental factors. Most commonly, type 2 diabetes appears to be inherited as a polygenic trait, with environmental factors also involved, often at a very young age.

Secretion of insulin, in response to rising blood glucose levels, occurs in two phases in healthy individuals:

1. The first (early acute) phase consists of a rapid rise in insulin immediately after an increase in blood glucose levels. This insulin is thought to originate from a small pre-formed reserve pool of insulin stored within the beta cells.

2. The second (late) phase response consists of a more gradual rise in insulin levels. The on-going biosynthesis of newly formed insulin in this phase occurs during the time that blood glucose levels are elevated.

Early in type 2 diabetes, β-cells begin to lose their initial response of increased insulin secretion (“first phase”). Sustained hyperglycaemia reduces β-cell function by “glucose toxicity”. With progression of the disease, the loss of first-phase secretion leads to early post-prandial hyperglycaemia, exaggerated late (“second-phase”) insulin secretion and late post-meal hypoglycaemia. Insulin secretory pulses become abnormal under basal conditions. The loss of first-phase insulin secretion, which leads to post-prandial glucose “spikes”, is associated with an increased risk of cardiovascular disease.

In insulin resistance, insulin is unable to produce its usual effects at concentrations that are effective in normal individuals. Its onset precedes the development of type 2 diabetes and may arise from a variety of genetic mutations. It is thought that the reduced action of insulin is linked closely with the cardiovascular risk factors, such as obesity, that are part of the insulin resistance syndrome (Reaven 1988).

Malnutrition in utero and during early infancy may be associated with an increased risk of developing type 2 diabetes later in life (the “thrifty phenotype” hypothesis) by affecting both β-cell function and insulin resistance. Regular physical exercise, when undertaken consistently from childhood, can protect against type 2 diabetes by improving insulin sensitivity.

BIOCHEMISTRY OF DIABETES COMPLICATIONS

Vascular tissues are freely permeable to glucose. Poor glycaemic control renders such tissues more vulnerable to insult, which starts insidiously and may eventually lead on to the failure of major systems. Abnormalities may occur in endothelial cells, vascular smooth muscle cells, glomeruli and mesangial cells, and cardiomyocytes. The clinical consequences of diabetic microvascular disease include visual impairment, chronic renal failure and neuropathic foot ulceration.

Four main mechanisms (not mutually exclusive) have been implicated in the pathogenesis of glucose-mediated vascular damage:

1. Non-enzymatic glycation (basis for the glycated haemoglobin assay)

2. Oxidative/reductive stress

3. Aldose reductase activation

4. Activation of diacylglycerol-protein kinase C signalling.

However, clinical complications are driven by raised blood pressure, dyslipidaemia and multiple other abnormalities, particularly those that contribute to insulin resistance. Some of these precede the onset of type 2 diabetes, by which time atheromatous changes have already been established. This is why prevention or early detection is important.

The precise pathogenesis of atheromatous change in patients with diabetes may vary not just between individuals, but also between sites and different calibre of arteries in the same individual. A better understanding of these mechanisms may help to identify potential therapeutic interventions, although lifestyle modification and vigorous correction of raised blood pressure and dyslipidaemia must remain central to reducing cardiovascular risk.

INTERMEDIATE HYPERGLYCAEMIC CONDITIONS

The term “pre-diabetes” has been used to categorise people with impaired glucose metabolism, but who are not diabetic. Many, but not all, progress to diabetes. The WHO term “intermediate hyperglycaemia” may be more accurate. Whichever term is used, identification followed by appropriate interventions (particularly aimed at optimising lifestyle) can achieve real benefit for this group.

IMPAIRED GLUCOSE TOLERANCE AND IMPAIRED FASTING GLUCOSE

These terms are not interchangeable and do not define identical groups of individuals. The rationale for establishing these intermediate categories of impaired glucose regulation is based on their value in predicting cardiovascular risk (IGT) and future diabetes mellitus (IFG).

Impaired glucose tolerance (IGT) refers to a glucose metabolic state that is intermediate between normal glucose homeostasis and diabetes mellitus. IGT only applies to a plasma glucose level in the range of 7.8 to 11.0 mmol/l at 2 hours after a 75 g glucose load. Patients can be labelled as having IGT only from an OGTT. Individuals with IGT are at increased risk of developing macrovascular disease. IGT progresses to type 2 diabetes in 37% at 5 years (Gillies 2007) and 50% at 10 years (Davies 2006). It is logical to regard IGT as a risk factor rather than as a disease entity, particularly as many individuals with IGT are asymptomatic and have normal plasma glucose levels in their daily lives. Some evidence suggests that the most cost-effective interventions to prevent or delay the onset of diabetes should target individuals with IGT, followed by high-risk groups.

Impaired fasting glucose (IFG) applies only to a fasting plasma glucose level in the range between a lower limit of 6.1 mmol/l (recommended by JBS2, WHO and the IDF) and an upper limit of 6.9 mmol/l. Some patients, particularly elderly or Indo-Asians, can have IFG, but fulfil the diagnostic criteria for diabetes because their 2 hour post 75 g load plasma glucose level is 11.1 mmol/l or greater. Thus, patients with IFG can have either diabetes or IGT or normal glucose homeostasis (based upon the 2 hour result). This has major implications for screening, because a fasting plasma glucose level below the diagnostic threshold for diabetes does not exclude current diabetes.

If IFG is defined as between 6.1 and 6.9 mmol/l, then it includes a much lower proportion of the population than is categorised as having IGT. One review found that of those who had IFG and/or IGT, 16% had both, 23% had IFG alone, and 60% had IGT alone, with significant age and gender differences between the glucose intolerance categories (Unwin et al 2002). IFG and IGT may be different metabolic states. Although the ADA has recommended a reduced lower limit for IFG to 5.6 mmol/l is, on balance, a better predictor for cardiovascular and metabolic outcomes (The Expert Committee 2003a), other guidance (WHO/IDF, JBS2, NICE/NSF still recommends that the lower limit should remain at 6.1 mmol/l (WHO/IDF 2006).

There has been considerable research into the factors that increase the likelihood of an individual with intermediate hyperglycaemia developing diabetes. It is also interesting to look earlier in the natural history at factors that may cause glucose intolerance. Smoking is thought to increase insulin resistance, but the evidence is inconclusive as to whether smoking is an independent risk factor for the development of diabetes. The CARDIA study found that, in young individuals with normal glucose tolerance, both active and passive smoking were associated (more so in whites) with the development of glucose intolerance (Houston et al 2006).

METABOLIC SYNDROME

Insulin resistance is associated with a collection of abnormal risk factors (obesity, impaired glucose tolerance, hypertension and dyslipidaemia) and is now recognised as a major underlying contributor to increased coronary heart disease (CHD) mortality. Metabolic syndrome is a defined cluster of abnormal cardiovascular risk factors; it doubles cardiovascular disease (CVD) mortality, trebles the onset of CVD events (Carr et al 2004), and predicts the development of not only type 2 diabetes, but also obstructive/sleep airways disease, gall stones, some cancers and chronic kidney disease.

The two major previous definitions of metabolic syndrome were from:

• the WHO (including evidence of insulin resistance and measurement of fasting insulin)

• the National Cholesterol Education Program (presence of three from the following clinical criteria: BP > 130/85, HDL-C < 1.04 mmol/l in men or < 1.29 mmol/l in women, TG > 1.69 mmol/l, fasting glucose > 5.6 mmol/l or abdominal obesity).

In 2005, the International Diabetes Foundation (IDF) proposed a new definition for metabolic syndrome for clinical use (set out in Table 1.4) that avoids the technical difficulties associated with insulin measurement (International Diabetes Federation 2005).

TABLE 1.4 International Diabetes Federation (IDF) definition of metabolic syndrome (IDF 2005)

Parameter	Qualifications
*In all cases:*
Central obesity (waist circumference)	In men: ≥ 94 cm in Europids, Sub-Saharan Africans and Arabs ≥ 90 cm in South Asians and Chinese ≥85 cm in Japanese

In women:

≥ 80 cm in all populations, except

≥90 cm in Japanese

Any 2 of the following 4: Raised triglyceride level

≥1.7 mmol/l

specific treatment for this lipid abnormality

Reduced HDL-C levels

≤ 0.9 mmol/l in males or ≥1.1 mmol/lin females

specific treatment for this lipid abnormality

Raised blood pressure

Systolic ≥130 mm Hg

or diastolic ≥80 mm Hg

treatment of previously diagnosed hypertension

Raised fasting plasma glucose

≥5.6 mmol/l

(OGTT recommended to confirm diabetes)

previously diagnosed type 2 diabetes

OGTT: oral glucose tolerance test

Epidemiological data about the metabolic syndrome are not entirely reliable due to the use of different definitions, but the prevalence varies amongst different ethnic groups, ranging from 22 to 39% of the adult population. It is commoner in an Indo-Asian population: a BMI greater than 23 kg/m² in Indo-Asians (as compared to 25 kg/m² in white Caucasians), is now thought to indicate increased CVD risk in this population, subject to the occasional confusion of obesity with being overweight (Hanif et al 2002).

Although the presence of metabolic syndrome does “predict” CVD events, there is no evidence that its criteria enhance the estimation of CVD risk in either diabetic or non-diabetic populations beyond existing tools, such as the New Zealand tables. However, the criteria associated with metabolic syndrome, particularly increased waist circumference, are better predictors of the development of type 2 diabetes (Sattar 2006).

Although no data are yet available from any large intervention trials in primary care for preventing or delaying diabetes or CVD in people with metabolic syndrome, there are good reasons for managing metabolic syndrome aggressively, addressing all relevant cardiovascular risk factors:

• Metabolic syndrome is clearly associated with a much greater risk of developing either CHD or type 2 diabetes or both.

• Its constituent adverse risk factors are also associated with increased cardiovascular mortality and morbidity, the result being often greater than additive.

• Metabolic syndrome itself is a significant predictor of total and cardiovascular mortality (Sundström et al 2006).

• If these constituent factors are properly addressed, then mortality and morbidity are delayed or reduced significantly.

• Many of the interventions involve improving lifestyle, where successful change is known to be beneficial for more than CVD prevention.

PREVALENCE OF TYPE 2 DIABETES

CURRENT PREVALENCE

Estimating the prevalence of type 2 diabetes mellitus risks underestimation, because many type 2 patients are asymptomatic, and these individuals and/or health-care professionals may not always recognise the symptoms of hyperglycaemia. Some of the attempts to estimate the prevalence of diabetes are listed in Table 1.5.

TABLE 1.5 Estimates of the prevalence of diabetes mellitus in the UK

The National Diabetes Audit (NDA) has used the estimates for England produced by the PBS Diabetes Population Prevalence Model for 2001 and predicted for 2010, with adjustments for age, gender, ethnicity, social deprivation and district. The PBS model’s overall estimate for the prevalence of diabetes in 2001 of 4.37% was higher than other estimates (YHPHO 2005). In 2005 the estimated prevalence of all diabetes (diagnosed and undiagnosed) in England was 4.67% or 2 350 000 persons (Kingdom & Ferguson 2006).

One of the quality indicators for the GMS contract is that each practice should keep a diabetic register: indicator DM1 in 2004–2006; now DM19 (see Appendix 3). Data from the monitoring process provided a “raw” national prevalence, based on the number of diabetics identified in every practice, of 3.544% in 2005, lower than the PBS model, indicating that there is still a substantial number of “missing” diabetics (about one-quarter of those with diabetes).

FUTURE PREVALENCE

It is predicted that the prevalence of type 2 diabetes will continue to increase rapidly. It has been estimated that the worldwide prevalence of type 2 diabetes will more than double from 98.9 million in 1994 to 215 million by 2010 (Amos et al 1997).

FACTORS THAT AFFECT THE PREVALENCE OF TYPE 2 DIABETES

Deprivation

Deprivation is associated with a higher prevalence in both sexes aged 35 to 74 years (Newnham et al 2002), although it is unclear whether this is independent of the above factors. The PBS model estimated the prevalence of diabetes to be 35% higher than the mean in the most socio-economically deprived fifth of the English population.

CONSEQUENCES OF TYPE 2 DIABETES MELLITUS

DIABETIC CONTROL, MORTALITY AND COMPLICATIONS

The diagnosis of type 2 diabetes mellitus has a profoundly adverse effect on morbidity and mortality. The long-term complications are mostly commonly vascular in origin and can be divided into macrovascular (ischaemic heart disease, cerebrovascular disease, and peripheral arterial disease) or microvascular (retinopathy, nephropathy and neuropathy).

In its 2004 report, Diabetes UK listed some truly frightening statistics (Diabetes in the UK 2004):

• Reduced life expectancy by up to 10 years in type 2 diabetics.

• Evidence of complications at diagnosis in 50% of type 2 diabetics.

• In the UK diabetics spend 1.1 million days in hospital every year.

• Cardiovascular disease will be the cause of death in 80% of diabetics.

• Diabetics are two to three times more likely to have a stroke.

• 1000 diabetics start kidney dialysis every year in the UK.

• Diabetes is the leading cause of blindness in people of working age in the UK.

• The rate of lower limb amputation in diabetics is 15 times higher than in people without diabetes.

Overall, 5-year mortality in type 2 diabetics increases two- to threefold and age-adjusted life expectancy is reduced by 5 to 10 years compared to the general population (Panzram 1987), with cardiovascular disease being the predominant cause of mortality.

The adverse effects of sustained hyperglycaemia and poor management on mortality and morbidity in type 2 diabetes have been studied extensively. The UKPDS has now identified poor glycaemic control among the risk factors for coronary artery disease in type 2 diabetics (Turner et al 1998). The UKPDS also demonstrated that improving glycaemic control significantly reduced the risk of microvascular complications (e.g. retinopathy, nephropathy), with lesser reductions in macrovascular disease (e.g. CHD, major stroke) and no effect on diabetes-related mortality in type 2 diabetics (UKPDS 1998a, Stratton et al 2000).

As many type 2 patients at diagnosis have complications, particularly vascular, and/or adverse cardiovascular risk factors, early diagnosis (including targeted screening) and high-quality management can improve their future well-being.

FINANCIAL CONSEQUENCES OF TYPE 2 DIABETES

Several attempts have been made to ascertain precisely how much is spent both directly and indirectly on diabetes care.

In 2002, the Diabetes NSF estimated that the NHS spent £3.5 billion per year on treating diabetes and its complications (Department of Health 2002). Since then the actual costs will have increased, due to economic and health-care inflation and to the rising prevalence of type 2 diabetes. T2ARDIS estimated that diabetes was responsible for 5% of the total NHS expenditure (British Diabetic Association 2000). The prevalence of type 2 diabetes, with its associated morbidity and mortality, continues to increase, as do its costs and the pressures on those responsible for financing health care. Diabetes UK predicted that the proportion of NHS expenditure on diabetes will rise to 10% by 2011 (Diabetes in the UK 2004).

The UKPDS calculated the annual manpower cost of implementing its findings for glycaemic and blood pressure control in patients with type 2 diabetes to be £264 per patient in 1998 (UKPDS 1998a, b).

The Type 2 Diabetes: Accounting for a major Resource Demand in Society in the UK (T2ARDIS) project looked at the financial impact of type 2 diabetes and presented its results to Diabetes UK’s Annual Professional Conference in March 2000 (British Diabetic Association 2000). T2ARDIS calculated that the average annual cost of treating each person with type 2 diabetes incurred direct costs to the country of £2152, comprising:

• £1738 (81%) to the NHS, the total annual cost being £2 billion, 4.7% of the total NHS expenditure, with hospitalisation consuming more than 40% of diabetes-related NHS expenditure.

• £285 (13%) borne by the private individual.

• £129 (6%) to the social services. Over 75% of social services costs for people with type 2 diabetes are associated with residential and nursing care.

T2ARDIS also showed that the presence of both microvascular and macrovascular complications in an individual increased the average NHS costs in the calculation above fivefold, personal expenditure threefold, and social services costs fourfold.

Published in 2002, the CODE-2 study gathered data on 7000 type 2 diabetics from eight European studies: it estimated the average annual cost per patient to be £1934, of which 55% could be attributed to hospital admissions and 7% to the cost of blood glucose-lowering medication (Jonsson 2002); these estimates are not too dissimilar to T2ARDIS.

At Diabetes UK’s Annual Professional Conference in April 2005, Professor Rhys Williams presented predictions that, by 2025, diabetes care worldwide could cost between US$153 and 286 billions, accounting for 7.5 to 13.5% of total health-care costs: a terrifying prospect.

Wanless estimated that the additional annual cost of implementing fully the Diabetes NSF in England by 2010–2011 would be £600 million; however, an annual saving of £200 million would result from measures that reduce the medical effects of diabetes (by prevention, earlier detection or improved care to reduce complications).

SCREENING FOR TYPE 2 DIABETES MELLITUS

GENERAL POINTS AND DEFINITIONS

NSF

Diagnostic testing and screening are two distinct processes. In diabetes, although both use the same clinical tests (blood glucose measurements), diagnostic tests are performed when an individual exhibits suggestive clinical features of diabetes, whereas screening aims to identify asymptomatic people who may already have or be at imminent risk of developing diabetes. The initial screening process may be followed by the appropriate diagnostic tests.

Screening for diabetes in primary care is an important public health issue. Despite the findings of important studies, such as the recent DHDS, there are still large gaps in the evidence needed to inform the decisions that must be made about any screening programme.

HOW STRONG IS THE CASE FOR SCREENING FOR TYPE 2 DIABETES?

Since setting up and implementing any screening programme can be a major undertaking with considerable costs, it is necessary to be able to evaluate the benefits of any proposal against recognised criteria (Davies 1997, The Expert Committee 2003a). The role of the National Screening Committee (NSC) is to provide advice about established and newly proposed screening programmes, with the aim of evaluating these against specified criteria, published in 1998 and divided into four areas: the condition, the test, the treatment and the screening programme (National Screening Committee 1998).

While the NSC crirteria are primarily designed to evaluate the cost-effectiveness of a screening programme, it may be more useful at practice level to evaluate the benefits of screening against established and recognised criteria (The Expert Committee 2003a). These can be summarised as follows:

1. The prevalence of the disease in the population to be screened must be sufficiently large to make mass screening practical.

The UK prevalence of diabetes in 2005 was estimated to be at least 4.67% and is rising. This prevalence is greater with increasing age, particular ethnicity and/or the presence of one or more other risk factors (hypertension, obesity, family history of diabetes) (Lawrence et al 2001).

2. The disease must be clearly defined to allow an accurate diagnosis.

There are diagnostic criteria for type 2 diabetes.

3. There should be evidence that the disease is undiagnosed in a significant proportion of individuals, so that a prompt early diagnosis is not inevitable in most of these cases.

It is estimated that about a quarter of diabetics are undiagnosed. There is good evidence that most type 2 diabetics have had their disease at least 4 to 7 years prior to clinical diagnosis (Harris et al 1992). Complications are already evident in many patients at diagnosis.

4. Early diagnosis of the disease must result in improved outcomes, including disease treatment and reduced impact of complications.

At least one-third of patients with type 2 diabetes have at least one complication present at diagnosis (UKDIABS 2000). However, type 2 diabetics identified by screening have:

• a significant cluster of cardiovascular risk factors, which can be improved by early intervention (Davies et al 1996)

• a lower prevalence of microvascular complications at diagnosis (UKDIABS 2000)

• a lower risk of death than in known type 2 diabetes patients after 4 years in a UK study (Croxson et al 1994).

Patients found to have IGT may benefit from interventions to reduce both their risk of becoming diabetic in the future and of their existing propensity to develop cardiovascular disease. Ultimately, the practice must be able to offer high-quality, appropriate and effective care on a consistent basis to newly diagnosed diabetics.

Diagnosis should not cause physical harm. There are numerous interventions in diabetes for which the benefits outweigh harm, although any intervention carries some degree of risk.

5. The screening test(s) used must be readily available and must have an acceptable specificity, sensitivity and predictive value in the population to be screened.

Reliable blood-glucose estimation is readily available by sending samples to an accredited local chemical pathology laboratory. The three methods of diagnosing diabetes (random, fasting and OGTT) have a high sensitivity (i.e. the likelihood of a positive result in patients with disease) and positive predictive value, because they define the disease, which is a lifelong diagnosis. However, a negative OGTT has a higher specificity (i.e. the likelihood of a negative result in patients without disease) than a negative fasting or random plasma glucose, because a proportion of diabetics can be missed by the latter two screening tests. The predictive value of a negative result is not conclusive, since currently non-diabetic individuals, particularly those from higher risk groups, may develop diabetes in the future. A result that demonstrates IGT, although negative for diabetes, does predict an increased future risk of developing diabetes.

6. The screening test(s) used must be acceptable to individuals being screened, and not cause significant adverse effects.

Venepuncture is invasive, but low risk when performed competently. Most would probably find it acceptable, provided informed consent is obtained. Screening, however, may cause anxiety, which needs to be minimised.

7. Case finding and treatment must be cost-effective, in relation to health expenditure as a whole.

This is unlikely if screening the general population solely for diabetes (Goodyer 2006). In addition to a lower yield of new cases, whole-population screening requires considerable resources. One study calculated the workload as 1 hour per week for a year to screen 620 patients during that period (Lawrence et al 2001). However, a screening programme that recruited well elderly Americans found that the presence of certain factors (age, gender, ethnicity, raised BMI, greater waist:hip ratio, hypertension) increased the yield to as much as one new case of diabetes diagnosed for every six individuals screened (Franse et al 2001).

Targeted screening of “high-risk” groups would require less resource and produce a higher yield (thus, be more practical) than screening the whole population. If additional information needed to estimate cardiovascular risk was gathered at screening, then all screened individuals, even the majority who screen negative for diabetes, could benefit from this estimation, particularly if interventions are available to reduce this risk.

8. Facilities and resources must be available to treat newly diagnosed cases.

Screening is likely to be undertaken using existing staff and facilities. It is unlikely that additional resources will be provided in the near future, either for screening or for the treatment of new cases; these will need to be found from existing resources.

9. Screening should be a systematic on-going process.

It should not merely be an isolated one-time effort. Depending upon what the National Screening Committee recommends, any screening programme will need to be managed, with an agreed set of quality assurance standards. Primary care is well placed to develop the accurate population and disease registers required to identify at-risk subjects, and to undertake regularly both recall and screening tests, provided that adequate systems are in place and maintained.

In 2001 it was argued that the case for whole-population screening for diabetes could not be made (Wareham & Griffin 2001). In the light of current evidence, the case for a cost-effective screening programme can be made only if either:

• a selected “high-risk” population is targeted

• other parameters contributing to cardiovascular risk are included in the screening.

PRACTICAL ISSUES IN SCREENING FOR TYPE 2 DIABETES

Any practice seeking to screen a population for type 2 diabetes needs to consider its answers to five practical questions:

1. Who to screen?

2. Which screening method(s) should be used?

3. How often to screen?

4. Should anything else be done at the time of screening?

5. Is the practice able to provide high-quality care to newly diagnosed diabetics without detriment to other patients?

Additional evidence is required to resolve fully the uncertainties that surround the first three questions, a point recognised by the Diabetes NSF.

Who to screen?

Instead of whole-population screening, the current consensus is for a targeted screening approach that concentrates on those at “high risk” with:

• previous IGT, metabolic syndrome (particularly obese or those with large waist circumference) or gestational diabetes (or high-birth-weight babies)

• existing cardiovascular disease

• strong family history of diabetes, i.e. first-degree relative

• increased-risk ethnicity, particularly older Indo-Asians and Afro-Caribbeans

• recurrent or major sepsis, or peripheral skin breakdown

• autoimmune endocrinopathies (e.g. primary hypothyroidism) or organ-specific disorders (e.g. polycystic ovary disease).

It is sensible to create a register of these “high-risk” individuals to facilitate recall.

Which screening method to use?

The logistics and greater resource demands of the OGTT render it unsuitable to be undertaken as the first-line test on a large scale in primary care. Therefore, a preliminary screening test needs to be performed to identify those who require an OGTT, as a second stage. The two tests that might serve as a first-stage screen are:

1. Fasting plasma glucose

2. Glycosylated haemoglobin (HbA1c).

Fasting plasma glucose test is less sensitive and specific, but may be a sensible compromise as the initial screening test. However, fasting plasma glucose has a reduced specificity because many individuals, particularly Indo-Asians, with a non-diabetic fasting level will have a diabetic level at 2 hours post glucose load. The STAR study suggests that a reasonable cut off for Indo-Asians under the age of 40 years is below 5.9 mmol/l. For Indo-Asians aged 40 to 75 years, an HbA1c below 6.2% is a reasonable cut-off for excluding diabetes (Davies 2006). Urine testing is simple, quick and cheap, but does not fulfil the diagnostic criteria for diabetes and a raised renal threshold may miss a diabetic blood glucose level.

In the event of diagnostic uncertainty or if the individual being screened is considered to be at high risk of diabetes, then an OGTT should be performed.

How often to screen?

Diabetes UK recommends (but supporting evidence is not cast-iron) that screening should not be performed more often than every 5 years in subjects with no risk factors, and no more often than every 3 years in those with one risk factor.

Should anything else be done at the time of screening?

For greater cost-effectiveness, individuals who are invited for screening should have a cardiovascular risk assessment that includes also blood pressure, lipids, smoking status and waist circumference.

Is the practice able to provide high-quality care to newly diagnosed diabetics without detriment to other patients?

While recognising the strong arguments in favour of targeted screening for diabetes, a practice should not forget that:

• screening will require additional resources, particularly of staff time

• the newly diagnosed diabetic patient should receive high-quality, appropriate and effective care on a consistent basis

• the delivery of care to other patients with existing diabetes and/or other significant problems must be maintained and not suffer when additional tasks are taken on.

THE LIKELY FUTURE SHAPE OF DIABETES SCREENING PROGRAMMES

The Diabetes Heart Disease and Stroke (DHDS) Project was set up to explore these practical issues. In its 2006 report, the overall “pick-up” rate for newly diagnosed diabetics was 4.3% of those screened (Goodyer 2006), but the rate varied, being higher in those populations where the prevelance of diabetes is greater. Although the project did not evaluate cost-effectiveness, the results do not support this in a screening programme searching solely for new cases of diabetes in a general population; rather, the screening for diabetes in a general population is best incorporated into a general vascular risk assessment.

This point both informs and underlies the Department of Health’s 2006 White Paper (DoH 2006) which proposes health MOTs at various stages of life. At the time of writing, a diabetes screening programme is unlikely to stand alone, but would be incorporated into a population-wide assessment, probably at significant age milestones (i.e. 40, 50 and 60 years), of parameters (that may include waist circumference, blood pressure, fasting lipids and glucose) which will identify a subset with metabolic syndrome that may need more frequent (every 3 years?) fasting blood glucose checks.

Ultimately, any systematic and sustained programme requires sufficient additional resources for screening, diagnostic tests and treatment; and their provision must be balanced against other demands made upon the practice. A practice’s decisions need to be based upon what is most feasible, both logistically and clinically.

PREVENTION OF TYPE 2 DIABETES MELLITUS

NSF

WHY PREVENT TYPE 2 DIABETES?

The predicted rise in prevalence makes all the greater the need to consider and to implement effective preventive strategies. Not all people are at equal risk of developing diabetes. Certain factors, such as age, ethnic origin and family history, are unalterable. However, many individuals who develop type 2 diabetes have a less than optimal lifestyle, over which they do have control. Factors associated with both increased cardiovascular risk and greater insulin resistance are already present in British children, particularly of Indo-Asian origin (Whincup et al 2002). A healthier diet and increased exercise reduces insulin resistance and cardiovascular risk.

POSSIBLE INTERVENTIONS TO PREVENT TYPE 2 DIABETES: EVIDENCE BASE

A number of published studies demonstrate that type 2 diabetes in high-risk individuals can be prevented either by improving lifestyle or by prescribing medication.

Modifying lifestyle

The Da Qing IGT and Diabetes Study (China), which followed up individuals with IGT over 6 years, showed a reduction by at least one-third of the incidence of new diabetes with either diet or exercise interventions or both (Pan et al 1997).

The Finnish Diabetes Prevention Study Group, followed up overweight individuals with IGT over a mean of 3.2 years, and compared “individualised” counselling aimed at reducing weight, modifying diet and increasing physical activity against controls. The cumulative incidence of new diabetes was reduced by 58% in the intervention group, or one case of diabetes was prevented for every 22 overweight individuals with impaired glucose tolerance “treated” for 1 year (Tuomilehto et al 2001). The benefits were sustained (APR 3.1%; RRR 43%) during a median follow-up of 3 years after the intervention was discontinued (Lindstom 2006).

The Diabetes Prevention Program (DPP) in the USA, followed up “high-risk” subjects over a mean of 2.8 years, and compared three arms: standard lifestyle recommendations plus placebo, standard lifestyle recommendations plus metformin, and an intensive programme of lifestyle modification (16 lessons covering diet, exercise and behaviour modification). The reduction in incidence of new cases of diabetes was 31% in the metformin group (treat 41.7 subjects for 1 year to prevent one new case) and 58% in the intensive programme group (treat 20.7 subjects for 1 year to prevent one new case) (Diabetes Prevention Program Research Group 2002).

To apply any of the above programmes on a sustained basis to a large population will require considerable effort and resources. However, the evidence does support modifications of lifestyle:

• increased levels of physical activity, along the lines of 30 minutes of “moderate” exercise (e.g. brisk walking) at least five times per week

• reduced fat and limited calorific intake

with the aim of achieving and maintaining weight reduction, which has considerable health benefits beyond preventing type 2 diabetes.

Medication

Any drug shown to be effective in preventing diabetes could potentially create a huge market for pharmaceutical companies. Even if proved effective, the colossal financial costs of treating large populations with some of these drugs may be prohibitive.

The following studies investigated drug therapy to prevent diabetes using different classes of glucose-lowering agents:

• The DPP (cited above) demonstrated a reduction in incidence of type 2 diabetes in high-risk subjects who used metformin, but lifestyle changes were even more effective as prevention (Diabetes Prevention Program Research Group 2002).

• In a multicentre randomised trial, Study to Prevent NIDDM (STOP-NIDDM), conversion of impaired glucose tolerance into type 2 diabetes was less in patients treated with acarbose than with placebo. However, 31% of the patients allocated to the acarbose group discontinued treatment early (Chiasson et al 2002). The use of acarbose is limited by frequent gastrointestinal side effects.

• In the Troglitazone in Prevention of Diabetes (TRIPOD) study, Hispanic women with “recent” gestational diabetes were randomised to receive either placebo or troglitazone (a glitazone now withdrawn from commercial sale). In the treated group there was a 56% relative reduction in the incidence of diabetes (Buchanan et al 2005).

In addition:

• Orlistat was added to lifestyle change in a group with BMI equal to or greater than 30 kg/m², with or without IGT, in the XENical in the prevention of Diabetes in Obese Subjects (XENDOS) study. After 4 years of treatment, the addition of orlistat corresponded to a 45% risk reduction of new diabetes in the IGT group, with no effect observed in those without IGT (Torgerson et al 2004).

Finally, there are data to suggest that blockade of the renin-angiotensin system may prevent diabetes:

• An incidental finding in the Heart Outcomes Prevention Evaluation (HOPE) trial was the lower rate of new diagnosis of diabetes in vascular high-risk patients aged over 55 years treated with the ACE inhibitor ramipril (Yusuf et al 2001). In the Diabetes Reduction Assessment with Ramipril and Rosiglitazone Medication (DREAM) study, rosiglitazone, unlike ramipril when compared to placebo, reduced significantly (hazard ratio 0.40, 95% Cl 0.35–0.46) the risk of patients with impaired fasting glucose or IGT developing diabetes. Although the study was well conducted, the results need to be interpreted and applied with caution, as rosiglitazone did increase fluid retention and heart failure and it is still an expensive drug being used to treat here something for which changes in lifestyles are still the mainstay.

• In the Losartan Intervention for Endpoint Reduction in Hypertension (LIFE) study, which looked at cardiovascular morbidity and mortality in patients with essential hypertension and left ventricular hypertrophy and compared treatment using the ARB losartan with atenolol, new-onset diabetes was less common in the losartan-treated group (Dählof et al 2002).

• In the VALUE study, although designed to evaluate cardiovascular outcomes in high-risk hypertensive subjects, new-onset diabetics was significantly less in the ARB valsartan treatment group compared with the CCB amlodipine treatment group (Julius et al 2004).

• A recent meta-analysis of 12 RCTs found that, in high-risk individuals (pre-diabetic conditions, such as IFG, metabolic syndrome), ACE inhibitors and ARBs were associated with relative risk reductions in the incidence of newly diagnosed diabetes by 25% in the pooled analysis (Abuissa et al 2005). However, further research is needed to determine the mechanism of action.

The HOPE and LIFE studies found that inhibition of the renin-angiotensin system may improve glucose tolerance. The LIFE study authors suggested an effect on insulin resistance, but other hypotheses suggest an increased insulin secretory response, either by inducing a raised pancreatic islet blood flow to secure a better early insulin response (Carlsson et al 1998) or as a result of the higher serum potassium levels associated with ACE-inhibitor use augmenting insulin secretion (Santoro et al 1992).

However, a recent systematic review concluded that, despite data that some drugs lowered the incidence of diabetes compared to placebo, no single pharmacological intervention can be definitively recommended for diabetes prevention. Further research is needed, using studies designed with the incidence of diabetes as the primary outcome and of sufficient duration to differentiate between real prevention and the delay or masking of the condition (Padwal et al 2005). Whilst the benefits of lifestyle changes may last beyond the period of intervention, it is unclear whether the benefit of medication is sustained. Although evidence supports the effectiveness of both lifestyle and pharmacological interventions in reducing the risk of IGT progressing to type 2 diabetes (Gillies 2007), selecting the optimal intervention requires careful consideration of benefit vs. harm, patient preference, likely concordance, available resources and other issues that are, as yet, unresolved.