Diabetes mellitus represents a cluster of metabolic disorders distinguished by hyperglycemia. This condition arises from impairments in insulin secretion, insulin action, or a combination of both. Persistent hyperglycemia in diabetes is linked to long-term damage, dysfunction, and eventual failure of various organs, notably the eyes, kidneys, nerves, heart, and blood vessels. Understanding the diagnosis and classification of diabetes is crucial for effective diabetes care and management. This article delves into the definition, description, and classification of diabetes mellitus, drawing upon established criteria and guidelines relevant to the 2006 and subsequent understandings of the disease.
The development of diabetes involves several pathogenic processes, ranging from autoimmune destruction of pancreatic β-cells leading to insulin deficiency, to abnormalities causing resistance to insulin’s action. At the core of metabolic disturbances in diabetes—affecting carbohydrate, fat, and protein metabolism—is the insufficient action of insulin on target tissues. This deficiency stems from either inadequate insulin secretion or reduced tissue responsiveness to insulin at various points in the hormone’s action pathways. Often, both impaired secretion and action coexist in a patient, making it challenging to pinpoint the primary cause of hyperglycemia.
Symptoms of significant hyperglycemia include increased urination (polyuria), excessive thirst (polydipsia), unexplained weight loss, sometimes increased hunger (polyphagia), and blurred vision. Chronic hyperglycemia may also hinder growth and increase susceptibility to infections. Acute, life-threatening complications include hyperglycemia with ketoacidosis or the nonketotic hyperosmolar syndrome.
Long-term diabetes complications are extensive and severe. They encompass retinopathy, potentially leading to blindness; nephropathy, which can progress to renal failure; peripheral neuropathy, increasing risks of foot ulcers, amputations, and Charcot joints; and autonomic neuropathy, causing gastrointestinal, genitourinary, and cardiovascular issues, along with sexual dysfunction. Individuals with diabetes also face a heightened risk of atherosclerotic cardiovascular, peripheral arterial, and cerebrovascular diseases. Hypertension and lipoprotein metabolism abnormalities are frequently observed in diabetic populations.
Diabetes cases predominantly fall into two broad etiopathogenetic categories: type 1 and type 2 diabetes. Type 1 diabetes is characterized by an absolute deficiency in insulin secretion. Individuals at higher risk can often be identified through serological evidence of autoimmune processes in pancreatic islets and genetic markers. Type 2 diabetes, the more common form, results from insulin resistance combined with an insufficient compensatory insulin secretion. In type 2 diabetes, a level of hyperglycemia sufficient to cause pathological and functional changes may exist for a considerable period without noticeable clinical symptoms. This asymptomatic phase allows for the detection of carbohydrate metabolism abnormalities through plasma glucose measurements in fasting states or post-oral glucose load.
Glycemic levels in diabetes can fluctuate over time, influenced by the progression of the underlying disease and treatment efficacy, as illustrated in Figure 1. A disease process might be present without yet causing hyperglycemia, or it may manifest as impaired fasting glucose (IFG) and/or impaired glucose tolerance (IGT), conditions that do not fully meet diabetes diagnostic criteria but indicate increased risk. Effective glycemic control in some diabetic individuals can be achieved through lifestyle modifications like weight reduction and exercise, or with oral glucose-lowering agents, negating the need for insulin. Conversely, some individuals with residual insulin secretion may require exogenous insulin to maintain glycemic control, yet can survive without it, while those with extensive β-cell destruction and no insulin secretion are insulin-dependent for survival. The severity of metabolic abnormalities in diabetes is dynamic, capable of progression, regression, or stabilization. Therefore, the degree of hyperglycemia is more reflective of the underlying metabolic process’s severity and its management than the nature of the process itself.
Figure 1. Disorders of glycemia: etiologic types and stages. Illustrates the progression and stages of different etiologic types of glycemia disorders, highlighting the variability in disease progression and the potential for different interventions at each stage.
CLASSIFICATION OF DIABETES MELLITUS AND CATEGORIES OF GLUCOSE REGULATION DISORDERS
Assigning a specific type of diabetes can be context-dependent at diagnosis, with many individuals not fitting neatly into a single category. For instance, gestational diabetes mellitus (GDM) may evolve into persistent hyperglycemia post-delivery, leading to a diagnosis of type 2 diabetes. Similarly, diabetes induced by exogenous steroids might resolve upon discontinuation, only to reappear years later due to conditions like recurrent pancreatitis. The effect of thiazide diuretics on diabetes development is another example; while they rarely cause severe hyperglycemia alone, they can exacerbate underlying type 2 diabetes. For clinicians and patients, understanding the pathogenesis of hyperglycemia and effectively treating it is more critical than merely labeling the diabetes type.
Type 1 Diabetes: β-cell Destruction Leading to Insulin Deficiency
Type 1 diabetes is characterized by β-cell destruction, typically resulting in absolute insulin deficiency. It is further divided into immune-mediated and idiopathic forms.
Immune-Mediated Diabetes
This form, accounting for 5–10% of diabetes cases, was previously known as insulin-dependent diabetes, type 1 diabetes, or juvenile-onset diabetes. It results from autoimmune destruction of pancreatic β-cells. Markers of this autoimmune process include islet cell autoantibodies, insulin autoantibodies, GAD (GAD65) autoantibodies, and autoantibodies to tyrosine phosphatases IA-2 and IA-2β. One or more of these autoantibodies are present in 85–90% of individuals at initial diagnosis of fasting hyperglycemia. The disease also shows strong HLA associations, particularly with DQA and DQB genes, and is influenced by DRB genes, with HLA-DR/DQ alleles being either predisposing or protective.
The rate of β-cell destruction varies in immune-mediated diabetes. It can be rapid, especially in children, or slower, particularly in adults. Some patients, especially children and adolescents, may initially present with ketoacidosis. Others may have mild fasting hyperglycemia that can quickly escalate to severe hyperglycemia and/or ketoacidosis under stress, such as infection. Adults may retain enough residual β-cell function to prevent ketoacidosis for years, eventually becoming insulin-dependent and at risk for ketoacidosis. At advanced stages, insulin secretion is minimal or absent, indicated by low or undetectable plasma C-peptide levels. While commonly diagnosed in childhood and adolescence, immune-mediated diabetes can occur at any age.
Genetic predispositions and environmental factors, still not fully understood, contribute to autoimmune β-cell destruction. Although patients are typically not obese at diagnosis, obesity does not preclude this diagnosis. These patients are also susceptible to other autoimmune disorders like Graves’ disease, Hashimoto’s thyroiditis, Addison’s disease, vitiligo, celiac sprue, autoimmune hepatitis, myasthenia gravis, and pernicious anemia.
Idiopathic Diabetes
Idiopathic type 1 diabetes lacks known etiologies and evidence of autoimmunity. Patients with this form experience permanent insulinopenia and are prone to ketoacidosis, but without autoimmune markers. Although a minority of type 1 diabetes cases are idiopathic, it is more prevalent among those of African or Asian descent. This form is characterized by episodic ketoacidosis and varying degrees of insulin deficiency between episodes. It shows strong inheritance patterns but lacks immunological evidence of β-cell autoimmunity and HLA associations. Insulin replacement therapy needs may fluctuate in affected individuals.
Type 2 Diabetes: Insulin Resistance and Relative Insulin Deficiency
Type 2 diabetes, encompassing 90–95% of diabetes cases, was formerly referred to as non–insulin-dependent diabetes or adult-onset diabetes. It is characterized by insulin resistance and a relative, rather than absolute, insulin deficiency. These individuals typically do not require insulin for survival, especially at onset, and often throughout their lives. The exact causes are varied and not fully understood, but it is not associated with autoimmune β-cell destruction or other specific causes of diabetes.
Obesity is a major factor in type 2 diabetes, causing insulin resistance. Even non-obese individuals may have increased abdominal fat, contributing to insulin resistance. Ketoacidosis is rare in type 2 diabetes, usually occurring under severe stress like infection. The condition often remains undiagnosed for years due to gradual hyperglycemia development and initially mild symptoms. However, these patients are at increased risk of macrovascular and microvascular complications. While insulin levels may appear normal or elevated, they are insufficient to compensate for insulin resistance, given the elevated blood glucose levels. Insulin secretion is therefore defective. Insulin resistance can improve with weight loss and medication but is rarely fully reversed. The risk of type 2 diabetes increases with age, obesity, physical inactivity, prior GDM, hypertension, and dyslipidemia, and varies across racial and ethnic groups. It has a strong genetic predisposition, more complex and less defined than type 1 autoimmune diabetes.
Other Specific Types of Diabetes
Beyond type 1 and type 2, several other specific types of diabetes exist, including those due to genetic defects, exocrine pancreas diseases, endocrinopathies, drug or chemical induction, infections, uncommon immune-mediated forms, and genetic syndromes.
Genetic Defects of β-cell Function
Several monogenetic defects in β-cell function lead to specific diabetes forms, often with onset before age 25. These are known as maturity-onset diabetes of the young (MODY), characterized by impaired insulin secretion with minimal insulin action defects, inherited in an autosomal dominant pattern. Six genetic loci on different chromosomes are implicated. The most common form involves mutations in the hepatocyte nuclear factor (HNF)-1α on chromosome 12. Another form is linked to glucokinase gene mutations on chromosome 7p, resulting in defective glucokinase, the “glucose sensor” for β-cells. Less common forms involve mutations in transcription factors like HNF-4α, HNF-1β, insulin promoter factor (IPF)-1, and NeuroD1.
Mitochondrial DNA point mutations, notably at position 3,243 in the tRNA leucine gene, are associated with diabetes and deafness. This mutation is also found in MELAS syndrome, but diabetes manifestation varies. Genetic defects preventing proinsulin conversion to insulin and mutant insulin molecules with impaired receptor binding are rare autosomal dominant conditions causing mild glucose intolerance.
Genetic Defects in Insulin Action
Rare genetic abnormalities in insulin action can cause diabetes. Insulin receptor mutations range from hyperinsulinemia with mild hyperglycemia to severe diabetes, sometimes with acanthosis nigricans. In women, virilization and cystic ovaries may occur. Pediatric syndromes like leprechaunism and Rabson-Mendenhall syndrome involve insulin receptor gene mutations, extreme insulin resistance, and severe outcomes. Lipoatrophic diabetes, also insulin-resistant, likely involves postreceptor signal transduction pathway defects.
Diseases of the Exocrine Pancreas
Diffuse pancreatic injury from pancreatitis, trauma, infection, pancreatectomy, or pancreatic carcinoma can cause diabetes. Damage must be extensive, except in pancreatic cancer cases where diabetes may arise from smaller lesions, suggesting other mechanisms. Cystic fibrosis and hemochromatosis, if extensive, can impair insulin secretion. Fibrocalculous pancreatopathy, with abdominal pain and pancreatic calcifications, also contributes to diabetes.
Endocrinopathies
Excess hormones like growth hormone, cortisol, glucagon, and epinephrine can antagonize insulin action, causing diabetes, especially in those with pre-existing insulin secretion defects. Conditions like acromegaly, Cushing’s syndrome, glucagonoma, and pheochromocytoma can induce diabetes, which typically resolves with hormone level correction. Somatostatinoma- and aldosteronoma-induced hypokalemia can also cause diabetes by inhibiting insulin secretion, generally resolving post-tumor removal.
Drug- or Chemical-Induced Diabetes
Many drugs can impair insulin secretion or action, potentially precipitating diabetes, especially in those with insulin resistance. Toxins like Vacor and intravenous pentamidine can permanently damage β-cells. Drugs such as nicotinic acid and glucocorticoids can impair insulin action. Alpha-interferon has been linked to diabetes with islet cell antibodies and insulin deficiency. Table 1 provides a classification of drug-, hormone-, and toxin-induced diabetes.
Table 1. Etiologic classification of diabetes mellitus. This table categorizes the various causes of diabetes mellitus, from type 1 and type 2 to other specific types including genetic defects, pancreatic diseases, endocrinopathies, drug-induced diabetes, infections, and gestational diabetes.
Infections
Certain viruses, such as congenital rubella, coxsackievirus B, cytomegalovirus, adenovirus, and mumps, have been associated with β-cell destruction and diabetes.
Uncommon Forms of Immune-Mediated Diabetes
Rare conditions like stiff-man syndrome, an autoimmune CNS disorder, and anti-insulin receptor antibodies can cause diabetes. Stiff-man syndrome patients often have GAD autoantibodies, with about one-third developing diabetes. Anti-insulin receptor antibodies can block insulin binding or, paradoxically, act as insulin agonists, causing hypoglycemia. These antibodies are sometimes found in systemic lupus erythematosus and other autoimmune diseases, often associated with acanthosis nigricans and termed type B insulin resistance.
Other Genetic Syndromes Sometimes Associated with Diabetes
Several genetic syndromes, including Down syndrome, Klinefelter syndrome, Turner syndrome, Wolfram’s syndrome, and others listed in Table 1, are associated with increased diabetes incidence. Wolfram’s syndrome, for example, is an autosomal recessive disorder featuring insulin-deficient diabetes and other severe conditions.
Gestational Diabetes Mellitus (GDM)
GDM is defined as glucose intolerance first recognized during pregnancy. It may resolve post-delivery but can indicate pre-existing or concurrent glucose intolerance. The rising prevalence of obesity and type 2 diabetes has increased the number of pregnant women with undiagnosed type 2 diabetes.
In 2008–2009, the International Association of Diabetes and Pregnancy Study Groups (IADPSG), including the American Diabetes Association (ADA), recommended diagnosing overt diabetes, not GDM, in high-risk pregnant women at their first prenatal visit using standard diabetes criteria (Table 3). GDM complicates about 7% of pregnancies, with over 200,000 cases annually.
Table 3. Criteria for the diagnosis of diabetes. This table outlines the diagnostic criteria for diabetes mellitus, including A1C levels, fasting plasma glucose (FPG), 2-hour plasma glucose during an OGTT, and random plasma glucose in patients with classic hyperglycemia symptoms.
CATEGORIES OF INCREASED RISK FOR DIABETES
In 1997 and 2003, expert committees recognized an intermediate group with glucose levels above normal but not meeting diabetes criteria, termed pre-diabetes. This includes impaired fasting glucose (IFG) with FPG levels of 100–125 mg/dl (5.6–6.9 mmol/l) and impaired glucose tolerance (IGT) with 2-h OGTT values of 140–199 mg/dl (7.8–11.0 mmol/l).
IFG and IGT indicate a high risk for future diabetes and cardiovascular disease, associated with obesity, dyslipidemia, and hypertension. Lifestyle interventions and certain medications can prevent or delay diabetes onset in IGT individuals, though their impact on mortality and cardiovascular disease is still under study. In 2003, ADA lowered the IFG threshold from 110 to 100 mg/dl to align IFG prevalence with IGT, though WHO and others did not universally adopt this change.
A1C is increasingly used to identify those at risk for diabetes. The International Expert Committee in 2009 noted a diabetes risk continuum across glycemic measures but did not formally define an A1C-equivalent pre-diabetes category. However, they acknowledged increased risk at A1C levels above the lab “normal” range but below the diabetes cutoff (6.0–6.4%). Incidence rates in this A1C range are significantly higher than in the general US population. Studies suggest an A1C of 5.5–6.0% accurately identifies IFG or IGT, and interventions in the Diabetes Prevention Program were effective across A1C levels below and above 5.9%. Thus, an A1C range of 5.5–6% is likely appropriate for initiating preventive measures.
Defining a lower limit for pre-diabetes A1C is somewhat arbitrary due to the continuous risk of diabetes with glycemia levels. A 5.7% A1C cutoff is less sensitive but more specific than a fasting glucose cutoff of 100 mg/dl for diabetes risk identification. Research indicates a 5.7% A1C has 66% sensitivity and 88% specificity for 6-year diabetes incidence prediction. Analyses suggest a 5.7% A1C correlates with diabetes risk similar to high-risk DPP participants. Therefore, an A1C range of 5.7–6.4% reasonably identifies individuals at high risk for future diabetes, termed pre-diabetes.
Individuals with A1C of 5.7–6.4% should be informed of their increased diabetes and cardiovascular disease risk and counseled on risk-reduction strategies like weight loss and physical activity. Risk increases disproportionately with A1C levels, necessitating more intensive interventions and vigilant follow-up for those above 6.0%. However, even those below 5.7% may still be at risk, depending on A1C level and other risk factors.
Table 2. Categories of increased risk for diabetes. This table summarizes the categories defining increased risk for developing diabetes, based on fasting plasma glucose (FPG), 2-hour plasma glucose in the OGTT (IGT), and A1C levels.
Table 2 summarizes categories of increased diabetes risk. Risk evaluation should include a comprehensive assessment for both diabetes and cardiovascular disease, considering patient comorbidities, life expectancy, lifestyle change capacity, and overall health goals.
DIAGNOSTIC CRITERIA FOR DIABETES MELLITUS
Diabetes diagnosis has historically relied on glucose criteria, FPG or 75-g OGTT. In 1997, diagnostic criteria were revised based on the association between FPG and retinopathy. Studies showed glycemic thresholds below which retinopathy prevalence was low and above which it increased linearly. Retinopathy onset deciles were consistent across glucose measures and populations. These analyses informed the new FPG diagnostic cutoff of ≥126 mg/dl (7.0 mmol/l) and confirmed the 2-h PG value of ≥200 mg/dl (11.1 mmol/l).
A1C is a key marker for chronic glycemia, reflecting average blood glucose over 2–3 months. It is crucial in diabetes management, correlating with microvascular and macrovascular complications, and is a standard for glycemic management assessment. Past committees did not recommend A1C for diagnosis due to assay standardization issues. However, with improved standardization, an International Expert Committee in 2009 recommended A1C for diabetes diagnosis at a threshold of ≥6.5%, a decision affirmed by ADA. This A1C cutoff, like FPG and 2-h PG cutoffs, aligns with retinopathy prevalence inflection points. Diagnostic A1C tests must be NGSP-certified and standardized to the DCCT reference assay. Point-of-care A1C assays are currently not sufficiently accurate for diagnosis.
Using A1C for diagnosis offers advantages over FPG, including convenience (no fasting), greater preanalytical stability, and less variability during stress or illness. However, it is more expensive, less available in developing regions, and has incomplete correlation with average glucose in some individuals. A1C can also be misleading in certain anemias and hemoglobinopathies. For hemoglobinopathies without abnormal red cell turnover, assays without hemoglobin interference should be used. In conditions with abnormal red cell turnover, glucose criteria must be used exclusively.
Established glucose criteria (FPG, 2-h PG) remain valid. Random plasma glucose ≥200 mg/dl (11.1 mmol/l) in patients with severe hyperglycemia symptoms or hyperglycemic crisis also confirms diagnosis. While A1C would likely be measured and be above diagnostic cutoffs in such cases, in rapidly developing diabetes, A1C may not be elevated despite frank diabetes.
Concordance between A1C and glucose-based tests is not complete. NHANES data suggest that A1C ≥6.5% identifies fewer undiagnosed diabetes cases than FPG ≥126 mg/dl (7.0 mmol/l). However, A1C’s greater practicality might increase overall diagnoses due to wider application.
Further research is needed to understand discrepancies between A1C and glucose tests. Discordance may arise from measurement variability, time changes, or different physiological processes measured by each test. Elevated A1C with non-diabetic FPG might indicate higher postprandial glucose or glycation rates. High FPG with non-diabetic A1C could suggest increased hepatic glucose production or reduced glycation rates.
Diabetes diagnosis should be confirmed with repeat testing, ideally using the same test, unless clinically evident (e.g., hyperglycemic crisis). If different tests are used and both are above diagnostic thresholds, diabetes is confirmed. In discordant cases, the test above the threshold should be repeated, and diagnosis is based on the confirmed test. If a repeat test falls below the diagnostic cutoff, close patient monitoring and repeat testing in 3–6 months may be appropriate, especially if initial values were near diagnostic thresholds.
Healthcare professionals should decide which test to use, considering test availability and practicality. Ensuring diabetes testing is performed when indicated is crucial. Many at-risk patients still lack adequate testing and counseling. Table 3 summarizes current diagnostic criteria for diabetes.
Diagnosis of GDM
GDM diagnostic criteria at the time of this publication were those of Carpenter and Coustan. The Fourth International Workshop-Conference on Gestational Diabetes Mellitus in 1997 supported Carpenter/Coustan criteria and the 75-g 2-h OGTT.
Testing for Gestational Diabetes
Previous recommendations included universal GDM screening. However, low-risk women meeting all criteria below may not require screening:
- Age <25 years
- Normal body weight
- No family history of diabetes (first-degree relative)
- No history of abnormal glucose metabolism
- No history of poor obstetric outcome
- Not from a high-risk ethnic/racial group (Hispanic American, Native American, Asian American, African American, Pacific Islander)
Risk assessment for GDM should occur at the first prenatal visit. High-risk women (marked obesity, prior GDM, glycosuria, strong family history) should be tested immediately. If initial tests are negative, retesting should occur at 24–28 weeks gestation. Average-risk women should be tested at 24–28 weeks.
FPG >126 mg/dl (7.0 mmol/l) or random plasma glucose >200 mg/dl (11.1 mmol/l) meets diabetes diagnostic threshold and requires confirmation. Below these levels, GDM evaluation follows a one-step or two-step approach.
One-Step Approach
Diagnostic OGTT without prior screening, potentially cost-effective in high-risk groups.
Two-Step Approach
Initial 50-g glucose challenge test (GCT). Diagnostic OGTT for women exceeding a glucose threshold on GCT. A threshold >140 mg/dl (7.8 mmol/l) identifies ~80% of GDM cases, increased to 90% at >130 mg/dl (7.2 mmol/l).
GDM diagnosis via either approach relies on OGTT. 100-g OGTT criteria are derived from O’Sullivan and Mahan, modified by Carpenter and Coustan (Table 4, top). Alternatively, 75-g glucose load criteria are used for fasting, 1-h, and 2-h values (Table 4, bottom), though less validated than 100-g OGTT.
Table 4. Diagnosis of GDM with a 100-g or 75-g glucose load. Presents the glucose concentration thresholds in mg/dl and mmol/l for diagnosing Gestational Diabetes Mellitus (GDM) using either a 100-g or 75-g glucose load during an Oral Glucose Tolerance Test (OGTT).
For a positive GDM diagnosis, two or more venous plasma concentrations must meet or exceed the thresholds. Testing should be in the morning after an 8–14 hour overnight fast, following at least 3 days of unrestricted diet (≥150 g carbohydrate/day) and activity. Subjects should remain seated and not smoke during the test.
The Hyperglycemia and Adverse Pregnancy Outcomes (HAPO) study demonstrated continuous risk increase of adverse outcomes with maternal glycemia at 24–28 weeks, even within previously normal ranges. IADPSG recommended 75-g OGTT for all women at 24–28 weeks without prior diabetes, establishing diagnostic cut points for fasting, 1-h, and 2-h plasma glucose linked to increased adverse outcome odds.
At the time of this publication, ADA was considering adopting IADPSG diagnostic criteria in collaboration with obstetrical organizations, recognizing that increased GDM prevalence may result, but evidence suggests treating even mild GDM reduces maternal and infant morbidity.
Acknowledgments
The American Diabetes Association acknowledged the writing group members for updating sections on diagnosis and risk categories.
