Types 1 and 2 diabetes together account for approximately 98% of cases, with type 2 diabetes making up approximately 85–90% of these. Rarer causes of diabetes collectively account for only 2–3% of cases, but it is crucial they are diagnosed correctly to ensure that the correct treatments are pursued (Pickup and Williams, 2002).
The current classifications of diabetes were introduced by the American Diabetes Association (ADA; Expert Committee on the Diagnosis and Classification of Diabetes Mellitus, 1997) and the World Health Organization (WHO; Alberti et al, 1998) in the late 1990s. They replaced previous WHO guidelines (Expert Committee on Diabetes Mellitus, 1980; 1985) which were based on the extent of a patient’s insulin dependency.
The current classification is based on both the clinical stage of glucose intolerance and the underlying aetiological cause (see Table 1). A key advantage of this system is the recognition that many different processes can result in hyperglycaemia. The emphasis on aetiology can help clinicians consider the underlying cause of the diabetes they see in their surgery and tailor therapy more appropriately. It also creates a framework that will evolve as more specific causes of diabetes are identified and characterised.
With respect to nomenclature, both the ADA and WHO re-introduced the terms “type 1” (replacing insulin-dependent diabetes mellitus or juvenile onset) and “type 2” (replacing non-insulin-dependent diabetes mellitus or mature onset) diabetes. This terminology will be used in here.
Diabetes in young adults
The biochemical diagnosing of diabetes is straightforward but should be only one part of clinical assessment. Making an aetiological diagnosis can sometimes be more challenging, and may not be immediately possible. A number of aetiological possibilities should be considered, these are represented graphically in Figure 1 and discussed below.
Type 1 diabetes
Clinical features at diagnosis suggesting type 1 diabetes include a short duration (weeks) of intense osmotic symptoms, marked weight loss, and the presence of moderate or heavy ketonuria. If available, islet cell antibody screening may be positive in most but not all people with type 1 diabetes. Case study 1 illustrates an example of type 1 diabetes, and case study 2 illustrates the diagnostic challenge of type 1 diabetes complicated by acromegaly (see Appendix).
Type 2 diabetes
Type 2 diabetes is a growing problem in Western countries associated with rising levels of obesity (Hannon et al, 2005). Recent conservative estimates suggest that approximately 20000 obese children in the European Union have type 2 diabetes, and a further 400000 are glucose intolerant (Lobstein and Jackson-Leach, 2006). Features pointing to a diagnosis of type 2 diabetes in youth would include a family history, weight 20% or more above the ideal, a high waist–hip ratio and other features of the metabolic syndrome, mild or absent traditional type 2 symptoms, a lack of significant weight loss at presentation and acanthosis nigricans. Certain ethnic groups are at high-risk of type 2 diabetes and are over-represented, as are young women.
Maturity onset diabetes of the young
Maturity onset diabetes of the young (MODY) is a sub-type of diabetes associated with autosomal dominant inheritance, first described by Tattersall (1974). To date, six MODY sub-types have been described, all of which result in insulin secretory defects of varying degrees of severity (Hattersley and Pearson, 2006). Insulin resistance is not a feature of MODY and the clinical course is determined by the specific genetic defect. MODY 2, which results from a defect in the glucokinase gene, is associated with life-long mild fasting hyperglycaemia. People with MODY 2 do not require tablets or insulin to treat their diabetes, and as their HbA1c levels are at or only slightly above that of healthy people and remain stable, they are at low risk of diabetic complications (Hattersley and Pearson, 2006).
In contrast, the more common MODY 1 and MODY 3 sub-types, resulting from genetic defects of hepatic nuclear transcription factors (HNF), are associated with progressive beta-cell failure and require progressive intensification of treatment. The risk of long-term complications is significantly greater than for MODY 2. Recent reports have suggested that those with MODY 3 (HNF1-alpha mutation) have enhanced sensitivity to sulphonylureas (Pearson et al, 2003). This observation has allowed some insulin-treated people – many of whom had been misdiagnosed with type 1 diabetes – to replace their insulin with sulphonylurea tablets without risk of ketoacidosis, enabling them to achieve equivalent or sometimes improved glycaemic control (Shepherd et al, 2003).
MODY should be suspected in those under the age of 25 with impaired fasting glucose or diabetes, and with a family history of diabetes for at least two consecutive generations. Markers of auto-immunity are negative. Genetic tests are available for suspected cases.
Maternally inherited diabetes and deafness
Maternally inherited diabetes and deafness (MIDD) was first described in 1992 (van den Ouweland et al, 1992). It is a rare cause of diabetes that should be considered if sensorineural deafness and a family history of diabetes and deafness in the maternal lineage is present. Hearing loss usually occurs in early adult life and commonly precedes the onset of diabetes. Other features can include stroke-like episodes, retinal pigmentary changes, myopathy and renal disease (Murphy et al, 2008). If suspected on clinical grounds, the diagnosis can be confirmed with genetic testing for the mitochondrial DNA point mutation at position 3243A>G. Diagnostic confirmation enables genetic counselling to be provided. Metformin therapy is not recommended as people with MIDD are at increased risk of lactic acidosis.
Types 1 and 2 diabetes, MODY and MIDD, as well as all the other possibilities, present diagnostic challenges to the clinician when diagnosing diabetes. Two key points to consider are whether the person in question has (i) marked insulin deficiency, or (ii) risk factors for type 1 diabetes.
Diabetes in older adults
The UK Prospective Diabetes Study investigators found that 12% of people with an initial clinical diagnosis of type 2 diabetes were later found to have positive autoantibodies, suggesting slow-onset type 1 diabetes (Desai and Clark, 2008). This condition has been termed latent autoimmune diabetes of adults (LADA). Although still a controversial area, it is likely that LADA represents an extension of type 1 diabetes into adult life with a less-intense, and therefore slower, autoimmunological destruction of beta-cells (Desai and Clark, 2008). People with LADA are more likely to progress to insulin treatment (usually within 6 years of diagnosis) than antibody-negative people, but they can often be managed with conventional oral hypoglycaemic agents for months or even years (Turner et al, 1997).
Clinical features that should lead clinicians to suspect LADA include age <50 years, a body mass index of <25kg/m2, acute symptoms and a personal or family history of autoimmune disease. Two or more of these features have a 90% sensitivity and 71% specificity for identifying LADA (Fourlanos et al, 2006). Such an assessment can be of help in deciding when insulin initiation is needed. Case study 3 (see Appendix) illustrates a case of LADA.
The current classification of diabetes challenges clinicians to consider the underlying aetiology, as well as the clinical stage of glucose intolerance, when diagnosing the condition. After looking at the aetiology of diabetes, one can think more widely than just ‘type 1’ or ‘type 2’ during their consultations, in both newly diagnosed and known cases of diabetes.
Case study 1: Type 1 diabetes
A 27-year-old male fitness instructor (body mass index of 23.4kg/m2) presented for an army medical and was found to have glycosuria. A subsequent oral glucose tolerance test revealed fasting and 2-hour glucose concentrations of 8.4 and 17.9mmol/L, respectively. He had no symptoms and screening for urinary ketones was negative. Without further assessment it would only be possible to confirm the diagnosis of diabetes, not the underlying cause.
He had no family history of diabetes (which makes monogenic diabetes less likely), he did not have an insulin-resistant phenotype (making type 2 diabetes less likely) and his islet antibody screening test was positive. This confirmed a diagnosis of type 1 diabetes diagnosed, by chance, in the pre-symptomatic phase.
Case study 2: Type 1 diabetes complicated by acromegaly
A 22-year-old woman with a 4-year history of type 1 diabetes diagnosed elsewhere was referred for insulin intensification and the Dose Adjustment For Normal Eating education programme. Her HbA1c was 10.8%. She was awaiting further investigation of visual problems, and had been found to have a bitemporal hemianopia by her optitian.
On review, a bitemporal field defect was confirmed together with bilateral optic atrophy. The patient had features of acromegaly, confirmed by grossly raised growth hormone and insulin-like growth factor levels. Pituitary imaging revealed a macroadenoma. Following transsphenoidal surgery and normalisation of growth hormone levels, her HbA1c fell to 5.1% and she required only 40% of her pre-operative insulin dose.
Case study 3: LADA
A 67-year-old woman was admitted to hospital with urinary sepsis and poorly controlled diabetes. She had been diagnosed with type 2 diabetes 6 years previously and was taking gliclazide and metformin. Her glucose level on admission was 33.3mmol/L and she was treated with a standard insulin sliding scale, fluids and antibiotics. Once the sepsis was controlled, several attempts were made to re-start oral hypoglycaemic agents, but capillary glucose levels increased. She was eventually discharged on her admission diabetes medication, with blood glucose levels ranging between 10 and 18 mmol/L.
Her GP subsequently referred her urgently to the diabetes clinic. Clinical assessment revealed a body mass index of 21kg/m2 and an HbA1c of 13%, a personal history of autoimmune thyroid disease and a twin sister who also had diabetes and who had commenced insulin 6 weeks previously. Her islet cell antibodies were positive and she commenced insulin the same day with an excellent clinical response.