Classification
Diabetes can be classified into the following general categories[1]:
1. Type 1 diabetes: due to autoimmune β-cell destruction, usually leading to absolute insulin deficiency, including latent autoimmune diabetes of adulthood
2. Type 2 diabetes: due to a non-autoimmune progressive loss of adequate β-cell insulin secretion frequently on the background of insulin resistance and metabolic syndrome
3. Specific types of diabetes: due to other causes, e.g., monogenic diabetes syndromes (such as neonatal diabetes and maturity-onset diabetes of the young), diseases of the exocrine pancreas (such as cystic fibrosis and pancreatitis), and drug- or chemical-induced diabetes (such as with glucocorticoid use, in the treatment of HIV/AIDS, or after organ transplantation)
4. Gestational diabetes mellitus: diabetes diagnosed in the second or third trimester of pregnancy that was not clearly overt diabetes prior to gestation
Figure 1 Type of diabetes (WHO 2019)[2]
Assigning diabetes type in clinical settings
The Expert group decided to focus on providing a practical clinical guide for clinicians faced with the challenge of assigning a type of diabetes to individuals at the time of presentation with hyperglycaemia to help choose an appropriate treatment, particularly whether insulin treatment should be started. With the limited access in most countries throughout the world to laboratory investigations which might assist with the classification of an individual, the Expert group opted for providing guidance on identifying clinical subtypes, while recognizing the limitations of this approach[2].
Countries and centres able to test for genes, for islet autoimmunity and for endogenous insulin production can use these to increase the accuracy of clinical subtyping of diabetes.
Steps in clinical subtyping an individual first diagnosed with diabetes:
1. Confirm diagnosis of diabetes in an asymptomatic individual
2. Exclude secondary causes of diabetes
3. Consider the following which may assist in differentiating subtypes:
· age at diagnosis
· family history
· physical findings, especially presence of obesity
· presence of features of ketosis or ketoacidosis
4. Note presence or absence of ketosis or ketoacidosis
5. Perform diagnostic tests if available (β-cell autoantibodies, C-peptide)
It may not be possible to definitively subtype an individual at the time of presentation and classification may only become possible over time when, for example, the longer term insulin requirement to manage glycaemia has been established. Consequently, the individual should be monitored and the appropriateness of the assigned diabetes classification should be regularly reviewed, especially in individuals without the classical features of diabetes subtypes.
Age at diagnosis as a guide to subtyping diabetes
The following presents the major diagnostic diabetes subtypes according to usual age at diagnosis. The group with highest heterogeneity and risk of misclassification is that of young adults aged 20 to 40 years.
Age < 6 months
Types of diabetes:
· Monogenic neonatal diabetes – transient or permanent
· Type 1 diabetes – extremely rare
The majority presenting in this age group will have monogenic neonatal diabetes – either transient or permanent. This can only be definitely diagnosed by genetic testing, but where this is not possible, careful assessment of the clinical course should establish either the transient nature of the diabetes or that insulin therapy is not required. While mostly occurring before 6 months of age, neonatal diabetes may present up to 12 months of age. T1DM is extremely rare in the first year of life.
Age 6 months to < 10 years
Types of diabetes:
· Type 1 diabetes
· Monogenic neonatal diabetes – transient or permanent
· Type 2 diabetes – rare before puberty
T1DM is the most common type of diabetes in this age group. The diagnosis of T1DM is relatively straightforward in normal-weight individuals who present with ketoacidosis aged 10 years or younger and who are autoantibody-positive. Both T2DM (even if the individual is obese) and monogenic diabetes (except for neonatal diabetes and GCK) are rare before puberty.
Age 10 to < 25 years
Types of diabetes:
· Type 1 diabetes
· Type 2 diabetes
· Monogenic diabetes
The relative proportions of different types of diabetes in this age group differ by ethnic group. T2DM with onset in youth occurs most often during the second decade of life and there is usually a strong family history of T2DM in first- and second-degree relatives. While all ethnic groups can be affected, prevalence is much lower in populations of European descent. It is the predominant form of youthonset diabetes in some countries – 90% in Hong Kong, 60% in Japan and 50% in Taiwan. The presence of obesity varies with ethnicity and is found in virtually all people with youth-onset T2DM in the USA and Europe; by contrast, half of Asian children with T2DM have a normal weight. The presentation of youth-onset T2DM can vary from asymptomatic hyperglycaemia detected through screening or during routine physical examination, to ketoacidosis in up to 25% of patients or hyperglycaemic hyperosmolar state.
Features favouring a diagnosis of T2DM rather than T1DM at diagnosis include:
· Overweight or obesity
· Age above 10 years
· Strong family history of T2DM
· Acanthosis nigricans
· Undetectable islet autoantibodies (if measured)
· Elevated or normal C-peptide (if assessed)
The peak age at diagnosis of people with monogenic diabetes is around 15–20 years of age (derived from UK referrals data) and this subtype of diabetes should be considered in this age group.
Clinical prediction models have been developed that can be used to discriminate between monogenic diabetes and T1DM and T2DM in individuals diagnosed between the ages of 1 and 35 years. The models are available online at www.diabetesgenes.org.
Age 25 to 50 years
Types of diabetes:
· Type 2 diabetes
· Slowly evolving immune-mediated diabetes
· Type 1 diabetes
Slowly evolving immune-mediated diabetes in adults usually presents after the age of 25 years and mostly after the age of 35 years. It is found in around 10% of people presenting as T2DM in Europe, North America, and Asia. Pancreatic autoantibodies (especially GAD autoantibodies) are a feature. Subtypes of diabetes other than T2DM should be considered in adults who have a normal weight and are without other metabolic syndrome features. These features are present in 33% of adults with slowly evolving immune-mediated diabetes, compared to 83% in people with T2DM.
T1DM accounts for an estimated 5% of diabetes diagnosed between the ages of 31 and 60 years. T1DM should be especially considered on clinical grounds in adults presenting with one or more of the following: ketosis, rapid weight loss, age of onset below 50 years, BMI below 25 kg/m2, or personal and/or family history of autoimmune disease.
Age > 50 years
Types of diabetes:
· Type 2 diabetes
· Slowly evolving immune-mediated diabetes
· Type 1 diabetes
The same considerations apply in this age group as those that apply to individuals aged 20–50 years, although T1DM presenting in this age group is less common (see section “Age 25 to < 50 years”).
Differential diagnosis of individuals presenting with ketosis or ketoacidosis
Types of diabetes:
· Type 1 diabetes
· Ketosis-prone type 2 diabetes
· Type 2 diabetes with onset in youth
· Type 2 diabetes with onset in adult
In people with diabetes diagnosis before the age of 20 years, diabetic ketoacidosis at the time of diagnosis occurs in approximately 25% but varies within and across populations. The prevalence decreases with age from 37% in children aged 0 to 4 years to 15% among those aged 15 to 19 years. Diabetic ketoacidosis prevalence is significantly higher in people with T1DM (about 30%) compared with T2DM (about 10%). However, higher proportions (nearly 20%) have been reported in young-onset T2DM in the US.
The possibility of ketosis-prone type 2 diabetes should be considered in adults of all ethnicities (except caucasian populations) who present with ketosis, but otherwise have most features of T2DM. However, diagnosis can only be established over time.
The increasing popularity of ketogenic diets may also influence the clinical presentation and should be considered.
Various schemes have been proposed for classifying patients with ketosis-prone diabetes. A comparison of these systems reported that the AB system based on the presence of autoimmunity (A +/−) or preserved β-cell function (β +/–) performed best with regard to accuracy and value in predicting long-term insulin dependence and clinical phenotype. However, this scheme is not widely used and requires assessments not generally available in most clinical care settings.
References:
Diagnosis and test method
1. Diagnostic tests
1.1 Blood glucose and HbA1c:
The diagnosis of diabetes may be based on venous plasma glucose rather than capillary glucose measurements. The diagnostic criteria for diabetes are shown in Figure 1[1]. Fasting glucose refers to the measurement of blood glucose levels following a minimum 8-hour period of fasting, during which only water is allowed. The 2-hour postprandial glucose test involves consuming a glucose load equivalent to 75-g of anhydrous glucose dissolved in water. To diagnose diabetes in the absence of hyperglycemia symptoms, it is necessary to have two abnormal test results.
In 2011, the World Health Organization (WHO) recommended the use of glycosylated hemoglobin (HbA1c) for the diagnosis of diabetes in countries and regions where it is available, with a diagnostic cut point of HbA1c greater than and equal to 6.5%. However, diabetes can only be diagnosed based on venous plasma glucose levels in the following cases: sickle cell disease, pregnancy (intermediate and late), glucose-6-phosphate dehydrogenase deficiency, AIDS, hemodialysis, recent blood loss or transfusion, and erythropoietin therapy. In addition, HbA1c is not recommended to screen for cystic fibrosis-associated diabetes. HbA1c testing should be conducted in a certified laboratory using methods that are endorsed by the National Glycohemoglobin Standardization Program (NGSP). These standardized tests are essential for assessing diabetes control and detecting complications related to the condition.
Fasting plasma glucose, 2-h plasma glucose values after 75-g oral glucose tolerance test (OGTT), or HbA1c can be used alone for epidemiological investigations or population screening. If the purpose of OGTT is only to clarify the state of glucose metabolism, only fasting and 2-h post-glucose load glucose need to be measured. Certain epidemiological data show that the rate of missed diagnosis of diabetes is high when only fasting glucose is checked, and the ideal investigation is to test fasting glucose, 2-h glucose after OGTT, and HbA1c at the same time. Blood glucose is not used as diagnostic criteria except for the OGTT time points. It is recommended that people whose blood glucose levels have reached impaired glucose regulation should undergo OGTT to improve the diagnosis of diabetes mellitus.
The same tests may be used to screen for and diagose diabetes and to detect individuals with prediabetes. Prediabetes refers to Impaired Glucose Tolerance (IGT) and/or Impaired Fasting Glucose (IFG). The American Diabetes Association recommends prediabetes can be diagnosed when an HbA1c value between 39-47 mmol/mol (5.7-6.4%) or fasting blood glucose between 5.6-6.9 mmol/L (100-125 mg/dL), or a 2-hour glucose tolerance between 7.8-11.0 mmol/L (140-199 mg/dL). People with prediabetes are at increased risk for type 2 diabetes and diabetes-related complications[2]. Five years after IGT or IFG diagnosis, the cumulative incidence of progression to type 2 diabetes is estimated to be 26% and 50%, respectively[1].
Figure 1 Diabetes diagnosis criteria
Image source: IDF Global Diabetes Map (10th edition)
Diagnosis
In a patient with classic symptoms, measurement of plasma glucose is sufficient to diagnose diabetes (symptoms of hyperglycemia or hyperglycemic crisis plus a random plasma glucose ≥200 mg/dL [11.1 mmol/L]). In these cases, knowing the plasma glucose level is critical because, in addition to confirming that symptoms are due to diabetes, it will inform management decisions. Some health care professionals may also want to know the HbA1c to determine the chronicity of the hyperglycemia. The criteria to diagnose diabetes are listed in Figure 1 [2].
Differential diagnosis for T1DM
Adults with new-onset type 1 diabetes can present with a short duration of illness of 1–4 weeks or a more slowly evolving process that can be mistaken for type 2 diabetes. Several other types of diabetes, for example monogenic diabetes, can be misdiagnosed as type 1 diabetes. In older adults, pancreatic cancer may present with diabetes and weight loss. A new and emerging issue is the development of profound insulin deficiency associated with the use of immune check-point inhibitors, which may present with hyperglycaemia and diabetic ketoacidosis (DKA).
Most of the available data discussed below are derived from White European populations and may not be representative of other ethnic groups. The clinical presentation may differ, but the classical triad of thirst and polydipsia, polyuria and weight loss are common symptoms of type 1 diabetes. Accurate classification of the type of diabetes has implications beyond the use of insulin treatment; education, insulin regimen, use of adjuvant therapies, access to newer technologies, need for psychosocial support to address the profound psychological impact of the diagnosis of diabetes and concurrent disease screening may all depend on the diagnosis an individual receives. Furthermore, accurate diagnosis allows an assessment of the risk of diabetes in first-degree relatives and appropriate counselling. Although profound insulin deficiency is the hallmark of type 1 diabetes, some adults with type 1 diabetes maintain some insulin secretion for years after diagnosis and may not require insulin treatment at diagnosis, leading to diagnostic uncertainty about the type of diabetes and its management[3].
Differentiating T1DM from T2DM
Identifying whether an adult with newly diagnosed diabetes has type 1 diabetes may be challenging where the individual has features pointing towards both type 1 diabetes and type 2 diabetes, such as an older adult with a low or normal BMI or young adult with an elevated BMI. Ketoacidosis, once considered pathognomonic of type 1 diabetes, may occur in ketosis-prone type 2 diabetes. Misclassification of type 1 diabetes in adults is common and over 40% of those developing type 1 diabetes after age 30 years are initially treated as having type 2 diabetes. From a patient perspective, a misdiagnosis of type 2 diabetes can cause confusion and misunderstanding, especially for those with type 1 diabetes who have overweight or obesity. This can impair the acceptance of the diagnosis and future management plans. No single clinical feature confirms type 1 diabetes in isolation. The most discriminative feature is younger age at diagnosis (<35 years), with lower BMI (<25 kg/m2), unintentional weight loss, ketoacidosis and glucose >20 mmol/L (>360 mg/dL) at presentation also being informative. Other features classically associated with type 1 diabetes, such as ketosis without acidosis, osmotic symptoms, family history or a history of autoimmune diseases are weak discriminators.
The very strong relationship between type 2 diabetes incidence and age means that even "classical" features of type 1 diabetes may have a limited predictive value in older adults, as type 2 diabetes in this age group is so common. The majority of older adults with low BMI will have type 2 diabetes, even more so when a person’s ethnicity is associated with high type 2 diabetes risk. Rapid progression to insulin treatment (<3 years) is strongly suggestive of type 1 diabetes at any age. The diagnosis of type 1 diabetes can be more difficult in adults who progress to insulin therapy more slowly. Controversy remains as to whether latent autoimmune diabetes of adulthood (LADA) is a discrete subtype, a milder form of type 1 diabetes, or a mixture of some individuals with type 1 diabetes and others with type 2 diabetes[3].
Differentiating T1DM from monogenic diabetes
Monogenic diabetes is found in approximately 4% of those diagnosed with diabetes before the age of 30 years; the likelihood of monogenic diabetes rises to 20% where islet antibodies are negative and C-peptide secretion is maintained. Monogenic diabetes is commonly mistaken for type 1 diabetes because of the young age at onset. A diagnosis of monogenic diabetes allows specific treatment with discontinuation of insulin in many cases and has implications for family members and screening for concurrent conditions[3].
Investigation of an adult with suspected T1DM
An algorithm for the investigation of adults with suspected type 1 diabetes is shown in Figure 2[3].
Figure 2 Flow chart for investigation of suspected type 1 diabetes in newly diagnosed adults, based on data from White European populations
Image source: A consensus report by the American Diabetes Association (ADA) and the European Association for the Study of Diabetes (EASD)
1.2 C-peptide measurement:
Beyond 3 years after diagnosis, where there is uncertainty about the type of diabetes, a random C-peptide measurement (with concurrent glucose) within 5 hours after a meal is recommended. If a person is treated with insulin, this test should always be performed before discontinuing insulin to rule out severe insulin deficiency.
Persistent C-peptide >600 pmol/L (non-fasting) is strongly suggestive of type 2 diabetes. Patients with C-peptide in this range are often able to replace insulin with other medications. Routine C-peptide testing in individuals with clinically diagnosed type 1 diabetes for at least 3 years has led to reclassification in 11% of those with adult-onset diabetes. In contrast, low or absent C-peptide confirms the diagnosis of type 1 diabetes. Although low C-peptide concentrations may occur in some types of secondary diabetes and very long-standing type 2 diabetes, these situations are unlikely to be confused with type 1 diabetes; however, in some cases, investigation of other types of diabetes may be appropriate.
Plasma C-peptide is the recommended test when available, with slightly higher performance than urine measurement. The latter may be confounded by impaired renal function. When a urine C-peptide: creatinine ratio is used, a value <0.2 nmol/ mol can be used to define severe insulin deficiency[3].
2. Islet autoantibodies
An evaluation of islet autoantibodies at diagnosis is recommended as the primary investigation of an adult with suspected type 1 diabetes. GAD should be the primary antibody measured and, if negative, should be followed by islet tyrosine phosphatase 2 (IA2) and/or zinc transporter 8 (ZnT8) if these tests are available. Measurement of Islet cell antibody (ICA) is no longer recommended because it is an imprecise biological assay that has been superseded by the direct measurement of single antibodies.
In people with clinical features suggestive of type 1 diabetes, the presence of one or more positive islet autoantibodies is highly predictive of rapid progression and severe insulin deficiency and these individuals should be considered to have type 1 diabetes, even if they did not require insulin at diagnosis. Because positive GAD antibodies may be found at a low level in adults without autoimmune diabetes and false positives can occur, GAD should only be measured in those suspected of having type 1 diabetes.
The absence of autoantibodies does not rule out type 1 diabetes, as approximately 5–10% of white European people with new-onset type 1 diabetes are negative for islet antibodies, and further consideration of the diagnosis is necessary. Furthermore, antibodies may disappear over time. In those diagnosed under the age of 35 years, type 1 diabetes is still the most likely diagnosis, especially in the absence of clinical features of type 2 diabetes or monogenic diabetes. In those over 35 years of age, type 2 diabetes is increasingly likely in the absent islet autoantibodies and older age.
It is important to make a clinical decision about how to treat a person with diabetes. Regardless of any features of type 2 diabetes or the absence of islet antibodies, if there is a clinical suspicion of type 1 diabetes, the individual should be treated with insulin. However, for some individuals whose clinical course is more suggestive of type 2 diabetes, a trial of non-insulin therapy may be appropriate. Those whose diabetes is treated without insulin will require careful monitoring and education so that insulin can be rapidly initiated in the event of glycemic deterioration. Type 2 diabetes and other types of diabetes should be considered in all age groups, but in those younger than 35 years, negative islet antibodies should raise the suspicion of monogenic diabetes[3].
3. Genetic testing
Because monogenic diabetes was less likely to be considered in the past, molecular genetic testing for neonatal diabetes should be considered for all people with type 1 diabetes, regardless of current age, who were diagnosed before the age of 6 months, as more than 80% have monogenic neonatal diabetes, and the 30–50% with ATP-sensitive potassium (KATP) channel mutations can replace insulin with sulfonylureas.
Monogenic diabetes should be considered in individuals with one or more of the following features: age at diagnosis of less than 35 years, HbA1c <58 mmol/mol (7.5%) at diagnosis, a parent with diabetes, and features of a specific monogenic cause (e.g. renal cysts, partial lipodystrophy, maternally inherited deafness, severe insulin resistance in the absence of obesity). A Monogenic Diabetes Prediction Model risk calculator (www.diabetesgenes.org/mody-probability-calculator; accessed August 20, 2021) may also be used to determine which individuals diagnosed between 6 months and 35 years of age are at increased risk for monogenic diabetes. Those at increased risk should undergo islet autoantibody and C-peptide testing. Molecular genetic testing should only be considered if the antibodies are negative and the non-fasting C-peptide is >200 pmol/L. Molecular genetic testing is not widely available[3].
References:
