Epidemiology, presentation, and diagnosis of type 2 diabetes mellitus in children and adolescents

INTRODUCTION — The incidence of type 2 diabetes mellitus (T2DM) in youth has increased in many countries since the early 1990s and is linked to the rise in childhood obesity. T2DM and its comorbidities are risk factors for vascular disease later in life and premature mortality. As a result, it is imperative for clinicians to identify and treat children and adolescents with this disease.

The epidemiology, presentation, and diagnosis of T2DM in children and adolescents are presented here. Other aspects of T2DM in youth are discussed in separate topic reviews:

●(See "Prevention of type 2 diabetes mellitus".)

●(See "Management of type 2 diabetes mellitus in children and adolescents".)

●(See "Chronic complications and screening in children and adolescents with type 2 diabetes mellitus".)

PATHOGENESIS — T2DM is characterized by hyperglycemia, insulin resistance, and also relative impairment in insulin secretion. The presence of insulin resistance explains the strong clinical association of T2DM with obesity and other insulin-resistant states (see 'Risk factors' below). Impaired insulin secretion has been demonstrated in a variety of studies, which show that patients usually have lost approximately 80 percent of their pancreatic beta cell function before the diagnosis of T2DM [1,2]. In most cases, the impaired insulin secretion does not appear to be mediated by antibodies against the pancreatic islet cells. (See "Pathogenesis of type 2 diabetes mellitus".)

By contrast, type 1 diabetes mellitus (T1DM) is primarily characterized by absolute insulin deficiency, usually caused by T cell-mediated immune destruction of pancreatic beta cells. Insulin resistance may be present to a variable degree. (See "Pathogenesis of type 1 diabetes mellitus".)

Despite these apparent differences, there is considerable clinical overlap between T2DM and T1DM regarding both insulin resistance and presence of pancreatic autoantibodies. There may be considerable clinical overlap between T2DM and T1DM, and, at the time of presentation, it may be difficult to distinguish between them. Insulin resistance can be present in T1DM. The presence of pancreatic autoantibodies, however, is considered diagnostic of T1DM. (See 'Type 2 versus type 1 diabetes' below and "Classification of diabetes mellitus and genetic diabetic syndromes".)

EPIDEMIOLOGY

Incidence and trends

●Diabetes – The incidence of T2DM among youth is increasing in many countries, coinciding with increasing prevalence of obesity and severe obesity. As an example, in the United States, there was a sharp rise in T2DM from 9.0 cases per 100,000 in 2002-2003 to 13.8 cases per 100,000 in 2014-2015, with an adjusted annual increase of T2DM of 4.8 percent (figure 1), based on a large representative dataset from the SEARCH for Diabetes in Youth study [3]. Similarly, the proportion of new cases of diabetes in youth classified as T2DM was negligible before 1980s, then rose substantially in Japan [4], Thailand [5], Argentina [6], and other countries [7].

The increasing trends in T2DM among youth appear to lag, as expected, behind the increase in obesity among adolescents. One explanation is that there is a latency period between the onset of obesity and the related risk for T2DM [8]. Consistent with this model, studies have shown sharp increases in T2DM among young adults [9]. Thus, the rates of obesity among adolescents may cause a modest increase in T2DM among adolescents and a more substantial increase in T2DM when the cohort reaches adulthood. Furthermore, T2DM may be most closely correlated with severe obesity, which continues to rise among adolescents in the United States [10]. Finally, changes in demographics also may contribute to the increased prevalence of T2DM in the United States, with increases in racial and ethnic groups that experience higher rates of T2DM and/or ongoing increases in the prevalence or severity of obesity in some segments of the population. (See "Definition, epidemiology, and etiology of obesity in children and adolescents", section on 'Epidemiology'.)

The burden of T2DM varies markedly among racial/ethnic groups (figure 2) [11]. As an example, the incidence of T2DM in United States children ages 10 to 19 years in 2014-2015 was [3]:

•Non-Hispanic White youth – 4.5 per 100,000

•Asian/Pacific Islander youth – 11.9 per 100,000

•Hispanic youth – 20.9 per 100,000

•Native American youth – 32.8 per 100,000

•Non-Hispanic Black youth – 37.8 per 100,000

There were significant increases in incidence from 2002-2003 through 2014-2015 among all groups except for non-Hispanic White youth, with the greatest adjusted annual increases among Asian/Pacific Islander youth (7.7 percent) and Hispanic youth (6.5 percent) [3]. (See 'Risk factors' below.)

●Prediabetes – In a nationally representative sample in the United States, the overall prevalence of prediabetes in youth 12 to 18 years of age was 18 percent; the prevalence was significantly higher (26 percent) among those with obesity and in males versus females [12]. This study defined prediabetes as an abnormality in fasting plasma glucose (FPG), hemoglobin A1c (A1C), and/or glucose tolerance. The most common abnormality was impaired fasting glucose, present in 11 percent of samples. There was little overlap among the subgroups with abnormal A1C, FPG, or glucose tolerance, indicating that prevalence estimates will vary substantially depending on which of these tests is used. (See 'Prediabetes' below.)

Risk factors

●Obesity – Obesity and excess adipose tissue are the most important risk factors for T2DM. Body mass index (BMI) percentiles are used to define overweight and obesity in children and adolescents (table 1). The association between obesity and T2DM is even stronger in youth than in adults [13]. As an example, in a registry study from the United States, nearly 80 percent of youth with T2DM had obesity and an additional 10 percent were overweight [14]. In addition, BMI during adolescence predicts the risk for T2DM during adulthood [15]. (See "Overview of the health consequences of obesity in children and adolescents", section on 'Type 2 diabetes mellitus'.)

The risk for T2DM is associated with an abdominal fat distribution (also called central adiposity, visceral, android, or male-type obesity) [16,17]. Obesity predisposes to T2DM by increasing peripheral resistance to insulin-mediated glucose uptake. (See "Type 2 diabetes mellitus: Prevalence and risk factors", section on 'Obesity' and "Pathogenesis of type 2 diabetes mellitus", section on 'Role of diet, obesity, and inflammation'.)

Glucose tolerance appears to respond to changes in health behaviors. In one study, a multicomponent intervention aimed at reducing obesity and risk for T2DM in middle school students improved fasting insulin levels (suggesting greater insulin sensitivity) despite minimal effect on overall weight status in the intervention group compared with the control group, although BMI Z-scores were reduced [18]. These benefits were evident in those youth with overweight or obesity at study entry.

●Genetic susceptibility – T2DM is caused by a complex interaction of environmental and genetic factors in susceptible people, based on studies in adults. In the majority of patients with T2DM, genetic susceptibility appears to be due to the expression of multiple genes (ie, the risk is a polygenic trait). (See "Pathogenesis of type 2 diabetes mellitus", section on 'Genetic susceptibility'.)

Evidence for a strong genetic component for T2DM includes the increased risk of T2DM in close relatives of an affected patient:

•Among monozygotic twins in which one twin has T2DM, the other twin has a 90 percent chance of developing the disease [19].

•For children or adolescents with T2DM, one-half to three-quarters have at least one affected parent [6,20,21].

•For parents with T2DM, children have an estimated 20 percent risk of developing T2DM by late adulthood (approximately 3.5 times the risk of an individual without parental T2DM) [22]. The risk rises to 30 percent if both parents are affected (approximately six times the risk of an individual with unaffected parents). These figures come from a longitudinal study in a predominantly White population; it is possible that familial risk varies between different racial and ethnic groups.

Identification of the relevant genes has been difficult because of the complex inheritance patterns and interactions with the environment [23]. Investigations of genetic susceptibility are ongoing, some of which specifically study patients with early onset of T2DM (under 18 years old). Candidate genes that have been linked to T2DM are discussed separately. (See "Pathogenesis of type 2 diabetes mellitus", section on 'Genetic susceptibility'.)

●Ethnicity – In the United States, T2DM is more common in Native American, African American, Hispanic, Asian American, and Pacific Islander youth than in the general population (figure 2) [11,20,24-33].

•Native American youth – The overall prevalence of T2DM among Native American youth (15 to 19 years) is 4.5 per 1000 [20], with substantial variation among groups:

-Pima Indian – 51 per 1000 [20]

-Canadian Cree and Ojibway Indians in Manitoba – 2.3 per 1000 [20]

-Navajo – 2.2 per 1000 [34]

The particularly high prevalence among Pima Indian youth may be partly attributable to systematic population screening for this group [20]. The incidence of T2DM among Native American youth is approximately seven times that of non-Hispanic White youth [3].

•Black non-Hispanic youth – Several studies have demonstrated that T2DM disproportionately affects Black non-Hispanic individuals [11,26]. In a registry study in the United States, the prevalence of T2DM among Black non-Hispanic youth was 1.06 per 1000 individuals, five times the rate among White non-Hispanic youth [11]. Similarly, several studies suggest that insulin resistance is greater in Black children compared with White children [35-39].

•Hispanic youth – Hispanic children also are at increased risk for T2DM compared with the general population. In a registry study in the United States, the prevalence of T2DM among Hispanic youth was 0.79 per 1000 individuals, four times that of White non-Hispanic youth [11]. Hispanic children have been reported to have lower insulin sensitivity [40,41] and reduced pancreatic beta cell function (decreased insulin secretion) [42], both of which increase the risk of diabetes.

●Age and pubertal status – Many patients with pediatric T2DM present at the onset of or during puberty, a stage of development when there is physiologic insulin resistance [43-45]. Accordingly, virtually all patients recruited for a large trial (TODAY study) were pubertal or postpubertal at study entry [21]. Approximately 40 percent of pediatric cases occur between 10 and 14 years of age, and the remaining 60 percent between 15 and 19 years [46]. During puberty, insulin sensitivity decreases by approximately 30 percent [47], related to the increased activity of growth hormone.

●Sex – Females are 1.2 to 1.7 times more likely than males to progress from prediabetes and develop T2DM during adolescence [3,11,20,46,48,49]. Although the reason for this increased risk in females is probably multifactorial, it may be related to an increased risk of insulin resistance, as seen in adolescents with polycystic ovary syndrome (PCOS), and could be related to the increase in adiposity versus lean body mass observed in female compared with male adolescents, along with less physical activity in females versus males in late adolescence (albeit based on a study in adolescents with type 1 diabetes mellitus [T1DM]) [50]. (See "Etiology and pathophysiology of polycystic ovary syndrome in adolescents", section on 'Insulin-resistant hyperinsulinism' and "Treatment of polycystic ovary syndrome in adolescents", section on 'Treatment of obesity and insulin resistance'.)

●Prenatal exposures – One hypothesis suggests that prenatal exposure to maternal undernutrition or gestational diabetes causes metabolic and hormonal changes in the offspring that promote obesity and insulin resistance and increase the risk for T2DM risk in adulthood. This phenomenon has been termed "metabolic programming."

•Small for gestational age – Low birth weight and/or length for gestational age (a sign of intrauterine undernutrition) is associated with insulin resistance. The combination of low birth weight and weight gain in adult middle age increases insulin resistance and the risk for T2DM (figure 3) [51]. Individuals with the lowest birth weight and the highest prepubertal body weight appear to be at the greatest risk for insulin resistance and T2DM [52-54]. (See "Pathogenesis of type 2 diabetes mellitus", section on 'Role of intrauterine development' and "Infants with fetal (intrauterine) growth restriction".)

•Gestational diabetes – The abnormal intrauterine metabolic environment of a pregnancy complicated by diabetes appears to increase the risk of T2DM. Intrauterine exposure to hyperglycemia and hyperinsulinemia may affect the development of adipose tissue and pancreatic beta cells, leading to future obesity and altered glucose metabolism. In the SEARCH for Diabetes in Youth study, gestational or pre-gestational diabetes and intrauterine exposure to maternal obesity were independently associated with T2DM in adolescents, and both risk factors were present in 47 percent of the SEARCH cohort [55]. The effects of gestational diabetes on the risk of developing T2DM in the offspring are discussed separately. (See "Infants of mothers with diabetes (IMD)", section on 'Diabetes'.)

●PCOS – Insulin resistance is a component of PCOS and may play a role in its pathogenesis. Patients with PCOS are at increased risk for developing T2DM. (See "Diagnostic evaluation of polycystic ovary syndrome in adolescents", section on 'Additional evaluation of PCOS patients' and "Etiology and pathophysiology of polycystic ovary syndrome in adolescents", section on 'Insulin-resistant hyperinsulinism'.)

CLINICAL PRESENTATION — The presentation of T2DM in children and adolescents is variable, ranging from incidental detection to symptoms of hyperglycemia, diabetic ketoacidosis (DKA), or hyperosmolar hyperglycemia [56].

Asymptomatic — Approximately 40 percent of children and adolescents with T2DM are identified while asymptomatic [26,57]. These patients may be tested for T2DM because of risk factors or because glycosuria was detected on a urinalysis obtained as part of a routine physical examination [48]. (See 'Screening' below.)

Common symptoms — Approximately 60 percent of children and adolescents with T2DM are symptomatic at presentation [26,57]. The main symptoms are due to hyperglycemia and commonly include polyuria, polydipsia, and nocturia, similar to those in patients with type 1 diabetes mellitus (T1DM). Symptomatic patients also may have experienced recent weight loss, although this is not a common presenting complaint, and the percentage of weight loss is typically less than in patients with T1DM [48,49].

Occasionally, the presenting complaint may be vaginal discharge or Candida vulvovaginitis in females (typical symptoms are pruritus or burning) or tinea cruris in adolescent males [48,58]. (See "Candida vulvovaginitis: Clinical manifestations and diagnosis", section on 'Risk factors'.)

Diabetic ketoacidosis — Occasionally, children with T2DM present with DKA (hyperglycemia, ketonuria, and acidosis). Symptoms are due to hyperglycemia and include polyuria (due to the glucose osmotic diuresis), polydipsia (due to increased urinary losses), fatigue, and lethargy. Children with DKA require hospitalization, rehydration, and insulin replacement therapy. (See "Diabetic ketoacidosis in children: Clinical features and diagnosis", section on 'Signs and symptoms' and "Diabetic ketoacidosis in children: Treatment and complications".)

The reported frequency of DKA as the initial presentation for childhood T2DM varies from 5 to 12 percent depending on the population studied [26,59-61], with further increases during the coronavirus disease 2019 (COVID-19) pandemic [62-64]. Presentation with DKA appears to be most common in Black or African American youth, among whom up to 25 percent present with DKA [49,65]. In a case series that included 28 non-Hispanic Black patients with T2DM, 25 percent presented with DKA and an additional 43 percent presented with ketonuria but without acidosis [65]. In contrast, none of the 12 White patients in the same study had ketonuria or DKA. In an observational study of 21 Mexican American children with newly diagnosed T2DM, approximately one-third had ketonuria [66].

Nonetheless, most patients who first present with DKA have T1DM rather than T2DM: among more than 1500 youth presenting with DKA, approximately 95 percent had T1DM and 5 percent had T2DM [67]. (See "Diabetic ketoacidosis in children: Clinical features and diagnosis", section on 'Epidemiology' and 'Type 2 versus type 1 diabetes' below.)

Hyperosmolar hyperglycemic state — Rarely, adolescents with T2DM may present with hyperosmolar hyperglycemic state (HHS), also referred to as hyperosmolar hyperglycemic nonketotic syndrome (HHNK) [68-72]. This condition is a medical emergency; it is characterized by [68]:

●Marked hyperglycemia (plasma glucose >600 mg/dL)

●Hyperosmolality (serum osmolality >330 mOsm/kg)

●Severe dehydration

●Little or no ketonuria

Initially described in adult patients with poorly controlled T2DM, the condition has also been reported in adolescents, most of whom were African American with newly diagnosed T2DM [69-74]. In one report, HHS was the presenting feature of T2DM in approximately 2 percent of adolescents [59]. Recognition of HHS is important because it is characterized by more severe dehydration than typical DKA and has high morbidity and mortality if not adequately treated. Management of HHS is discussed in a separate topic review. (See "Diabetic ketoacidosis and hyperosmolar hyperglycemic state in adults: Treatment", section on 'Fluid replacement'.)

SCREENING

●Indications – For asymptomatic children and adolescents we use the following criteria for initial screening for T2DM (table 2), consistent with recommendations from the American Diabetes Association and International Society for Pediatric and Adolescent Diabetes [75-77]:

•Overweight or obesity (body mass index [BMI] ≥85^th percentile) and

•One or more of the following additional risk factors:

-T2DM in a first- or second-degree relative

-Member of a high-risk racial/ethnic group: Native American, African American, Latino, Asian/Pacific Islander, Canadian First Nations, Australian indigenous

-History of maternal diabetes or gestational diabetes during the child's gestation

-Signs of insulin resistance or conditions associated with insulin resistance (eg, acanthosis nigricans, hypertension, dyslipidemia, polycystic ovary syndrome [PCOS], or small for gestational age birth weight and/or length)

-Current use of weight-promoting atypical antipsychotic drug [77]

●Method – Screening should begin at the onset of puberty or ≥10 years (whichever is earlier). We use hemoglobin A1c (A1C) and/or fasting plasma glucose (FPG) to screen for T2DM in asymptomatic adolescents. Patients with abnormal results require further evaluation. (See 'Asymptomatic with abnormal screening results' below.)

We repeat the screening approximately annually in patients with marked obesity or strong evidence of insulin resistance (eg, acanthosis nigricans), or in patients with borderline results. This is consistent with the American Diabetes Association recommendations to repeat screening at a minimum of every three years or more frequently if BMI is increasing [78].

●Evidence – In adolescents, targeted testing of risk groups has a low yield for identifying T2DM but a high yield for identifying patients with prediabetes. As an example, one study that tested a population of 8^th grade students in the United States found T2DM in 0.5 percent, while 43 percent met criteria for prediabetes [41]. Most of the subjects had multiple risk factors for T2DM, including high-risk race/ethnicity. Estimates of prediabetes prevalence vary substantially among studies, depending on the risk factors in the population and the tests used for the evaluation.

No studies in children have examined whether early diagnosis improves the long-term outcome of T2DM. However, indirect evidence suggests that identifying individuals with prediabetes may be valuable because studies in adults indicate that lifestyle intervention can prevent or delay the onset of T2DM [79]. For these reasons, targeted testing of adolescents with multiple risk factors for T2DM is recommended, as outlined below.

EVALUATION

Asymptomatic with abnormal screening results — For asymptomatic children with abnormal screening results, the next step depends upon the level of concern:

●For those with laboratory abnormalities in the prediabetes range (table 3), it is reasonable to repeat the screening test in three to six months. If there is a high clinical suspicion of diabetes (eg, severe obesity and/or multiple risk factors for diabetes (table 2)), repeat the screening test (eg, hemoglobin A1c [A1C]) sooner and/or proceed to an oral glucose tolerance test (OGTT).

●For those with more severe abnormalities (eg, A1C ≥6.5 percent), repeat the A1C and/or at least one additional laboratory test for diabetes and perform a urinalysis. (See 'Laboratory tests' below.)

Symptomatic — All patients with symptoms suggesting possible diabetes (polydipsia, polyuria, blurred vision, unexplained weight loss) should be tested for diabetes, regardless of risk factors. (See 'Clinical presentation' above.)

The usual first step for symptomatic patients is to measure random plasma glucose, A1C, and urinalysis to evaluate for glucosuria and/or ketonuria. (See 'Laboratory tests' below and 'Diagnosis' below.)

Laboratory tests — Testing for diabetes can be done by measuring A1C or fasting plasma glucose (FPG) or by performing an OGTT [75]. In our practice, we measure either FPG or A1C and confirm abnormal results by either repeating the initial test on another day or by performing a different test. A1C should be included at least once to provide a baseline and is often repeated for confirmation of the diagnosis in patients without symptoms.

The characteristics of each of these laboratory tests are outlined below, and diagnostic criteria for prediabetes and diabetes are summarized in the tables (table 3 and table 4). (See 'Diagnosis' below.)

●Random (nonfasting) plasma glucose — The primary indication for a random plasma glucose is for urgent evaluation of a patient with symptoms suggesting hyperglycemia (eg, polydipsia, polyuria, blurred vision, unexplained weight loss). Results are interpreted as follows:

•Normal – Random plasma glucose <140 mg/dL (<7.8 mmol/L) suggests that the current symptoms may not be related to hyperglycemia but does not entirely exclude the possibility of diabetes.

•Moderately elevated – Random plasma glucose 140 to 199 mg/dL (7.8 to 11 mmol/L) is concerning for impaired glucose tolerance and should be followed by the other tests below for more accurate evaluation of glycemia.

•Elevated – Random plasma glucose ≥200 mg/dL (11.1 mmol/L) in a patient with classic symptoms of hyperglycemia is diagnostic of diabetes.

These glucose measurements should be obtained in a clinical laboratory and not based solely on a point-of-care glucose meter.

●A1C – A1C (glycated hemoglobin) is accepted as a screening test for diabetes mellitus. It is useful in clinical practice because it does not require fasting, but its performance as a screening test has not been fully validated in adolescents. Diagnostic criteria using A1C for adult and pediatric patients are [77,78,80]:

•Diabetes – A1C level ≥6.5 percent (48 mmol/mol), on two occasions

•Prediabetes – A1C level between 5.7 and 6.4 percent

Use of A1C as a screening test was supported by an international consensus statement and endorsed by the American Diabetes Association, provided that the A1C assay is certified by the National Glycohemoglobin Standardization Program and performed in a certified laboratory and not by using a point-of-care device [78,81]. Although some point-of-care A1C devices are certified for glycemic monitoring, most are not appropriate for initial screening and diagnosis of diabetes. Although traditionally reported as a percent, A1C is increasingly reported as mmol/mol (calculator 1). (See "Clinical presentation, diagnosis, and initial evaluation of diabetes mellitus in adults", section on 'A1C'.)

Although A1C may be used for screening in adolescents, the results should be interpreted with some caution because A1C has not been well validated as a screening test in this population [75]. Almost all of the data on A1C as a screening test are based on studies in adults. One study in a population-based sample of adolescents concluded that a single measurement of A1C ≥6.5 percent had modest specificity (75 to 76 percent) and sensitivity (86 to 91 percent) for detecting undiagnosed diabetes mellitus [82]. These findings should be interpreted with caution because they are based on only five cases of undiagnosed T2DM in a sample of 14,119 youth. An earlier study with similar design estimated specificity and sensitivity of 99 and 75 percent, respectively [83]. These observations suggest that an elevated A1C (on two occasions, using a certified laboratory assay) can be used to diagnose diabetes, but a normal A1C does not exclude T2DM.

●FPG – Compared with the OGTT, FPG is more convenient, less expensive, and less invasive but also less sensitive [41]. Diagnostic criteria using FPG for adult and pediatric patients are [78]:

•Diabetes – FPG ≥126 mg/dL (7.0 mmol/L), if confirmed by another test on a different day (or if classic symptoms of diabetes are present)

•Prediabetes – FPG ≥100 mg/dL to 125 mg/dL (≥5.6 mmol/L to 6.9 mmol/L)

These glucose measurements should be performed in a clinical laboratory and not based solely on a point-of-care glucose meter.

Children with FPG values in the upper one-half of the normal range (between 86 and 99 mg/dL) have 2.1 times the risk for developing diabetes during adulthood and 3.4 times the risk for developing prediabetes, independent of the child's weight status [84].

●OGTT – The OGTT is helpful when there is a need for more accurate evaluation of dysglycemia, such as when the clinical suspicion for diabetes remains high despite an FPG or A1C that is nondiagnostic.

The standard glucose load used for the OGTT in children is 1.75 g/kg, up to a maximum dose of 75 g. Results are interpreted as follows [78]:

•Diabetes – Plasma glucose ≥200 mg/dL (11.1 mmol/L) measured two hours after the glucose load

•Prediabetes – Plasma glucose ≥140 to 199 mg/dL (≥7.8 to 11.0 mmol/L) measured two hours after the glucose load

The OGTT is a more sensitive test than FPG because OGTT detects patients with diabetes early in the development of their disease when the FPG may not be elevated [85]. Evidence from adult patients with diabetes shows that approximately 30 percent of individuals will have a nondiabetic fasting glucose concentration when their OGTT is diagnostic for diabetes [86].

There are several limitations to the use of the OGTT in children and adolescents. First, it is unknown whether the diagnostic cut points used for adults are applicable to children. Furthermore, the available evidence to support the glucose dose used for children (1.75 g/kg; maximum dose 75 g) is not strong and its validity has been challenged [23]. Finally, there is some evidence that the OGTT may be less reliable in children than in adults. One study has shown poor reproducibility of results in 60 obese children who had two OGTTs performed 1 to 25 days apart [87]. Performing an OGTT in patients currently treated with metformin may mask the diagnosis of diabetes, suggesting the need for a one- to two-week washout period off of metformin before performing the OGTT [88].

●Tests for ketoacidosis – Evaluate for diabetic ketoacidosis (DKA) as soon as diabetes is confirmed or strongly suspected, because 5 to 10 percent of adolescents with T2DM present with DKA. If the patient is asymptomatic, measuring serum bicarbonate and urinary ketones is sufficient. If the results are abnormal or if the patient is symptomatic (polydipsia, polyuria, or signs of dehydration), measure venous pH and serum beta-hydroxybutyrate (or urinary ketones). (See 'Diabetic ketoacidosis' above.)

●Tests to distinguish type 1 diabetes mellitus (T1DM) from T2DM

•Pancreatic autoantibodies – As soon as diabetes is confirmed, test all patients with a panel of pancreatic (islet) autoantibodies even if the patient has a classic presentation of T2DM [76,77]. The panel should include anti-glutamic acid decarboxylase (GAD) and tyrosine phosphatase insulinoma-associated antigen 2 (IA2) antibodies. Insulin autoantibodies (IAA) should be included provided that insulin replacement therapy has not yet been initiated or has been used for less than two weeks. Testing for a beta cell-specific autoantibody to zinc transporter 8 (ZnT8) is also valuable and is becoming more widely available [89]. (See 'Type 2 versus type 1 diabetes' below.)

•C-peptide – If the diabetes type remains uncertain, a C-peptide level (and/or a serum insulin level if insulin therapy has not yet been initiated) obtained after glycemic control has been established may also provide clinically useful information. A low C-peptide level supports a diagnosis of T1DM but should be interpreted with caution because this can also be low at the time of diagnosis of T2DM [90].

DIAGNOSIS

Prediabetes — Prediabetes refers to a state of increased risk for future development of diabetes, and is typically defined by any of the following (table 3) [75]:

●Hemoglobin A1c (A1C) 5.7 to 6.4 percent (39 to 47 mmol/mol)

●Fasting plasma glucose (FPG) 100 to 125 mg/dL (5.6 to 6.9 mmol/L)

●Oral glucose tolerance test (OGTT) with plasma glucose 140 to 199 mg/dL (7.8 to 11.0 mmol/L) two hours after a glucose load

For adolescents with A1C or FPG results consistent with prediabetes, or with discordant results for either of these tests, we suggest further testing with an OGTT, especially in patients with multiple risk factors for T2DM, because OGTT is more sensitive than A1C and FPG. Subsequent management is discussed below. (See 'Subsequent management' below.)

Diabetes mellitus — The first step is to diagnose diabetes; the second step is to differentiate T2DM from other types of diabetes. (See 'Type 2 versus type 1 diabetes' below and 'Other causes of diabetes' below.)

Diabetes mellitus is diagnosed based upon any one of following (table 4) [77,78]: (See 'Laboratory tests' above.)

●A1C ≥6.5 percent (48 mmol/mol)

●FPG ≥126 mg/dL (7 mmol/L)

●Random plasma glucose ≥200 mg/dL (11.1 mmol/L) in a patient with classic symptoms of hyperglycemia

●OGTT with plasma glucose ≥200 mg/dL (11.1 mmol/L) two hours after a glucose load

Unless unequivocal symptomatic hyperglycemia is present, the diagnosis should be confirmed by repeat testing on a different day. Hyperglycemia that is associated with acute stress (eg, acute infection or surgery) is not diagnostic of diabetes; it should be treated but may be transient. These diagnostic criteria are the same as those used in adults. (See "Clinical presentation, diagnosis, and initial evaluation of diabetes mellitus in adults".)

Glucosuria is suggestive of diabetes but not diagnostic. For example, patients with renal glucosuria or Fanconi syndrome present with glycosuria but will be normoglycemic.

Severe hyperglycemia (plasma glucose ≥600 mg/dL [33 mmol/L]) suggests the possibility of hyperosmolar hyperglycemic state (HHS), which is a medical emergency. (See 'Hyperosmolar hyperglycemic state' above.)

DIFFERENTIAL DIAGNOSIS

Type 2 versus type 1 diabetes — It is important to distinguish between T2DM versus type 1 diabetes mellitus (T1DM), because management differs. T2DM is characterized by hyperglycemia, insulin resistance, and relative impairment in insulin secretion. By contrast, the primary characteristic of T1DM is absolute insulin deficiency, which usually is caused by immune-mediated destruction of pancreatic beta cells, as indicated by the presence of autoantibodies. Insulin resistance can also be present in T1DM [91]. At the time of presentation, it may be difficult to distinguish between T1DM and T2DM. The presence of pancreatic autoantibodies, however, is considered diagnostic of T1DM. (See 'Pathogenesis' above.)

No set of criteria or diagnostic test can consistently distinguish between T1DM and T2DM. Therefore, differentiating between the two types is based upon a combination of the clinical presentation and history, often supported by laboratory studies (table 5) (see "Epidemiology, presentation, and diagnosis of type 1 diabetes mellitus in children and adolescents"):

●Clinical characteristics:

•Body habitus – Patients with T2DM are usually obese (body mass index [BMI] ≥95^th percentile for age and sex). In contrast, children with T1DM are often not overweight and usually have a recent history of weight loss, although 25 percent or more may have overweight or obesity (BMI ≥85^th) at diagnosis [14,92,93].

•Age – Youth with T2DM generally present after the onset of puberty, and almost all present after 10 years of age (figure 4) [46]. By contrast, approximately 50 percent of youth with T1DM present prior to 10 years of age [94]. (See 'Risk factors' above.)

•Insulin resistance – Patients with T2DM usually have clinical features associated with insulin resistance such as acanthosis nigricans, hypertension, dyslipidemia, and polycystic ovary syndrome (PCOS), which are not commonly seen in children with T1DM. As an example, in studies in the United States, 50 to 90 percent of youth diagnosed with T2DM have acanthosis nigricans [20,91]. Among those clinically diagnosed with T1DM, up to 25 percent have biochemical evidence of insulin resistance and approximately 12 percent have acanthosis nigricans [91].

•Family history – 75 to 90 percent of those with T2DM have an affected first- or second-degree relative, whereas only up to 10 percent of patients with T1DM have an affected first- or second-degree relative. (See 'Risk factors' above.)

•Ethnicity – In the United States, most pediatric patients with T2DM belong to certain racial and ethnic groups including non-Hispanic Black, Hispanic, Native American, Asian American, and Pacific Islander youth, although either type of diabetes can occur in any of these racial or ethnic groups. (See 'Risk factors' above.)

●Laboratory results (see 'Laboratory tests' above):

•Ketoacidosis – The presence of diabetic ketoacidosis (DKA) supports a diagnosis of T1DM, but does not exclude T2DM. DKA is present at diagnosis in approximately 30 percent of children with T1DM, compared with 5 to 10 percent of those with T2DM. (See 'Diabetic ketoacidosis' above and "Diabetic ketoacidosis in children: Clinical features and diagnosis".)

•Pancreatic autoantibodies – The presence of pancreatic autoantibodies (glutamic acid decarboxylase [GAD], tyrosine phosphatase insulinoma-associated antigen 2 [IA2], insulin autoantibodies [IAA], or zinc transporter antibodies [ZnT8]) supports a diagnosis of T1DM. However, up to 10 percent of adolescents with T2DM diagnosed on the basis of clinical phenotype have evidence of beta cell autoimmunity [90]. Thus, the pathophysiologic features of both types of diabetes may coexist in the same patient, particularly if the patient has overweight or obesity.

•Insulin and C-peptide levels – A C-peptide level (and/or a serum insulin level if insulin therapy has not yet been initiated) obtained after glycemic control has been established may also provide clinically useful information. Low C-peptide levels support a diagnosis of T1DM but should be interpreted with caution because they can also be low at the time of diagnosis of T2DM. Fasting serum insulin and C-peptide levels have not been standardized to distinguish between type 1 and 2 diabetes [95,96].

Some patients may have mixed features and are difficult to classify. As an example, some patients with features suggesting T2DM also have atypical characteristics, such as presenting with ketosis ("ketosis-prone T2DM") or ketoacidosis, and/or the presence of pancreatic autoantibodies. The challenges of categorizing types of diabetes were illustrated in a multicenter study in which pediatric diabetes was classified based upon the presence or absence of beta cell autoimmunity and the presence or absence of insulin sensitivity [91]. More than 70 percent of patients fell into traditional categories of autoimmune and insulin-sensitive T1DM (55 percent) or nonautoimmune and insulin-resistant T2DM (16 percent). An additional 20 percent had both autoimmunity and insulin resistance, a pattern typical for obese patients with T1DM. The remaining 10 percent of patients were insulin sensitive in the absence of pancreatic autoimmunity, most of whom were clinically categorized as T1DM (ie, type 1B diabetes) and the remainder as T2DM. Of these 10 percent of patients with insulin sensitivity and without markers of pancreatic autoimmunity, nearly 10 percent (or approximately 1 percent of the entire study cohort) had evidence of a genetic basis for their diabetes [96]. (See 'Other causes of diabetes' below.)

In these patients with atypical features, evaluation for other forms of diabetes (eg, monogenic diabetes) should be considered, as discussed in the next section. (See "Classification of diabetes mellitus and genetic diabetic syndromes", section on 'Distinguishing type 1 from type 2 diabetes' and 'Other causes of diabetes' below.)

Other causes of diabetes — Other forms of diabetes should be considered in patients with atypical features, mixed features of T2DM and T1DM, or underlying diseases. These include:

●Monogenic diabetes – Consider monogenic diabetes in patients with a multigenerational family history of early-onset diabetes (eg, across three or more generations), clinical features of T2DM, positive C-peptide, and negative pancreatic autoantibodies (especially in patients without overweight or obesity) [96-98]. (See "Classification of diabetes mellitus and genetic diabetic syndromes", section on 'Monogenic diabetes (formerly called maturity onset diabetes of the young)'.)

●Diseases of the exocrine system – Cystic fibrosis, hereditary hemochromatosis, and chronic pancreatitis.

●Endocrine abnormalities in glucose regulation – Cushing syndrome, glucagon-secreting tumors, catecholamine excess in pheochromocytoma.

●Drug-induced diabetes – eg, glucocorticoids, HIV protease inhibitors and others (table 6).

These diseases are discussed in greater detail separately; see appropriate topic reviews and (see "Epidemiology, presentation, and diagnosis of type 1 diabetes mellitus in children and adolescents", section on 'Other causes of diabetes' and "Classification of diabetes mellitus and genetic diabetic syndromes")

SUBSEQUENT MANAGEMENT

Patients with prediabetes

●Monitoring – Laboratory screening for T2DM should be repeated annually unless there is a change in symptoms or signs (eg, weight change or polydipsia/polyuria) that indicates a need for earlier retesting. (See 'Prediabetes' above.)

●Weight management – Patients with prediabetes should be engaged in intensive lifestyle interventions similar to the nonpharmacologic therapy recommended for patients with established T2DM. For most such patients, weight reduction is an important goal.

Studies in adults suggest that intensive lifestyle interventions in patients with prediabetes can prevent or delay the onset of T2DM [99]. The natural history of prediabetes in adolescents is not well described. In a cohort of 526 adolescents with obesity and impaired glucose tolerance seen in a pediatric obesity clinic, 8 percent progressed to T2DM with median duration of follow-up of 2.9 years and 65 percent reverted to normal glucose tolerance [100]. The risk of progression to T2DM was markedly higher for non-Hispanic Black adolescents compared with White adolescents. Thus, estimates of progression vary depending on ethnicity, obesity, and other risk factors, as well as the test used to determine prediabetes. (See "Management of type 2 diabetes mellitus in children and adolescents", section on 'Nonpharmacologic therapy'.)

For adolescents with severe obesity, weight loss surgery results in durable weight loss and improvements in dysglycemia and other comorbidities. (See "Surgical management of severe obesity in adolescents", section on 'Comorbidity improvement'.)

●Role of medications – It is unclear whether metformin is beneficial for adolescents with prediabetes or with other evidence of insulin resistance. Available data from a few trials suggest that treatment with metformin improves markers of insulin sensitivity and slightly reduces body mass index (BMI) [101-103]. However, the improvements are modest, and there is no evidence that metformin alters long-term risks for developing prediabetes or diabetes.

In the Restoring Insulin Secretion (RISE) pediatric medication study, treatment with metformin for 12 months, or insulin glargine for three months followed by metformin for nine months, failed to prevent deterioration of beta cell function at 12 months and upon retesting at 15 months [104]. In the subgroup with prediabetes (impaired glucose tolerance) at baseline, 6 percent (3 of 54 patients) progressed to T2DM at 15 months despite the medications. Among subjects with prediabetes or early diabetes, the adolescents showed greater insulin resistance and a concomitant increase in beta cell function to compensate for the insulin-resistant state, and a more rapid decline in beta cell function over time compared with adults [105,106].

By contrast, studies in adults with prediabetes suggest that similar treatments improve insulin secretion and preserved glycemic control [107,108]. (See "Prevention of type 2 diabetes mellitus".)

These studies do not exclude the possibility of some protective effect of medication in pediatric subjects with prediabetes, because there was no placebo-treated group for comparison. However, they indicate a need for more aggressive approaches for the management of prediabetes in youth, including lifestyle management and the possible addition of medications that require further study.

Patients with diabetes — Management of patients with T2DM is discussed separately. (See "Management of type 2 diabetes mellitus in children and adolescents".)

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Diabetes mellitus in children".)

SUMMARY AND RECOMMENDATIONS

●Risk factors – The risk of type 2 diabetes mellitus (T2DM) is increased in children and adolescents who are overweight or obese (table 1), have an affected close relative, or are a member of a high-risk racial/ethnic group (eg, Non-Hispanic Black, Native American, Hispanic, Pacific Islander, and Asian American). Because puberty increases insulin resistance, most cases of childhood T2DM occur after the onset of puberty, which is a time of physiologic insulin resistance. (See 'Risk factors' above.)

●Clinical presentation – A majority of children and adolescents with T2DM have some symptoms of hyperglycemia, including polyuria, polydipsia, fatigue, and lethargy at presentation. A minority of T2DM patients present with diabetic ketoacidosis (DKA). Approximately 40 percent of patients are asymptomatic at the time of diagnosis and are identified by testing for the disease because of risk factors or by urinalysis obtained during routine care. (See 'Clinical presentation' above.)

●Screening – All children should be assessed for risk factors for T2DM (eg, obesity, family history, high-risk racial/ethnic group (table 2)) at routine well-child visits. For children who are overweight or obese and have one or more additional risk factor for T2DM, we suggest laboratory screening (Grade 2C). Screening consists of either fasting plasma glucose (FPG) or hemoglobin A1c (A1C). It is typically performed every three years beginning at 10 years of age or at the onset of puberty (whichever occurs first); it's performed more frequently in patients with marked obesity or strong evidence of insulin resistance (eg, acanthosis nigricans) or in patients with borderline results. (See 'Screening' above and 'Laboratory tests' above.)

●Evaluation

•Abnormal screening results – For asymptomatic children with abnormal screening results, the next step depends upon the level of concern. For those with abnormalities in the prediabetes range (table 3), it is reasonable to repeat the screening test in three to six months. If there is a high clinical suspicion of diabetes (eg, severe obesity and/or multiple risk factors for diabetes (table 2)), repeat the screening test (eg, A1C) sooner (ie, within three months) and/or proceed to an oral glucose tolerance test (OGTT). (See 'Asymptomatic with abnormal screening results' above.)

•Symptomatic patients – All patients with symptoms suggesting possible diabetes (polydipsia, polyuria, blurred vision, unexplained weight loss) should be tested for diabetes, regardless of risk factors. The first step for symptomatic patients is typically a random plasma glucose and A1C. If these are not diagnostic for diabetes, measure FPG and repeat A1C. An OGTT should be performed if the initial tests are equivocal and there is need for more accurate evaluation of dysglycemia. (See 'Symptomatic' above and 'Laboratory tests' above.)

●Diagnosis and initial management

•Prediabetes – Prediabetes refers to a state of increased risk for future development of diabetes and is defined by mild abnormalities in FPG, A1C, or glucose tolerance (table 3). Adolescents with laboratory abnormalities in this range should undergo focused lifestyle intervention to support weight loss and should be rescreened for T2DM at least annually. Youth with prediabetes or early T2DM generally experience progressive deterioration of beta cell function. (See 'Prediabetes' above and 'Subsequent management' above.)

•Diabetes – The diagnosis of diabetes is based upon the detection of one of four abnormalities of glucose metabolism; confirmatory testing is required unless symptoms are present (table 4). (See 'Diabetes mellitus' above.)

-A1C ≥6.5 percent (48 mmol/mol) on a laboratory assay (not point-of-care A1C testing)

-FPG ≥126 mg/dL (7 mmol/L)

-Symptoms of hyperglycemia and a random plasma glucose ≥200 mg/dL (11.1 mmol/L)

-Plasma glucose ≥200 mg/dL (11.1 mmol/L) measured two hours after a glucose load in an OGTT

Management of T2DM is discussed separately. (See "Management of type 2 diabetes mellitus in children and adolescents".)

●Differential diagnosis – T2DM is distinguished from other causes of diabetes, such as type 1 diabetes mellitus (T1DM), by the clinical presentation and history. Clinical features suggesting T2DM are the presence of overweight/obesity, signs and/or symptoms of insulin resistance (such as acanthosis nigricans, hypertension, dyslipidemia, and polycystic ovary syndrome [PCOS]), positive family history, and being a member of a high-risk racial/ethnic group (table 5 and table 7). Pancreatic autoantibodies are negative in most but not all patients with clinical features of T2DM. (See 'Type 2 versus type 1 diabetes' above and 'Other causes of diabetes' above.)