Chapter 544. Diabetes Mellitus

Until the 1970s, the care of childhood diabetes was pursued by internists, pediatricians, nephrologists, and general physicians. In 1971 it was estimated that visits for diabetes by those 0 to 15 years of age were equally divided among internists, general physicians, and general pediatricians.1 At that time there were few pediatric endocrinologists, and most of them did not consider diabetes to be an endocrine disorder. By 1993, pediatricians accounted for two thirds of all visits of 0- to 21-year-old diabetes patients, and over half of these were to pediatric endocrinologists; the remainder were to internists (most likely the older adolescents and young adults).2 The contemporary epidemic of obesity-related type 2 diabetes (T2D) in youth has confronted pediatricians and pediatric diabetes specialists with responsibility for a condition that was previously rare in the pediatric age group. This obesity/T2D epidemic has had pediatricians dealing with various comorbidities of insulin resistance that were formerly the exclusive domain of physicians treating adults.3

The diagnosis of diabetes includes a wide variety of diseases characterized by hyperglycemia. Because insulin is the only physiologically important hypoglycemic hormone, hyperglycemia is the result of either impaired secretion of insulin from the beta cells of the pancreas (type 1 diabetes) or resistance to the effect of insulin in the liver, muscle, and fat cells exceeding a limited capacity of the pancreas to compensate (type 2 diabetes). Criteria for the diagnosis of diabetes were revised in 2005 and prediabetes categories of impaired glucose tolerance and impaired fasting glucose added; this change reflected the recognition that these preclinical glucose intolerance states are associated with increased cardiovascular morbidity.4 The information in Table 544-1 is based on the current recommendations of the American Diabetes Association.4

Etiologic Classification of Diabetes

Contemporary understanding of the pathogenesis of various forms of diabetes has made previous classification based on treatment inappropriate (non–insulin-dependent diabetes or NIDDM, insulin-dependent diabetes or IDDM for the principal forms).5Table 544-2 is adapted from the classification published by the American Diabetes Association expert committee first in 1997, then revised based on etiology in 1999.4

Type 1 Diabetes

Diabetes that occurs in childhood is predominantly immune-mediated type 1 diabetes (T1D) associated with histocompatibility locus (HLA) specificities. Idiopathic T1D may be difficult to distinguish from the immune-mediated form. Many, if not most, African American patients who have T1D without evidence of autoimmunity have what has been termed atypical diabetes mellitus or Flatbush diabetes.6,7 This condition has its onset throughout childhood. and rarely past age 40, and is not associated with HLA specificities. Further characteristics are described in Table 544-3.

Type 2 Diabetes

Type 2 diabetes (T2D) in childhood and adolescence accounts for 10% to 50% or more of new patients with diabetes in the 10- to 19-year-old age group, depending on the ethnic/racial mix of the population served.3,8

Monogenic Diabetes

Monogenic diabetes, formerly known as maturity onset diabetes of youth or MODY, is a stable form of youth-onset (under 25 years of age) diabetes inherited in an autosomal dominant fashion and affecting almost exclusively those of European ancestry (Table 544-3). Six molecular defects have been identified, affecting cell function. Its frequency among diabetes patients is reported to vary from less than 0.2% to 5% in various populations.9

Mitochondrial Mutations

Diabetes and deafness may be associated as a result of mutations in mitochondrial DNA.10-12

Genetic Defects in Insulin Action

Genetically determined abnormalities of insulin action as a cause of diabetes are very rare; the range of effects is from mild hyperinsulinemia with modest hyperglycemia to severe diabetes requiring massive amounts of insulin or treatment with recombinant insulin-like growth factor 1.13 In the insulin resistance of lipoatrophic diabetes, function of the insulin receptor is intact, indicating that the defect lies in postreceptor transduction.14

Diseases of the Exocrine Pancreas

Injury to the pancreas rarely results in diabetes because of the small volume occupied by the β cells, with the exception of carcinoma involving the pancreas or hemochromatosis (bronze diabetes), problems not seen in pediatrics. With improved survival, increasing numbers of patients with cystic fibrosis are developing diabetes.15

Endocrinopathies

Endocrine excess syndromes including acromegaly, Cushing syndrome, glucagonoma, and pheochromocytoma have been associated with diabetes in adults.16 There is concern about the use of pharmacologic doses of growth hormone in children who do not have growth hormone deficiency, which may be increasing the risk for type 2 diabetes.17,18

Drug- or Chemical-Induced Diabetes

Drugs frequently used to treat children can impair insulin secretion, increase gluconeogenesis, or increase insulin resistance, with resultant hyperglycemia or the precipitation of diabetes.19,20 The rat poison Vacor as well as pentamidine (given intravenously), can permanently destroy β cells.21-24 Atypical antipsychotic drugs (clozapine, olanzapine, quetiapine, risperidone) have been associated with precipitation of diabetes and ketoacidosis, often with clearing of the diabetes after the drug is stopped.25,26 Exact mechanisms are uncertain, but β cell toxicity and weight gain with insulin resistance have been implicated. Other drugs can impair insulin action, probably the most common being glucocorticoids.19,20

Infections

Numerous viruses have been implicated in the induction of diabetes; the strongest evidence of a direct causation comes from the experience with congenital rubella,27 fortunately now rarely seen.

Differential Diagnosis

Table 544-3 describes the characteristic features of the common types of diabetes seen in childhood, and Table 544-4 summarizes clinical characteristics helpful in distinguishing type 1 diabetes (T1D) and type 2 diabetes (T2D) in children and adolescents. In those with acute onset who are not obese and not African American, T1D is highly likely, and further testing is not necessary. Non-obese African American youngsters with acute onset diabetes who have a 3-generation family history of diabetes indicative of autosomal dominant transmission and who do not have islet autoimmunity markers are very likely to have atypical diabetes mellitus. Islet cell autoimmunity testing should be considered in obese patients having acute onset diabetes; if this is not practical or if the patient has acanthosis nigricans, the ability to reduce and stop the acutely required insulin over the first several months, with weight reduction, exercise, and oral hypoglycemic therapy as necessary, will clarify the diagnosis.

Obese individuals with insidious onset can be considered to have T2D. If the patient is lean, islet autoantibody testing will be helpful and, if positive, indicate early detection of T1D. Negative islet cell autoimmunity in the nonobese patient may indicate monogenic diabetes, in which case testing of family members will be productive. Unless apparently unaffected family members are tested, the pattern of autosomal dominant transmission may be missed. Fasting C-peptide or insulin measurements can be of value, with elevated levels indicative of T2D. Normal levels, however, may also occur during the honeymoon period of T1D and suppressed levels can be the result of glucose toxicity. Repeated testing after 1 year or more may be needed. A patient with persistent normal C-peptide levels can be treated as with T2D, regardless of the presence or absence of diabetes-related autoimmunity.

A study of 700 newly diagnosed 5- to 19-year-old patients from 3 university centers in Florida over a 5-year period indicated that 3% of those initially classified as T1D (17/605) were later classified as T2D and that 8% of those initially diagnosed as T2D were subsequently reclassified as having T1D (6/77).28 In this relatively sophisticated clinical setting, the proportion of patients in whom classification may be problematic was less than 5% of newly diagnosed children and youth.

The distinctions indicated by the typical features in Tables 544-3 and 544-4 are not always as certain as one would like. There are a number of reasons for classification difficulty. With the increased frequency of obesity in childhood, a substantial number of newly diagnosed T1D and atypical diabetes mellitus patients will be obese, suggesting the possibility of T2D. Insulin or C-peptide measurements at the onset of diabetes may not be helpful because there may be reasonable β-cell function during the recovery phase (“honeymoon”) of autoimmune diabetes, and conversely, glucose toxicity/lipotoxicity in non–autoimmune diabetes will impair insulin secretion at the time of testing. Because some T2D patients and almost all atypical diabetes mellitus patients have ketoacidosis, this feature is not helpful to distinguish non–T1D from T1D.

There are also some classification difficulties that suggest that there will be further refinement of diabetes classification in years to come and that the contemporary distinction between 2 major forms of diabetes, T1D and T2D, will come to be seen as simplistic. Challenges to the concept of 2 main and distinct forms of diabetes have accumulated for the last 40 years. There was little diabetes among Yemenite immigrants to Israel, but 25 years later they had a 40-fold increase in the frequency of diabetes with proportions of insulin dependency similar to that among Israelis of European origin, suggesting common environmental influences, most likely nutritional, for both T1D and T2D29 The specificity of a family history of T2D is low because of the high frequency of this disorder in the general population, particularly in minority populations. Furthermore, a family history of T2D is 3 times as common with T1D as in the general population, and conversely, T1D is more frequent in the relatives of patients with T2D30,31 The finding of islet autoimmunity markers at onset in a substantial percentage of children with typical T2D, and related to degree of obesity or treatment requirement, has been a further challenge to the notion of immune-mediated diabetes as a distinct and uniform clinical and pathologic entity.32-34

Despite uncertainties in our classification of diabetes, the clinician dealing with the individual patient should always base treatment on individual characteristics rather than absolute diagnostic classification.

Epidemiology

Rapid increase in knowledge of the epidemiologic patterns and natural history of type 1 diabetes (T1D) has heightened understanding of the disease, but has failed to provide an intervention to prevent T1D or to identify the environmental triggers.35

In 1990, the World Health Organization began The Multi-National Project for Childhood Diabetes (DIAMOND) study to investigate and monitor the patterns of incidence of T1D.36 This study involved surveillance of 4.5% of the world’s population 14 years of age or younger.37 From 1990 to 1994, nearly 20,000 cases of T1D were diagnosed in the 75 million sample population. Annual incidence rates (per 105 children per year) among the 100 populations studied ranged from 0.1 in China and Venezuela to 36.8 in Sardinia and 36.5 in Finland. Sweden, Norway, Portugal, the United Kingdom, Canada, and New Zealand also had very high rates (> 20) (eFig. 544.1).

Some neighboring areas in Europe and North America have shown great differences in incidence. Across the eastern border of Finland and Russia there is a sixfold gradient in the incidence of T1D between areas in geographic proximity, despite equal frequencies of high-risk and low-risk HLA genotypes.38 Alberta and Prince Edward Island, across the continent from each other in Canada, have comparable high rates of 24 and 24.5. Rates in the United States vary from 11.7 in Chicago to 17.8 in Pennsylvania. The rate in Puerto Rico, 17.4, is almost the same as that in Pennsylvania, while that in neighboring and ethnically similar Cuba is only 2.9.37 Estimate of T1D incidence in the United States is 15 to 17, with a prevalence of 1.7 to 2.5/1000.37

The environmental influence on T1D susceptibility can be demonstrated in migrant populations. Israeli children in Canada have a fourfold greater incidence than those in Israel.37 Indian children migrating to England from South Africa, where T1D in the Indian population is very low, developed incidence rates comparable to those of English children in the community.39 Japanese have a very low incidence of T1D, but those living in Hawaii are 5 times more likely to have T1D than those in Japan. Ethnic French and Italian children in Montreal have twice the incidence of diabetes as those in France and Italy.40-42

eFigure 544.1 shows the incidence of T1D diabetes in various countries worldwide. Incidence rates in European countries have steadily increased from the 1950s.43-45 Similar increases worldwide as countries undergo development emphasizes the influence of environment on T1D. In France, there was a 4.9% rise in incidence between 1980 and 1998,46 and Sweden and Lithuania had similar annual increases in incidence from 1983 to 2000 of 2.2% and 2.3%.47

The average age of onset of diabetes is thought to be decreasing.48-50 Modest seasonality has been noted, with a peak in winter months and a summertime dip in both the northern and southern hemispheres.44,51,52

Etiology and Prevention

Observations that type 1 diabetes (T1D) had a long latency period began almost 4 decades ago.53,54 Subsequent identification of genetic markers (HLA haplotypes), autoimmune markers reflecting islet cell destruction, and early biochemical abnormalities preceding frank diabetes have greatly improved predictability of T1D. These studies have reinforced the long-held view that T1D develops in genetically predisposed individuals who are exposed to one or more environmental influences resulting in autoimmune destruction of β cells by autoreactive T cells (Fig. 544-1).35,55 The injury to the β cells releases antigens perceived as foreign proteins by the immune system, which produces antibodies against them. These autoantibodies can be assayed as a reflection of the damage, but they are not the cause of it. Insulin autoantibodies (IAA), islet cell antibodies (ICA), glutamic acid decarboxylase (GAD65) antibodies, and antibodies to tyrosine phosphatase (IA-2 or ICA512) are indicators of ongoing autoimmune disease and may be present for many years before the development of overt diabetes. GAD65 antibodies can be detected in 75% of newly diagnosed diabetes patients56 and IA-2 in 50% to 60% of patients.57 The number of different autoantibodies and their titers are independent predictors of T1D risk, which also depends on the age at which they are detected. ICA titers of greater than 40 Juvenile Diabetes Foundation units convey a 60% to 70% risk of developing diabetes over 5 to 7 years.58 ICA in the first few years of life is associated with an approximate 90% 10-year risk of developing diabetes, whereas in a 40-year-old this would convey a 30% 10-year risk. With ongoing destruction of β cells, the antigen disappears, and after 10 years of overt diabetes, less than 5% of patients have detectable ICA. When IAA are found in combination with other autoantibodies, the risk for T1D increases significantly. In the large Diabetes Prevention Trial–Type 1), 5-year risk of T1D was 20% to 25% for subjects with 1 autoantibody marker, 50% to 60% for those with 2 markers, nearly 70% with 3 markers, and almost 80% with 4 autoantibody markers.59 The earliest metabolic abnormality is loss of first-phase insulin release during intravenous glucose tolerance testing (IVGTT). This is followed by abnormal oral glucose tolerance testing and, finally, overt diabetes.35,57,60

Immune markers measurable before the onset of diabetes has led to development of prediction models. Early assays that required O-type human pancreas substrate were reliable only in a few research laboratories.61,62 Immunoassays (GAD65, IA-2) are far more reliable and, have a high predictive value for T1D. Relatives of T1D subjects can be classified into groups with risks of diabetes ranging from 5% to 90% within 5 years.63 Other risk factors include young age, absolute number of autoantibodies, high titer antibodies, low first-phase insulin response to intravenous glucose tolerance testing, and high-risk human leukocyte antigen (HLA) genotypes (DR3/4, DQB0302/DQB0201) (Table 544-2).64,65 Loss of first-phase insulin secretion during IVGTT (first-phase insulin release) predicts progression to overt diabetes over the next 5 years in greater than 85%.66

Outside of the research setting and for the development of accurate type 1 diabetes (T1D) prediction models, diabetes-specific autoantibody testing is seldom necessary in diabetes practice. Such testing may be helpful in differentiating new onset T1D from T2D in the obese individual with characteristics of T1D, although even in this situation, individualized treatment is based on the clinical status rather than the specific diagnosis. Antibody testing may also be helpful for determining diabetes risk in individuals with stress- or drug-induced hyperglycemia.

The role of environmental factors in the development of T1D is supported by the finding that 30% of identical monozygotic twins are concordant for the disease67 and by the rising incidence with decreasing mean age of onset.37,47,49,50 Various environmental factors that have been implicated include congenital rubella infection,68 enteroviral infection,69,70 reduced exposure to early infection,71 and vitamin D deficiency.72 There is no evidence to support the notion that standard vaccination increases risk for T1D.73

An influence of cow’s milk protein on the development of T1D is based on the structural similarity of a 17-amino acid fragment (ABBOS) of bovine serum albumin to an islet antigen (ICA 69).74 The Diabetes Autoimmunity Study of the Young (DAISY) found no association between islet autoimmunity and early cow’s milk exposure.75 A more robust double-blind, controlled study is underway that will determine if avoidance of nutritional cow’s milk proteins for at least the first 6 months of life will reduce the incidence of diabetes.77 Investigators in Sweden and Finland found IgM antibodies to enterovirus more frequently in pregnant women whose children developed T1D than in control women.78,79 There is molecular mimicry between the P2-D protein of coxsackie virus and the GAD protein. Other viruses implicated include chickenpox, mumps, measles, ECHO, rotavirus, and maternal rubella.78,80 The evidence is most convincing for congenital rubella, with 20% of infected children islet cell antibodies-positive.68

Clinical Features

Type 1 diabetes (T1D) is usually diagnosed in a child with polydipsia, polyuria, weight loss, and decreased energy. Polyphagia is often absent, as ketosis can cause anorexia. With insulin deficiency, glucose utilization is impaired, resulting in elevated blood glucose concentrations that exceed the renal threshold for reabsorption (~ 180 mg/dL), producing osmotic diuresis and glycosuria. The tissues perceive starvation because of absence of intracellular glucose, causing the release of triglycerides from fat stores and protein from muscle. These substrates are processed by the liver to yield ketones and more glucose. The tissue breakdown leads to weight loss and glycosuria to polyuria and polydipsia. Insulin deficiency, dehydration, and ketosis will progress to diabetic ketoacidosis before diagnosis in 20% to 40% of all newly diagnosed T1D.81

The primary care physician is likely to see an child who appears to be ill for evaluation of weight loss and fatigue or nonspecific symptoms of dehydration, including constipation, headache, and abdominal pain. Diagnosis is often missed in emergency rooms or doctors’ offices because the symptoms are attributed to a viral syndrome or gastroenteritis if there is vomiting. Children who are vomiting but do not have diarrhea and are maintaining high urine output despite dehydration have T1D. Other causes for this combination of symptoms are rare (adrenal insufficiency, renal disease). Enuresis in a previously toilet-trained child, nocturia, pyogenic skin infections, recurrent candida diaper rash in babies and toddlers, and monilial vaginitis in teenagers also require consideration of diabetes.

Children in whom diabetes must be considered in addition to those who have typical signs and symptoms include those who have glucosuria without hyperglycemia and those with transient hyperglycemia.

Renal glucosuria may be an isolated congenital disorder, a feature of the Fanconi syndrome and other renal tubular disorders, due to severe heavy metal intoxication, due to ingestion of certain drugs (eg, outdated tetracycline), or result from inborn errors of metabolism such as cystinosis. Starvation ketosis may occur in these conditions with gastrointestinal disturbance, but the absence of hyperglycemia eliminates the possibility of diabetes. Stress of illness, usually acute asthma exacerbation or severe infection, especially in toddlers, may be associated with moderate to severe hyperglycemia and ketosis.82,83 A similar clinical picture can be seen with glucocorticoid administration.84,85 The presence of islet cell antibodies, IA-2, insulin autoantibodies, or GAD antibodies in those with stress or glucocorticoid-induced hyperglycemia may indicate limited β cell reserve and risk for future development of diabetes.

Complications of diabetic ketoacidosis account for most of the diabetes-related morbidity and mortality in childhood T1D81 diabetic ketoacidosis (DKA) is defined by venous pH less than 7.3 or serum bicarbonate concentration less than 15 mmol/L, with serum glucose concentration greater than 200 mg/dL (11 mmol/L) and is typically associated with glucosuria, ketonuria, and ketonemia.86 Rarely, DKA may occur with normal circulating glucose concentrations if there has been partial treatment or with pregnancy.87,88

Frequency

The frequency of DKA at onset of diabetes varies widely by geographic region, correlating inversely with the regional incidence (ie, level of awareness) of T1D. In the European diabetes study, this frequency varied from 11% to 67%.89 The frequency in Australia was 26%,90 and in New Zealand 63% in 1988/1989 and 42% in 1995/1996.91 DKA at diagnosis is more likely in children less than 5 years of age and in those social or economic situations that do not permit ready access to medical care.81,89-96 In patients from Colorado, DKA was seen at onset in 28.4%, with the odds ratio for uninsured subjects compared to insured subjects being 6.2 and with significantly greater severity in the uninsured group.94 Long-term reductions in frequency of DKA at onset have been reported following zealous efforts to educate the medical community.90-92

Among 1243 patients in Colorado, the risk of recurrent DKA was 8 episodes per 100 patient years; 80% of episodes involved 20% of patients. Factors increasing the risk of recurrent DKA were poor metabolic control or previous episodes of DKA, female gender (peripubertal or adolescent), psychiatric disorders including eating disorders, difficult or unstable family circumstances, limited access to medical services, and insulin pump therapy (because only rapid- or short-acting insulin is used in pumps, interruption of insulin delivery for any reason rapidly leads to insulin deficiency).97 In the 1970s and 1980s, the establishment of treatment teams with intensive education of families on sick day management and 24-hour availability demonstrated profound reduction in recurrent DKA.98,99

Pathophysiology

See Figure 544-2. Diabetic ketoacidosis (DKA) occurs when there is a critical net effective deficit of insulin, resulting in intracellular starvation, stimulating the release of the counterregulatory hormones glucagon, catecholamines, cortisol, and growth hormone. The counterregulatory hormonal responses stimulate hepatic and renal glucose production and reduce glucose utilization in insulin-dependent peripheral tissues (muscle, liver, adipose), which results in hyperglycemia, release of free fatty acids into the circulation from adipose tissue (lipolysis), and unrestrained hepatic fatty acid oxidation to ketone bodies.100

Morbidity and Mortality

Similar mortality risks for each episode of DKA have been reported from the United States (0.15%),101 Canada (0.18–0.25%),102,103 and the United Kingdom (0.31%).104 Cerebral edema accounts for 60% to 100% of this mortality.105-110 Other causes of death or disability with DKA in addition to cerebral edema include hypokalemia, hypophosphatemia, hypoglycemia, other intracerebral complications, peripheral venous thrombosis, mucormycosis, rhabdomyolysis, acute pancreatitis, acute renal failure, sepsis, aspiration pneumonia, and other pulmonary complications.81

Cerebral Edema in Diabetic Ketoacidosis

In 1967 two teenagers with fatal cerebral edema following DKA were described. Prior to that time there were only 3 other cases with a similar clinical history reported in the literature.126 By 1980 only 17 cases had been reported.127 Ten years later there were 40 reported cases, and an additional 29 were added in report and 11 in another.105,128 Since 2000, approximately another 160 cases have been added, collected over periods varying from 2 to 15 years. These reports have provided incidence data and have added to the appreciation of the idiosyncratic nature of the occurrence of cerebral edema and the absence of iatrogenic causation.129-131

Prevalence and Incidence

Bello and Sotos128 provided the first estimates of the incidence of cerebral edema in DKA based on their hospital experience with 1006 episodes in pediatric patients, of whom 11 had cerebral edema for an incidence of 1.1%; newly diagnosed patients had a cerebral edema frequency of 3.3% and established patients 0.23%. The interpretation of these data is limited because this was a biased population of more severely ill referred individuals, including some in advanced states of deterioration with cerebral edema who were transferred from other hospitals. A 3-year United Kingdom population-based study involving nearly 3 times as many episodes yielded a lower incidence of 0.7% overall, but also showed a marked discrepancy between the newly diagnosed, 1.2%, and established patients, 0.4%.106 A higher incidence of 4.3% in newly diagnosed, 1.3% in established patients, and 2% overall was reported from Australia.90 The North American multicenter study reported a cerebral edema incidence of 0.9%without differences between new onset and established patients.107 There were 13 verified cases of cerebral edema in a prospective surveillance study in Canada of patients younger than 16 years of age with DKA, for a rate of 0.51% of episodes of DKA.103 Variations in case definition and referral patterns likely explain some of these differences in frequency of cerebral edema.

Morbidity and Mortality

Case fatality and residual morbidity data from published case reports or litigation reflected ascertainment bias and likely later recognition of cerebral edema than with more recent population-based observations. Appropriate intervention was also less frequent in these early or litigated cases. In 1990 the first extensive series, analyzing 69 cases, including 40 from the literature, found a case fatality rate of 64% overall, with 14% surviving without disability, 9% with disability that did not preclude independence, and 13% with severe disability or vegetative state. The case fatality was only 30% in those who were treated before respiratory arrest, even if this was at the gasping stage, with 30% surviving without disability, 26% with disability not precluding independence, and 13% with severe disability or vegetative state.105 Although the intervention in these cases was often after severe neurologic compromise and inadequate, the outcome with treatment is similar to those in contemporary population-based studies (eTable 544.1). There had been a few case reports of cerebral edema developing before initiation of treatment of diabetic ketoacidosis and a recent report from Canada documented that 19% of cases of cerebral edema were present before treatment was undertaken.103,105-107

Risk Factors

Young age, especially younger than 5 years, carries an increased risk for the development of cerebral edema. This may reflect the greater severity at presentation in this age group due to delay in diagnosis because of the nonspecificity of symptoms. The younger brain might also be more susceptible to metabolic and vascular changes associated with diabetic ketoacidosis. Severity of acidosis, degree of hypocapnia, and elevated serum urea nitrogen, indicators of severity of ketoacidosis and dehydration, have been noted to be risk factors.81 Infusion of sodium bicarbonate has been a significant association with the development of cerebral edema in 1 population-based case-control study.107 It is uncertain if this finding was due to a greater degree of severity of ketoacidosis or was an effect of the bicarbonate itself.133,134

Cerebral edema associated with diabetic ketoacidosis is peculiar to the pediatric age group and is almost nonexistent in the adult population.131 In adult patients with hyperglycemic hyperosmolar state, this is true despite rapid rehydration and restoration of normal glycemia.132

In several studies increased risk for cerebral edema has been associated with an inadequate increase in serum sodium concentration during treatment for diabetic ketoacidosis. This may reflect dysregulation of antidiuretic hormone secretion as a result of changes in the brain in those who are in the early, presymptomatic, stages of cerebral edema.107,133 The absence of evidence of associations between volume or tonicity of fluids or the rate of change in serum glucose and risk of cerebral edema indicates that this might be the result of altered renal sodium handling due to cerebral injury, rather than the type or rate of fluid administration.81,103,105,128,133

In the 61 children with cerebral edema from the North American multicenter study,107 poor outcome was associated with greater neurologic depression at the time of diagnosis of their cerebral edema, and with initial serum urea nitrogen concentration. Intubation with hyperventilation to a Pco2 less than 22 mm Hg was also a risk factor, demanding caution in hyperventilating affected patients.81,134

Mechanisms of Cerebral Edema

Brain edema refers to an increase of cerebral tissue water causing an increase of tissue volume.135 The edema may be vasogenic, due to breakdown of the blood-brain barrier as around a tumor or with trauma; cytotoxic, from poisoning or metabolic derangement; or osmotic, as occurs with hyponatremia. Neither the cause nor the location of the fluid in the swollen brains of children with diabetic ketoacidosis–related or hyperglycemic hyperosmolar state–related cerebral edema is known. The mechanism is likely to be complex, and it may not be the same in all affected individuals, as reflected in the time of onset and the brain-imaging findings. The time of onset is distributed in a bimodal fashion, with approximately 2/3 of patients developing signs and symptoms in the first 6 to 7 hours and the rest from 10 to 24 hours after the start of treatment, with the early-onset individuals tending to be younger.107,133,134

A wide variety of imaging findings was reported in an analysis of initial brain computerized tomography (CT) scans done 2 to 44 hours after establishment of the diagnosis of cerebral edema; 39% had no acute abnormalities visible, and this did not differ significantly between early- and late-onset subjects; 26% had diffuse edema, which was also similar between early and late onset. Three of the 8 with diffuse edema also had hemorrhages, and 17% of the entire group had only subarachnoid or intraventricular hemorrhage. Focal brain injury in the mesial basal ganglia and thalamus, the periaqueductal gray matter, and the dorsal pontine nuclei was seen in 5 subjects (22%), and these localized injuries occurred only in the early-onset patients. These findings were not an artifact of the time that the studies were performed and thus truly reflect what appear to be widely varying pathology in the brain leading to the syndrome that is referred to as idiopathic cerebral edema.133 The observation that approximately 40% of initial brain-imaging studies in children who have cerebral edema are normal emphasizes that cerebral edema is a clinical diagnosis that requires treatment be started before imaging studies are undertaken.

Alterations in the brain with hyperglycemia and with diabetic ketoacidosis (DKA) have been explored using functional brain imaging. In thirty-one 10- to 18-year-old patients who were not in ketoacidosis and 7 controls, single photon emission tomography using technetium indicated significant brain hypoperfusion in those with diabetes, mainly in the basal ganglia and frontal regions, suggesting underlying vascular changes that might be contributory to the risk for cerebral edema with the advent of DKA.136 Diffusion MRI, quantified as the apparent diffusion coefficient (ADC), is highly sensitive for edema. Vasogenic edema, which expands the extracellular space, is characterized by increased ADC, whereas cellular swelling, such as occurs with hypoxic/ischemic injury or osmotic swelling, decreases the ADC. Perfusion MRI complements diffusion imaging by monitoring the first-pass intravenously administered contrast agent (gadolinium) to assess relative cerebral blood flow and cerebral blood volume. If cerebral edema during treatment of DKA were the result of osmotic shifts resulting in cellular swelling, ADC should be decreased and cerebral perfusion diminished or unchanged. Fourteen children aged 3.5 to 15.5 years were studied during DKA treatment and after recovery, an appropriate approach because several studies have documented subtle cerebral edema changes on CT scan in children without significant neurologic symptoms during their treatment for DKA.137-140 Increased water diffusion in all regions except the occipital gray matter was indicated by significant elevation of ADC during DKA treatment followed by reduction of ADC in the postrecovery MRI. Perfusion MRI during treatment showed significantly shorter mean transit times and higher peak tracer concentrations, suggesting increased cerebral blood flow. The predominant mechanism of edema formation suggested by these findings is a vasogenic process rather than osmotic cellular swelling.139,141 These observations further confirm the presence of subtle, asymptomatic cerebral edema during DKA treatment.137-140 The functional imaging studies139 suggest that cerebral hypoperfusion before DKA treatment,136 followed by reperfusion, might lead to cerebral edema. Ischemic brain injury is characterized in the acute to subacute phases by low ADC values (ie, restricted diffusion), likely a result of both astrocytic swelling and diminished intracellular microcirculation. Elevated ADCs in DKA do not contradict a possible role for cerebral hypoperfusion in the pathogenesis of DKA–related cerebral edema because the ADC may return to normal within 1 hour after reperfusion of ischemic tissue. The possibility remains that the initial and typical central nervous system (CNS) response to DKA is vasogenic edema, while symptomatic cerebral edema might require prolonged hypoxia or further insult (inappropriate vasopressin release, fluid overload, central hypoxia from rapid correction of peripheral acidosis with bicarbonate administration) leading to the superimposition of cellular edema.

Brain microvascular endothelium is directly affected by ketones, which may alter permeability of the blood-brain barrier and disrupt membrane structure by causing oxidative tissue damage through production of reactive oxygen species and by compromising free radical scavenging. These effects could contribute to cerebral edema. β-Hydroxybutyrate was detected in the brain by proton magnetic resonance spectroscopy in half of 25 children during or after treatment for DKA overall and in 90% of those tested within 4 hours of the start of treatment. Acetoacetate was detected in 60%, with greater frequency after 4 hours. This inverse relationship, as β-hydroxybutyrate is oxidized to acetoacetate, reflects changes in the concentrations of these substances with treatment. Ketones were detected in the brain more frequently in children with altered mental status during treatment.141

MRI and MR spectroscopy of the brain in 8 children (7 newly diagnosed) with hyperglycemia with or without DKA has provided further insight into the background on which cerebral edema develops. All had abnormal signal changes in the frontal region indicative of edema upon fluid-attenuated inversion recovery MR imaging. Spectroscopic abnormalities were increased taurine, myoinositol, and glucose. The abnormalities varied in severity, did not correlate with any clinical or biochemical measures, and were highly consistent between patients, thought to reflect homeostatic responses of the brain to prolonged hyperglycemia and associated osmotic pressure changes.142 Taurine has been considered important in neuroprotection and osmoregulation and has been implicated in animal models of cerebral edema.

Serum levels of inflammatory cytokines, as markers of cellular activation, were elevated in 6 children and adolescents before treatment of severe DKA. There were significant decreases of some cytokines during treatment, while inflammatory cytokines increased, reaching a peak consistent with the peak time of cerebral edema development. Such elevation could have effects on capillary function that could contribute to the expression of clinical cerebral edema.143

Monitoring

Because treatment modification has not prevented cerebral edema and with the strong evidence that early administration of mannitol prevents brain damage and death from this complication, Muir et al. developed a model for early detection.133 They applied signs of neurologic impairment based on 26 case reviews, which were then tested with an independent sample of 17 patients who had previously been reported with symptomatic cerebral edema105 for whom records were still available. The system allowed 92% sensitivity and 96% specificity for the recognition of cerebral edema early enough for intervention, using 1 diagnostic criterion, 2 major criteria, or 1 major plus 2 minor criteria (Table 544-5). This means that 5 youngsters will be treated for cerebral edema to prevent brain damage or death in 1 patient, a reasonable proposition considering the alternative of waiting for more stringent criteria to be met.

Treatment

While there can never be a controlled study of intervention, the response to rapid treatment with mannitol and associated measures as soon as cerebral edema is suspected, especially before respiratory arrest, has been repeatedly observed.81,105,117,128 Intervention measures include reduction in the rate of fluid administration, elevation of the head of the bed, and administration of intravenous mannitol in a dosage of 0.25 to 1.0 g/kg over 20 minutes, repeated as necessary in 1 to 2 hours. Although early intervention with mannitol treatment has improved outcome, it is difficult to determine to what degree, because increased recognition of the problem has undoubtedly led to less stringent case definition. The application of criteria developed by Muir et al. from the most severe cases, when applied to a series of 69 consecutive cases thought to be uncomplicated and experiencing full recovery, yielded 3 (4.3%) who would have been appropriately treated with mannitol according to these criteria.133 Indeed, this is the approximate percentage of diabetic ketoacidosis patients treated as cerebral edema in a recent report of 18 years’ experience.144

The effects of mannitol include lowering hematocrit and blood viscosity, which improves cerebral blood flow (CBF) and oxygenation, improvement in red cell deformability, vasoconstriction in areas of the brain with intact autoregulation or improved autoregulation from the effects of the mannitol on CBF and oxygenation, and direct osmotic effects with reduction in extracellular free water. Intensivists frequently express concern about the use of mannitol because of their experience with the risk of rebound edema and renal failure when mannitol is used over an extended period. There have been no reports, however, of any complications of mannitol use in this acute situation.

Hypertonic saline has been used to treat acute intracranial hypertension in head injury animal models and in patients undergoing surgical procedures in the supratentorial portion of the brain; in these circumstances it has been considered at least as effective as mannitol.145-151 Its use in diabetic ketoacidosis–associated cerebral edema was initially described in a 13-year-old girl with relatively mild acidosis who developed a severe headache 20 minutes after the start of treatment and who had CT scan evidence of diffuse cerebral edema with transtentorial herniation. Despite infusion of mannitol, she continued to deteriorate and was given 5 mL/kg of 3% saline rapidly; she awoke with recovered neurologic function within 5 minutes.152 The use of 3% saline in patients who have not responded adequately to mannitol infusion of a dose of 1.0 g/kg appears justified.

Intubation is obviously necessary for those who have progressed to respiratory compromise, but should not otherwise be undertaken simply because of the diagnosis of cerebral edema. Aggressive hyperventilation was a significant risk factor for poor outcome in the study of Marcin et al.,134 similar to detrimental effects reported in head trauma and high-altitude exposure.153-154

Other Complications of Diabetic Ketoacidosis

Rhabdomyolysis

Rhabdomyolisis occurs with diabetic ketoacidosis (DKA) and is of uncertain etiology. Risk factors include severe hyperglycemia, high osmolality, and hypophosphatemia. A recent report in a 15 month old has added to the few previously reported pediatric cases.109 A quarter of a century earlier, the occurrence of rhabdomyolysis in a 27-month-old patient led to a chart review of 133 children admitted with new onset diabetes; 12 of them had significant orthotolidine reactions in the urine. Urine myoglobin was measured in only 1 case and it was extremely high. Another patient had an elevated serum creatine kinase value 3 days after admission.110

Mucormycosis

Mucormycosis is an acute, rapidly progressing, and often fatal facultative fungal infection that occurs in young patients with diabetes who have poor glycemic control and acute ketosis. The principal infection locus may be rhinocerebral, pulmonary, cutaneous, gastrointestinal, or in the central nervous system, and may be disseminated. In 1 study of 5 patients, which included only 1 survivor who had severe neurologic disability, predictors of increased susceptibility to infection with mucormycosis were African American and had a history of poor compliance, as evidenced by poor clinic attendance, risk-taking behaviors, and high rate of hospital admission.111 Impaired neutrophil function, vascular insufficiency, and, in DKA, the acid environment and high blood glucose concentration favoring fungal growth, contribute to the predisposition to mucor infection in patients with diabetes.

Acute Pancreatitis

Pancreatitis is a well-recognized complication of DKA in adults but is unusual in childhood type 1 diabetes (T1D). Elevation of serum levels of amylase and lipase was common in 50 children with DKA at onset of T1D. The elevation of serum lipase was directly associated with degree of acidosis. Acute pancreatitis was demonstrated by abdominal computerized tomography scan in 1 patient with persistent abdominal pain.112 Acute pancreatitis must be considered with abdominal pain that does not resolve with correction of acidosis.

Deep Venous Thrombosis

Deep venous thrombosis (DVT) has been reported in 4 of 8 children with DKA who required femoral central venous catheter (CVC) placement113 and in 3 of 6 pediatric ICU patients with DKA who required femoral CVC (out of 113 DKA/pediatric ICU patients).114 The pediatric ICU patients were under 18 months of age and required long-term heparin therapy for persistent leg swelling; in contrast, there was complete absence of DVT in 7 comparably aged patients with shock who had femoral CVC for longer periods of time than did the DKA patients, suggesting that DKA is associated with a thrombotic diathesis beyond that which is attributable to dehydration. One of the ICU patients was heterozygous for factor V Leiden deficiency, which may have been contributory, but no other abnormalities were detected. Coagulation factor abnormalities have not been demonstrated in children with diabetes in the absence of DKA115 but von Willebrand factor activity remains elevated at 120 hours following admission for DKA.116 This suggests that DKA and its treatment may result in a prothrombotic state and activation of vascular endothelium, predisposing to thrombosis.

Greater than 50% of children who have venous thromboses will carry a genetic thrombophilic defect, and two thirds of such defects will be a mutation in the factor V gene referred to as factor V Leiden.117 Massive arterial thrombosis resulting in unilateral below-the-knee amputation occurred in a 12-year-old white girl with a heterozygous factor V Leiden mutation and a 2-year history of poorly controlled T1Dmellitus who was not in DKA.118

Intracerebral Complications Other Than Idiopathic Cerebral Edema

Neurologic complications occur in an estimated 10% of all instances of neurologic collapse during ketoacidosis.81,119-121 The causes include subarachnoid hemorrhage,105 basilar artery thrombosis,105 cerebral venous thrombosis,119-121 meningoencephalitis,122 and disseminated intravascular coagulation.123,124

One 5-year-old girl with established diabetes, who had mild ketoacidosis and microcytic anemia (a risk factor for thrombosis), developed thrombosis in the straight sinus and the vein of Galen with ischemic changes in the thalamus. She had a clinical picture indistinguishable from typical cerebral edema. With anticoagulant therapy, she made a full recovery, but with mild residual learning disability.120 An 11-year-old boy with deep venous thrombosis of the leg preceding admission for diabetic ketoacidosis at the onset of diabetes, in contrast, was found unconscious with fixed dilated pupils, but breathing. Respiratory arrest would be expected with unconsciousness and fixed dilated pupils as a reflection of brainstem herniation or compression if he had cerebral edema. Two hours after his collapse, MRI showed acute infarction affecting multiple areas of the brain without hemorrhage or edema; he did not survive. Heterozygosity for factor V Leiden deficiency was considered responsible for the unusual occurrence of deep venous thrombosis in a child.121 Case series of cerebral venous thrombosis in children from various acute causes, including dehydration, indicate that thrombosis is unlikely to occur in the absence of background coagulopathy.125

Management of Diabetic Ketoacidosis

The International Society for Pediatric and Adolescent Diabetes has recently completed consensus treatment guidelines.81 Children with established diabetes, who are not vomiting or severely ill and whose caregiver has been trained in sick day management, can be managed at home or an outpatient facility with appropriate supervision and follow-up.155-158 Children requiring intravenous rehydration over an extended period need to be admitted to a unit in which neurologic status and vital signs can be monitored frequently and blood glucose levels measured hourly.81,158 Nursing staff trained in monitoring and management of diabetic ketoacidosis (DKA) should be available, and there should be written treatment guidelines.81,158 Those with severe DKA (long duration of symptoms, compromised circulation, depressed level of consciousness) or with other factors increasing the risk for cerebral edema (ie, age under 5 years, low Pco2, high serum urea nitrogen) should be in an intensive care unit, preferably pediatric, or an equivalent facility.81,158

Evaluation

Differential diagnosis should include other causes of coma in children that might be present instead of or in addition to DKA (see Chapter 111). A body weight should be obtained for calculation purposes. The degree of dehydration can be estimated as 5% with reduced skin elasticity, dry mucous membranes, tachycardia, and deep breathing and up to 10% with capillary refill time greater than 3 seconds and sunken eyes. It is typical for calculations of fluid deficit to be based on 10% dehydration, which is a modest overestimate that does not appear to have clinical significance.159 The level of consciousness should be assessed using the Glasgow coma scale (Table 104-1).160 An initial blood sample should be tested for glucose, electrolytes, pH, urea nitrogen, creatinine, osmolality, ketones or β-hydroxybutyrate, hemoglobin and hematocrit or complete blood count, keeping in mind that diabetic ketoacidosis is associated with leukocytosis that does not correlate with the presence or absence of infection. As a substitute for measured osmolality, calculated osmolality may be determined as: 2× [Na + K] + glucose in millimoles per liter (or glucose in milligrams per deciliter divided by 18). A urine specimen should be checked for ketones. Appropriate culture specimens should be obtained as indicated. If there is any possibility of delay in obtaining a serum potassium result, electrocardiographic monitoring should be done for potassium status.161,162 Continuous electrocardiographic monitoring using standard lead II should be done for severely ill patients or if the initial potassium level is less than or equal to 3 mEq/L or more than or equal to 6 mEq/L; depressed T waves indicate hypokalemia, and peaked T waves hyperkalemia.163

Determination of glycohemoglobin (HbA1c) concentration will provide insight into the severity and duration of hyperglycemia in new patients and is useful for comparison to the previous determination in an established patient. It may also be useful to obtain a blood sample for measurement of free insulin to confirm an impression of failure to administer insulin as the cause of the diabetic ketoacidosis before administering any insulin.164

The severely obtunded child will require continuous nasogastric suction for the prevention of pulmonary aspiration and may require oxygen. Bladder catheterization should be performed only in individuals who are unable to void on demand; older patients may have atonic bladder with urinary retention requiring initial catheterization, but not indwelling catheterization.

Successful management and early intervention for complications require close monitoring. A flow chart should be maintained with as frequent entries as indicated by the patient’s condition (Table 544-6).165 Monitoring frequency recommendations are in Table 544-7.165

Fluid and Insulin Replacement

The goals of treatment with fluid and insulin are to (1) restore perfusion, which will increase glucose use in the periphery, restore glomerular filtration, and reverse the progressive acidosis; (2) arrest ketogenesis with insulin administration, reversing proteolysis and lipolysis while stimulating glucose uptake and processing, thereby normalizing blood glucose concentration; (3) replace electrolyte losses; and (4) avoid, insofar as possible, the complications of treatment or to intervene rapidly when they occur.

Fluid Therapy

• Dehydration can be assumed to be 5% to 10% (50–100 mL/kg). As noted above, clinical evaluation usually overestimates the level of dehydration. Serum urea nitrogen and hematocrit may be useful for determining severity of extracellular fluid (ECF) contraction. Serum sodium concentration is not reliable for determining ECF deficit because of the osmotic effect of hyperglycemia-induced dilutional hyponatremia166,167 and the low-sodium content of the elevated lipid fraction of the serum in diabetic ketoacidosis. Corrected sodium, that is, for normal glucose level, can be estimated by adding 1.6 mEq to the measured value for each 100 mg/dL blood glucose above normal.166
• NaCl (0.9%) at 20 mL/kg should be provided in the first 1 to 2 hours to restore peripheral perfusion.
• Maintenance can be calculated as 1000 mL for the first 10 kg body weight + 500 mL for the next 10 kg + 20 mL/kg over 20 kg.
• The remainder of replacement after the loading dose based on 5% to 10% dehydration, and maintenance, can be distributed over the subsequent 22 to 23 hours. Some prefer to provide half of replacement in the first 8 hours. It is important to consider fluids that have recently been administered orally at home (if not vomited) and parenteral fluids provided in the emergency room or referring institution.
• If osmolality (calculated or measured) is greater than 320 mosm/L, replacement can be calculated for correction in 36 hours, and if greater than 340 mosm/L, in 48 hours. Except for severely ill and very young individuals, oral intake will begin before 24 hours.
• While urinary output should be carefully documented, urinary losses should not be added to fluid requirements.81
• After initial 0.9% NaCl bolus, rehydration/maintenance should be continued with 0.45% NaCl. During rehydration, the measured Na can increase to the level of the corrected Na as glycemia declines and then decline to normal levels if the corrected level was elevated.
• The use of large amounts of 0.9% saline may result in hyperchloremic metabolic acidosis.168,169
• To prevent unduly rapid decline in plasma glucose concentration and hypoglycemia, 5% glucose should be added to the IV fluid when the plasma glucose falls to 300 mg/dL (17 mmol/L). An efficient method of providing glucose as needed without long delays entailed by the changing of IV solutions is to have 2 IV fluid bags connected, 1 containing 10% dextrose and the appropriate sodium and potassium concentrations and the other with the same salt concentrations but no dextrose, the so-called 2-bag system.170
• Potassium (20–40 mEq/L or up to 80 mEq/L as needed) can be provided as half KCl, half KPO4 (to replenish low phosphate levels and to decrease the risk of hyperchloremia) or as half KPO4 and half K acetate (which, like lactate, is converted to bicarbonate to help correct acidosis) after serum K is reported as less than 6 mmol/L or urine flow is established. Serum potassium concentration increases by approximately 0.6 mEq/L for every 0.1 decrease in pH, so that the serum potassium level does not reflect the large deficit from diuresis and vomiting. Both potassium and phosphate shift markedly from intracellular to extracellular compartments with acidosis and reenter cells rapidly with insulin-induced glucose uptake and rehydration.163,171
• Bicarbonate is very rarely indicated. There is no evidence that bicarbonate facilitates metabolic recovery. Restoring circulatory volume will improve tissue perfusion and renal function, increasing organic acid excretion and reversing lactic acidosis. The administration of insulin stops further synthesis of ketoacids and reactivates the Krebs cycle, permitting their metabolism and regeneration of bicarbonate. Bicarbonate therapy may cause paradoxical central nervous system (CNS) acidosis: Rapid correction of acidosis can cause hypokalemia, the additional sodium can add to hyperosmolality, and alkali therapy can increase hepatic ketone production, potentially slowing recovery from ketosis.81 The extremely rare circumstance that may benefit from cautious alkali therapy is severe acidemia, in which tachypneic response is inadequate, or persistent compromised perfusion following initial fluid expansion, which suggests a decreased cardiac contractility and peripheral vasodilatation.81 To avoid the risk of hypokalemia, safe administration is 1 to 2 mEq/kg body weight or 80 mEq/m2 body surface area over 2 hours. The NaCl concentration in the fluids should be reduced to allow for added Na ion.

Insulin

• Insulin should be started after the initial fluid expansion is completed for a more realistic starting glucose level.
• The most widely used system is 0.1 U/kg hourly as a continuous infusion, using a pump. Use 50 U of regular insulin diluted in 50 mL normal saline to provide 1 U/mL.
• A bolus dose is not necessary and may increase the risk of cerebral edema.172
• It may be more convenient in some settings to administer insulin subcutaneously. In adults, there did not seem to be any difference whether insulin was administered intravenously, intramuscularly, or subcutaneously after the first couple of hours of treatment.173-176 The first study of subcutaneous insulin in children with diabetic ketoacidosis using rapid-acting insulin analog (lispro) provided a dose of 0.15 U/kg every 2 hours, finding no essential differences from children randomized to receive 0.1 U/kg per hour intravenously.177 The administration of 0.1 U/kg subcutaneously every hour may be preferable and can be adjusted to maintain blood glucose concentration at approximately 250 mg/dL (11 mmol/L).
• During initial fluid expansion, a high blood glucose level may drop 10 to 15 mM/L, even without insulin infusion. High blood glucose levels should drop 50 to 150 mg/dL/hour (but not > 200 mg/dL/hour), and if they do not, the dose of insulin should be increased. This is rarely necessary.
• If the blood glucose level falls below 150 mg/dL with 10% dextrose solution running, the insulin dose should be reduced to 0.05 U/kg/hour.
• Insulin should not be stopped or reduced below 0.05 U/kg/hour, for a continuous supply of insulin is needed to prevent ketosis and permit continued anabolism.
• Persistent acidosis, defined as bicarbonate value less than 10 mmol/L after 8 to 10 hours of treatment, is usually caused by inadequate insulin effect, indicated by persistent hyperglycemia. Insulin dilution and rate of administration should be checked and a fresh preparation made if it is older than 6 to 8 hours. Too dilute a solution may enhance absorption to the tubing. If insulin is being given by subcutaneous injection, inadequate absorption may be occurring or there may be resistance due to unusually high counterregulatory hormones, as with concomitant febrile illness.165 Extremely rare causes of persistent acidosis are lactic acidosis due to an episode of hypotension or apnea or inadequate renal competency in the handling of hydrogen ion as a result of an episode of renal hypoperfusion.165

Transition

• IV fluids can be stopped l to 2 hours after substantial consumption of oral fluids without vomiting.
• Subcutaneous insulin injection can be started when the patient no longer needs IV fluids. It is most convenient to wait until the presupper or prebreakfast time to restart intermediate- or long-acting insulin. Until then, regular insulin 0.25 U/kg subcutaneously can be given approximately every 6 hours and the insulin infusion stopped 60 to 120 minutes after the first subcutaneous dose of regular insulin or 15 to 30 minutes after rapid-acting insulin analog.
• Patients should not be kept in the hospital simply to adjust insulin dosage, because food, activity, and psychosocial environment are not normal.

Prevention

The prevention of diabetic ketoacidosis (DKA) in children and adolescents requires a diligent health care system. The prevention of DKA at onset is possible when early diagnosis is made through genetic and immunologic screening of high-risk children.89,97 For the larger population, an example has been provided by the Italian school and physician awareness program directed at 6 to 14 year olds, which reduced the rates of DKA from 78% to near 0% over 6 years.92 Recurrent DKA was demonstrated to be dramatically reduced 30 years ago by a comprehensive program involving outreach clinics, frequent routine and emergency telephone contact, and a camping program supported by a state program for children with special health care needs. Private patients in the program had a reduction in hospital admission days from preintervention of 2.8/patient/year to 0.3, and in the second year to 0. The state program–sponsored children had a reduction from 4.9/patient/year to 1.8 and in the second year to 0.9.98 Patients with compliance problems resulting in recurrent DKA account for a disproportionate number of recurrent DKA episodes; in the UK surveillance study, 4.8% of patients accounted for 22.5% of all episodes.106 Insulin omission is the principal immediate reason for the development of DKA in children and adolescents, reflected in low or absent levels of free insulin.164 While psychosocial intervention is of importance, the necessity for assuring administration of insulin by responsible adults is paramount.

Type 1 diabetes (T1D) management poses a daunting challenge to children, families, and health professionals. Not only must the condition be constantly considered in every activity or plan a person makes, but numerous less-than-pleasant tasks must be performed daily, and unexplainable variations occur nonetheless. If complications ensue, after years of being told that good control will prevent them, the person with T1D must deal with guilt and self-recrimination. The fundamental problems in controlling glycemia remain, although services, tools, and insulin have improved the ability to achieve control. The basic problem is the inability to deliver insulin in the physiologic manner, that is, into the portal vein in immediate response to food intake, and turn it off as rapidly. Newer forms of insulin and delivery systems (slow- and fast-acting analogs, pens, pumps) have made the approximation better, but with the requirement that more attention be paid to the diabetes. Until the introduction of self-monitoring of blood glucose and glycated hemoglobin (HbA1c) measurement around 1980, the highly unreliable testing of urine provided little information about the control status, and one relied more heavily on well-being and function in school, work, and play, and growth, which was often inadequate.178-182 The Diabetes Control and Complications Trial demonstrated that the general level of control reached in this circumstance was inadequate to forestall microvascular complications, and the ability of patients to monitor their own blood glucose and physicians to estimate average control was by then standard of care.183

The final and perhaps most vexing limitation is the inability to control human behavior and social circumstances, which have a profound effect on compliance and metabolism. Nonetheless, progress has been made with this limitation, as well. The patient and family are recognized as being the major players on the treatment team, entitled to extensive diabetes and nutrition education, 24-hour access to expert assistance, psychological support as needed (enhanced by extensive psychosocial research in this area), and special programs such as camping.184-186 In sum, contemporary management of T1D is far more demanding of children, their families, and the health care system than before the modern technology era, but infinitely more satisfying for all members of the team. We realized how much children shared their doctors’ frustration with the lack of knowledge in the pre–self-monitoring of blood glucose era when we introduced the then primitive lancets and meters in camp. The prevailing wisdom was that this was a technology that could be applied to children. When asked if they would rather continue testing urine or lance their fingers several times a day to check blood glucose, 100% of the older campers (11 to 14 year olds) and 75% of the younger campers (8 to 11 year olds) preferred finger sticks.185

The Five Es

The five Es of diabetes control are education, emotions, eating, exercise, and enthusiasm/engagement/earnestness. Knowledge is power, but only when the knowledge is used. Each of the factors influences every other. For example, a child who is engaged in sports or otherwise exercising regularly, is likely to be eating better, feel better about herself, be less anxious, and be more enthusiastic about taking care of her diabetes.

Starting Treatment

Learning basic diabetes physiology and management can give the family a sense of control and confidence to cope with the anxiety and despair that accompanies the diagnosis. This is preferably carried out in the outpatient setting, which is much more comfortable for education than is the emotionally highly charged hospital environment, and outpatient management greatly reduces costs.98 Additionally, often the family obtains and reports glucose measurements and immediately follows through on recommendations conveyed over the telephone than occurs in hospital settings. Nevertheless, some children who are not in diabetic ketoacidosis may need to be admitted to the hospital because outpatient management is not available or the social situation is problematic. In such situations, referral to a pediatric diabetes center should be the strategy of choice.

The initial education efforts should be concise and simple, addressing only the most critical needs of checking blood sugar levels, mixing insulin, giving insulin injections, identifying and treating hypoglycemia, and testing for and treating ketonuria. This effort must be accompanied by reassurance that any questions that arise are appropriate for calling in. Over subsequent visits, more information about diabetes and advanced management techniques and goals for glycemic control are progressively discussed.

Treatment usually begins with a combination of neutral protamine Hagedorn insulin and a rapid-acting analog such as aspart or lispro. Most children will require a total insulin dose of 1 U/kg/day. Adolescents are less insulin sensitive and may require 1 to 1.5 U/kg/day. The daily insulin dosage is typically divided into two thirds of the total in the morning and one third in the evening. The beginning morning dose is two thirds neutral protamine Hagedorn insulin and one third rapid-acting analog and the presupper dose one half neutral protamine Hagedorn insulin and one half rapid-acting analog.

The long-acting nonpeak analogs (glargine, determir) are seldom used at the time of diagnosis, except in families with other members having type 1 diabetes. Newly diagnosed children and their families need time to adjust to the treatment, monitoring, and nutritional demands of the basal-bolus regimen and to using twice-daily insulin before being taught to keep count of carbohydrate intake and take 3 or 4 injections daily. Nevertheless, pump administration of insulin and glargine-based regimens are generally superior to neutral protamine Hagedorn use.

Multiple Daily Injection and Continuous Subcutaneous Insulin Injection

Multiple daily injection (MDI) using glargine and mealtime rapid acting insulin analogs or continuous subcutaneous insulin injection (CSII) using insulin pumps are offered to children who have already demonstrated the ability to manage their diabetes somewhat independently and to younger children who have parents who are committed to their care. Several studies have demonstrated the ability of CSII and MDI to improve metabolic control and reduce hypoglycemia frequency in young children and toddlers. Both of these approaches require more frequent blood glucose testing and insulin administration than do fixed doses given 2 to 3 times per day, but many children and parents accept the increased work required because of the lifestyle flexibility and improved control. All regimens will fail unless the child is committed to performing the diabetes tasks necessary to succeed. Thus, before changing therapy, children interested in MDI or CSII should attend a carbohydrate-counting or pump class, demonstrate that they will monitor their blood glucose levels 4 times per day, call weekly with blood glucose results, and record their carbohydrate intake and “virtual” rapid-acting insulin boluses based on a prescribed insulin-to-carbohydrate ratio and correction factor.

Initial glargine doses are calculated as 70% of the total intermediate-acting insulin previously given. Similarly, initial pump rates are usually 70% to 80% of the previous total daily amount of subcutaneous insulin, with 60% of that given as a basal dose divided over 24 hours. Meal coverage is determined by using insulin (units)-to-carbohydrate (grams) ratios. Ratios in children may vary from 1:30 for young children to 1:5 for adolescents at the peak of the physiologic insulin resistance of puberty.

Blood Glucose Monitoring

Blood glucose levels should be monitored at least 4 times per day, before breakfast, lunch, and dinner, and at bedtime. Intensive insulin therapy (multiple daily injections, insulin pump) also requires checking before snacks so that appropriate doses of rapid-acting insulin can be taken. More frequent monitoring throughout the day and during the night may also be needed when starting pump use or at times of other major management changes. In order to determine if the rapid-acting insulin dose is correct, 2-hour postprandial blood glucose levels are encouraged to assess the effectiveness of the insulin-to-carbohydrate ratio. If a child is waking up with high blood glucose levels, the caregiver needs to obtain a 2 or 3 am blood glucose level before adjusting the evening insulin dose upward to be sure that the child is not becoming hypoglycemic in the middle of the night and having rebound hyperglycemia in the morning: In this case the dose would need to be reduced instead of increased. It is also necessary to check the blood glucose levels more frequently during illness.

The various meters available all have memory for numerous readings and associated information and can all download data to the computer with specific programs from each manufacturer. Some can be downloaded from the Internet at no cost whereas others require software purchase. Meters may be downloadable with an infrared port or require a cable connection. Meters manufactured by the same company may require different cables for download. Reports generated include a logbook for a specified period, data list with all results, statistics, averages, pie charts, line graphs, and highest and lowest readings. Most useful in the clinic is the 14-day logbook and statistics. The 14 day logbook also includes averages for each time of day. The average number of blood glucose tests per day, which correlates with control, and the blood glucose ranges and averages, are recorded in the clinic note.

Patients and caregivers are encouraged to keep a logbook with blood glucose levels, insulin doses, and notes detailing events that may affect blood glucose readings, and to bring both the logbook and the meter for download to the clinic visit. Unless a daily or weekly record is kept, it is unlikely that patients and caregivers will recognize developing blood glucose patterns. While a meter can record the date, time, blood glucose reading, and with some meters insulin dose, amount of carbohydrates eaten, or exercise done, it provides less information and does not provide the immediate feedback that recording these events in a logbook does. The additional information recorded in the logbook, for example, might be that the patient ate pizza and ice cream for dinner and ended up with hyperglycemia at bedtime, or ate out and underestimated the carbohydrates eaten with hypoglycemia 1 hour after finishing the meal, or that the patient forgot to take a bolus for a midafternoon snack and had hyperglycemia before dinner. Logbooks are provided by the meter manufacturers.

Blood Glucose and Hba1c Goals

HbA1c is the hemoglobin fraction that is nonenzymatically glycated and therefore reflective of average blood glucose over the average half-life of the red cell hemoglobin, that is, 60 days. HbA1c goals must be age adjusted in children and adolescents to minimize the danger of serious hypoglycemia; fortunately, the risk for long-term complications is much less related to duration of diabetes before 15 years of age than after.187 Children who have recurrent hypoglycemia at an early age (< 5 years old) may be at increased risk for long-term neuropsychological impairment.188,189 Glycemic goals for infants and toddlers are 100 to 180 mg/dL (5.6–10 mmol/L) before meals, 110 to 200 mg/dL (6–11 mmol/L) overnight, and the HbA1c goal between 7.5% and 8.5%. School-aged children (6–12 years) should achieve better control, with HbA1c less than or equal to 8%.

In adolescents, the problems of adolescent rebellion and independence require a more individualized approach. Premeal blood glucose levels should be 90 to 130 mg/dL (5–7.2 mmol/L), and the goal should be to maintain a HbA1c value of less than 7%. This goal is difficult for most teens to achieve, however, and the ADA has recommended a HbA1c goal of 7.5% for this age group, with an ideal HbA1c value of less than or equal to 7% if achievable without hypoglycemia.

Nutrition

Children with diabetes should be encouraged to maintain a normal, healthy, age-appropriate diet, which includes avoiding concentrated sweets and excess calories. Daily caloric requirements for children can be estimated by using the formula 1000 kcal plus 100 kcal per year of age. Most children should obtain their caloric intake from a balanced diet containing 50% to 60% carbohydrates, 30% fat, and 10% to 20% protein. Fat should be limited to less than 30% of caloric intake if the child is overweight, with saturated fat less than 10% of daily calories and less than 7% of calories if the child has elevated low-density lipoprotein (LDL) cholesterol. Finally, cholesterol intake should be limited to 300 mg/day; if the child has elevated LDL cholesterol, this should be less than 200 mg/day.

Carbohydrate counting is important for children using multiple daily injection or continuous subcutaneous insulin injection. They may require simple carbohydrate ratios to minimize confusion, and many children require postprandial glucose monitoring to determine carbohydrate ratios for particular meals. Food should never be used to reward or punish children, especially those with diabetes.

Age-Specific Management Issues

Infants

Caring for an infant with diabetes can induce considerable emotional distress. It is difficult to traumatize infants and young children with multiple blood glucose level checks and insulin injections because the parent is unable to explain the reasons for the finger sticks and injections. Fortunately, infants adapt quickly to glucose monitoring and insulin administration. In addition to the anxiety of having an infant with diabetes, parents also have to cope with unpredictable eating and play activity level. They may need to give insulin after meals, based on intake, and this is now practical with the rapid-acting analogs.

Toddlers

Toddlers are able to scream, kick, bite, or make hurtful statements to caretakers who are giving insulin or checking blood glucose levels. As are infants, they remain incapable of understanding the need for these procedures. Thus, the parent or caregiver must carry out diabetes care tasks in a matter-of-fact, nonemotional way and provide enthusiastic praise when the toddler is cooperative. The combination of positive reinforcement and a matter-of-fact delivery of diabetes care will teach children that their cooperation is rewarded. Consistency and firmness are critical in teaching the child that diabetes care needs are nonnegotiable and are not subject to argument.

Elementary School Age Children

In the elementary school-age child, school absenteeism as the result of manipulative use of diabetes-related complaints can become a serious problem. Parents often worry about sending their child with diabetes to school when the child may actually be ill. Children should not be allowed to miss school unless there is good evidence of an illness, such as fever or vomiting.

As diabetes care has become more intensive, with blood glucose monitoring and insulin administration required during the school day, it is essential that school personnel be trained to handle diabetes-related tasks. Many schools do not have a nurse on site at all times, and educators and other school personnel often do not feel comfortable or competent dealing with the medical needs of the child. Nevertheless, parents and physicians must advocate and help educate school officials, and in cases in which a nurse is not available, teachers or coaches can be taught to provide basic diabetes care. Older children—those in middle and high schools—are usually developmentally able to check their blood glucose levels and administer insulin boluses independently. Diabetes should not keep children from participating in any school activities, including after-school sports. Conversely, school should not be a barrier for children’s ability to care for their diabetes.

Children with diabetes are covered under Section 504 of the Rehabilitation Act of 1973, the Individuals with Disabilities Education Act of 1991, and the Americans with Disabilities Act. Public schools, day care centers, and all schools that receive federal funding fall under the jurisdiction of these acts,190 requiring that the institutions provide adequate accommodation for children with diabetes. Parochial schools are exempt. (A sample 504 plan and an excellent book about diabetes in the schools are available on the National Diabetes Education Program (http://www.ndep.nih.gov/resources/resources.htm) and American Diabetes Association (http://www.diabetes.org/uedocuments/ad-504-adanasndredf-2007.pdf. and http://www.diabetes.org/advocacy-and-legalresources/discrimination/school/schooltraining.jsp) web sites. Thus, schools must be equipped to deal with children with diabetes. Children must be allowed to check blood glucose levels, give supervised insulin injections, and treat hypoglycemia in close proximity to the classroom. It is preferable for the school nurse to be familiar with diabetes management in order to oversee the child’s diabetes management. Many schools do not have a nurse, or the nurse may not be comfortable with diabetes management. In these circumstances training must be provided to a specified diabetes liaison. Teachers should also be aware of the symptoms and treatment of hypoglycemia, and fast-acting carbohydrates (eg, juice or insta-glucose) should be available. A glucagon emergency kit should also be kept in the nurse’s station. The child’s parent is responsible for providing all diabetes supplies.

In most schools, children must go to the school nurse’s station to check their blood sugar levels, administer insulin, and treat hypoglycemia. An aide or responsible classmate must accompany the child, especially if there are signs of hypoglycemia. School nurses or the diabetes liaison should be provided with specific protocols for treating hypoglycemia, checking for ketone levels, and administering insulin. The health care team is responsible for providing the school personnel with an individual diabetes management care plan.

If children are taking a combination of neutral protamine Hagedorn insulin and rapid-acting analogs, insulin injections at school are not usually required. The increasing number of children using continuous subcutaneous insulin injection and multiple daily injection, however, requires that school nurses or the diabetes liaison oversee lunch-time insulin injections. Parents and physicians must be sure that nurses or diabetes liaisons are aware of the child’s specific blood glucose goals and are appropriately trained to either administer or oversee insulin injections. Physical fitness is important for children with diabetes to maintain cardiovascular health and improve metabolic control. Ready access to fluids and carbohydrates is essential. If physical education is expected to last more than 30 minutes, a 10% to 20% reduction in the insulin dose or a juice box or small snack can be used to avoid hypoglycemia. Glucose testing should be made available at the site and should be allowed any time the child feels symptoms of hypoglycemia. The symptoms of hypoglycemia are clinically difficult to distinguish from those of stressful exercise, so it is better to err on the side of unnecessarily treating the symptoms as hypoglycemia, with carbohydrate ingestion, than run the risk of worsening hypoglycemia if glucose measurement cannot be immediately done. Full participation in physical education should be expected. Creating an exercise program for children with diabetes that provides for appropriate monitoring and the expectation that they participate fully teaches children that diabetes should not exclude them from any activities.

School-aged children begin to understand diabetes care and may want more independence in managing their diabetes. Although they may handle diabetes management tasks well, they still require supervision by an adults.

Adolescence

Poor adherence is a major problem for adolescents, who are usually managing their own care, during a period of development characterized by rebellion, rejection, and denial of vulnerability. Not all adolescents are ready for the responsibility of self-care, however, and all require some degree of supervision. Risk-taking behaviors such as smoking, alcohol and other substance use, sexual activity, and careless driving need to be addressed. Menstrual irregularities may be a sign of poor glycemic control. Discussion of contraception and pregnancy risks should be included in every adolescent office visit. Oral contraceptives have clinically insignificant effects on glucose tolerance, so that there is no reason to avoid their use in girls with type 1 diabetes.

The characteristic stresses of adolescence associated with social pressures, schoolwork, and puberty add an uncontrollable variable to the management of diabetes. The physiologic insulin resistance of adolescence, likely the result of increased growth hormone secretion, may be further intensified by elevations of cortisol, catecholamines, and sex steroids, resulting in increased insulin requirements (up to 1.5 to 2 U/kg/day) and extended periods of hyperglycemia and elevated HbA1c values during periods of stress and menses. Some pediatric endocrinologists use metformin as an adjunct therapy in adolescents with type 1 diabetes to enhance insulin sensitivity.

Sick Day Management

Sick day management, designed to prevent or treat early ketoacidosis (diabetic ketoacidosis) during illness or stress, requires repeated reinforcement with caregivers and adolescents. Provision of written instructions is important because of the infrequency of the situation. A common mistake is to skip insulin injection if the child refuses to eat or vomiting is present. The notion of continuing to take insulin despite no food intake is counterintuitive for many parents; in fact, the insulin requirements are typically increased because of the stress of illness, resulting in counterregulatory hormone-induced gluconeogenesis and ketogenesis. Blood glucose levels should be checked more frequently than usual, at least every 2 to 4 hours. Urine ketones should be checked often, as well. If the child is sick, ketones should be monitored even if the blood glucose level is low or normal. If vomiting occurs, ketones should be checked each time the child urinates, with repeated testing at least every 2 to 3 hours. Extra insulin may be necessary for the insulin resistance accompanying the illness or if ketones are present.

For negative, trace, or small ketonuria, increasing water intake, exercise, or insulin supplementation can be tried. For moderate or large ketonuria, extra insulin should be administered based on the total daily dose of insulin.

For children using multiple daily injection, temporary adjustments of the insulin/carbohydrate ratio, target blood glucose levels, and doses of the long-acting and/or rapid-acting insulins may be necessary.

Those using continuous subcutaneous insulin injection can program a temporary basal rate, usually a 20% to 30% increase for 3 to 4 hours.

While most children become hyperglycemic with illness despite not eating, some will have a decrease in glycemia under these conditions. We provide the following guide to our patients for appropriate fluids for various blood glucose ranges.

• If the blood sugar is below 120 mg/dL:
• Fluids need to contain extra glucose, such as soda or a sport drink with sugar added. The child should continue with glucose-containing drinks until the blood sugar is above 120 mg/dL.
• If the blood sugar is between 120 and 180 mg/dL:
• Fluids need to contain approximately 15 g of carbohydrates, such as one half cup of ginger-ale or cola, one half cup of regular Jell-O®, one half cup of juice, or one half of a popsicle.
• Foods need to have at least 15 g of carbohydrates per serving. If the child cannot eat normal meals, possible “sick day” foods include: one half cup of ice cream or sherbet, 1 slice of toast, or 6 saltines.
• If blood sugars are above 180 mg/dL:
• Fluids need to have no calories, such as diet soda, unsweetened tea or water.

Foods need to be small servings as tolerated. If the child is nauseated or vomiting, give small sips of clear fluids (such as ice chips, popsicles, clear broth or bouillon, gelatin, water, or tea). Avoid orange juice, milk, or milk products. Speak with your diabetes team about fluid replacement and advancing to foods. Nutritional needs and insulin doses can change hour by hour depending on blood glucose and urine ketone levels.

An over-the-counter product to control nausea, such as Emetrol, may prevent vomiting. Regular cola, ginger ale, or other soda with the bubbles shaken out may also be helpful. Vomiting is the signal that the diabetes team or primary physician needs to be called. If the child has not been ill for more than a few hours and has not been vomiting for more than couple of hours, a Tigan or Phenergan suppository may control the vomiting and permit oral fluid intake, thus avoiding admission to the emergency room or hospital for rehydration.

Diabetes cAMP

Diabetes camps provide a unique opportunity to educate and socialize children with diabetes. Children who are initially reluctant to attend a diabetes camp often leave camp with a new sense of acceptance of their diabetes and renewed interest in maintaining near-normal blood glucose control. Most diabetes camps are residential camps staffed by experienced pediatric diabetes nurses, psychologists, and physicians. Strict observation of blood glucose monitoring and insulin administration is required to assure maximum vigor for camp activities and prevent ketosis/ketoacidosis.191 Daily insulin adjustments must be made, as well, for most children are much more active at diabetes camp than at home, increasing the risk for hypoglycemia. Local Juvenile Diabetes Research Foundation and American Diabetes Association contacts typically have the best information on diabetes camps in the area.

Long-Term Complications

The Diabetes Care and Complications Trial and the longitudinal follow-up of these patients, the Epidemiology of Diabetes Interventions and Complications study, demonstrated that tight glycemic control is critical in reducing both microvascular and macrovascular complications of diabetes.183,192 Self-monitoring of blood glucose, HbA1c measurements, and improvements in insulins and delivery systems have made it possible to attain a degree of glucose control that will reduce the risk of long-term complications. The ability to normalize blood glucose is now limited only by the frequency and severity of hypoglycemia.193 Microvascular disease results in retinopathy, nephropathy, and neuropathy. Macrovascular complications affecting coronary, carotid, and other major vessels are now recognized as having their origins in childhood. Of particular interest in children are the effects of diabetes on growth and development and associated connective tissue complications.

Joint Problems

Limited joint mobility (LJM) is the earliest clinically apparent long-term complication of type 1 diabetes (T1D) in childhood and, following its initial description in 1974, was recognized to be a common feature of pediatric T1D.180 The initial description was of bilateral, painless, but obvious, contracture of the finger joints and large joints, with thick tight waxy skin and short stature in 3 older teenagers with long-standing diabetes and early microvascular complications.194 Milder expression was subsequently recognized as affecting approximately 30% of youngsters with diabetes, a third of whom had involvement of more than 2 fingers or 1 finger and 1 large joint bilaterally.180Variation in prevalence among various reports is related to the examination technique, the criteria, the age of the patients, and duration of diabetes. In older adult patients, examination is confounded by the presence of finger contracture related to age, occupation, and other problems affecting the joints more commonly in those with diabetes, such as Dupuytren disease, flexor tenosynovitis, or carpal tunnel syndrome.195

LJM is painless and only mildly disabling when severe (eFig. 544.2). A simple examination method is to have the patient attempt to approximate palmar surfaces of the interphalangeal joints. Passive examination is essential to confirm that inability to do so is due to LJM. Attained age seems more important than duration of diabetes or age of onset of diabetes in the evolution of LJM. With rare exception, LJM appears after the age of 10 years. The biochemical basis for LJM is likely glycation of protein with the formation of advanced glycation end products. This results in increased stiffness of the periarticular and skin collagen, with decreased range of motion (eFig. 544.2).

As might be expected, the presence of LJM is associated with increased risk for the development of microvascular disease. The initial report of the prevalence of LJM included 7 patients with severe changes, of whom 5 had clinically apparent retinopathy before the age of 18 years.183 In a longitudinal study of 169 patients with diabetes duration greater than 4.5 years, which was the shortest duration at which complications were noted, LJM was associated with an 83% risk for retinopathy, while the absence of LJM was associated with a 25% risk after 16 years of diabetes. The severity of microvascular changes also corresponded to the severity of LJM.196 Attempts to correlate the development of LJM with diabetes control as measured by HbA1c levels were unsuccessful in cross-sectional studies, but a longitudinal study of average HbA1c from the first year after diagnosis demonstrated unequivocally that the appearance of this complication was significantly related to diabetes control.197

Both the prevalence and severity of LJM have decreased since the initial recognition in the 1970s, most likely the result of improved glucose control during this era.198 These study results are pertinent to the question of the importance of prepubertal duration and control of diabetes for the development of long-term microvascular complications. Even for the youngest age group studied or those with the shortest duration of diabetes, there was a marked decrease in the prevalence of LJM between the 2 eras, suggesting that improved control from the first years of childhood diabetes will reduce long-term complication risks.

Skin Problems

Necrobiosis lipoidica diabeticorum is a rare pretibial lesion that is characterized by round or oval indurated plaques, with central atrophy and eventual ulceration. Unless they become infected, the lesions are painless. Relationship to diabetes control is uncertain, but there is an association with smoking, proteinuria, and retinopathy.199,200 The lesions do not heal well and need to be protected from trauma and treated vigorously if infected.

Lipoatrophy, indentation, or atrophy of the subcutaneous tissue resulting from insulin injection has become uncommon with the generalized use of recombinant derived human insulin. It is presumed that the impurities in animal extract insulin or the foreign nature of that insulin led to the problem. It was thought to be immune mediated, frequently following dermatomes and occurring symmetrically, distal from the actual injection sites. The lesions tended to resolve after a few years and were thought to be IgE mediated. Mild atrophy may still be seen occasionally with contemporary insulin preparations, affecting approximately 3% of patients 1 to 19 years old201 and has been reported with the use of insulin analogs.202,203 Injection site rotation may help decrease the risk of developing lipoatrophy. When lipoatrophy develops, giving insulin injections around the edges of atrophic areas or using an alternative insulin analog may be helpful.204

Lipohypertrophy, in contrast to lipoatrophy, always occurs at the site of injection as a local reaction to repeated injection trauma, usually reflecting favorite site injection, with scarring and decreased sensitivity to pain.205 Unlike lipoatrophy, lipohypertrophy continues to be a frequent problem with improved insulin forms, affecting as many as half of patients under 20 years of age.206 Both lipoatrophy and lipohypertrophy are associated with greater insulin antibody concentrations.201 Varying absorption of insulin from areas of hypertrophy can result in erratic glycemic control.206 Site rotation is also important in reducing the risk of developing lipohypertrophy. Once lipohypertrophy develops, injections into the area of induration should be discouraged. If injection of the affected area is avoided, most lipohypertrophy will resolve within 6 to 8 weeks. However, some areas of lipohypertrophy become permanent, even with dedicated avoidance of injections to the area.

Cataracts

Cataracts occur rarely during recovery from ketoacidosis, typically in the newly diagnosed patient; they usually disappear rapidly but may persist, requiring surgical removal.207 In contrast to these “sugar” cataracts, subcapsular juvenile cataracts, associated with chronic hyperglycemia and caused by sorbitol accumulation, do not regress and always require removal.

Growth Failure and Delayed Sexual Maturation

Insulin sufficiency is necessary for normal growth and development. During the nearly 6 decades between the discovery of insulin and the contemporary era of diabetes management, growth retardation and delayed puberty were frequently seen in children with diabetes. In twin pairs discordant for diabetes, the development of diabetes before the onset of puberty was associated with invariably shorter stature than that of the nonaffected twin.208 Nonetheless, despite the generally nonphysiologic traditional means of treating diabetes, most children with diabetes did not have serious growth or maturational delay problems. Most children with diabetes now experience normal timing of puberty, with normal growth and final adult heights.209,210

The effects of insulin on normal growth and development are largely mediated through the anabolic pathway and the GH-IGF-1 axis. Deficiency of portal insulin delivery to the liver results in reduced circulating growth hormone (GH) binding protein (the extracellular domain of the cell surface receptor for GH), GH resistance, and diminished IGF-1 production.211,212 The deficiency in insulin also results in decreased production of IGFBP-3, the principal binding protein for circulating IGF-1, and increased production of IGFBP-1 and IGFBP-2.213 These latter binding proteins, unlike IGFBP-3, do not deliver their bound IGF-1 to tissues. Furthermore, increased proteolysis of IGFBP-3 decreases its viability in the circulation. In addition to the GH resistance, which is reflected in the lower concentrations of circulating GH binding protein, there may be postreceptor defects in GH action mediated by the insulin deficiency. Diminished circulating total and free IGF-1 levels are associated with GH hypersecretion because of the absence of negative feedback from IGF-1. This hypersomatotropism increases glycemia and decreases insulin sensitivity, but is reversible with adequate insulin replacement.214

The Mauriac syndrome consists of growth failure, maturational delay with hepatomegaly, and abdominal distention in insulin-treated children.215 Hepatomegaly was apparently a common complication in children in the era when only short-acting insulin was available and aglycosuria was the objective. This was exemplified in a series from the Joslin Clinic in the late 1930s of 60 youngsters with hepatomegaly, growth failure, delayed sexual maturation, and severe uncontrollable diabetes with frequent hypoglycemia and ketoacidosis.216 Improvement was associated with the change to long-acting protamine zinc insulin. There appear to be 2 forms of the Mauriac syndrome. In 1 form there is associated Cushingoid obesity and documented wide fluctuation between hyperglycemia and hypoglycemia, suggestive of a pattern of over- and underinsulinization, which would be expected with treatment using only soluble insulin, with secondary hyperadrenalism.217 Periods of overinsulinization would appear to be essential for the development of obesity in this form of the syndrome and for the induction of hyperadrenalism.217 More recent reports are of patients with Mauriac syndrome who do not have obesity and who are without a history of alternating hypoglycemia and ketoacidosis; these patients are unmistakably inadequately insulinized continuously.218,219

Eating disorders are also more common in children with diabetes than in the general population and should be considered in the assessment of any child who is not gaining weight or growing normally.220-223 Adolescent females with diabetes are twice as likely to have eating disorders than are those without diabetes. Mean HbA1c levels were higher in adolescents with diabetes who had an eating disorder than in those who did not.224

Autoimmune Disorders

Associated autoimmune problems, such as hypothyroidism and celiac disease, can result in growth failure in children with type 1 diabetes (T1D). Approximately 20% of children with T1D have antibodies indicative of autoimmune thyroid disease.220 Anti-thyroid autoantibodies should be tested for in all youngsters with T1D at diagnosis, and those who are positive should have annual determination of TSH levels. It has recently been recommended that all children with T1D have TSH testing every 1 to 2 years and any time thyromegaly is noted or clinical thyroid dysfunction is suspected.221 If TSH is abnormal, free T4 should be measured.

Because celiac disease has been described in 5% to 8% of children with T1D, tissue transglutaminase antibodies should be sought in children with growth failure without other obvious cause, even in the absence of diarrhea.221 Children should be tested soon after T1D diagnosis for celiac disease, using tissue transglutaminase antibodies and serum IgA levels. Retesting should be done if growth failure, failure to gain weight, weight loss, or gastrointestinal symptoms occur. Small bowel biopsy is required to confirm the diagnosis. Management is with a gluten-free diet.222

Microvascular Complications

Thickening of the capillary basement membrane with alterations in membrane permeability and occlusion of small blood vessels is the hallmark of diabetic microvascular disease. Although some of the long-term complications of T1D discussed above are likely partially due to microvascular disease, renal disease and retinopathy are more clearly linked. Several mechanisms have been proposed to explain the pathophysiology of these complications, all with the basic concept of hyperglycemia-induced tissue damage that appears to result from the increased production of reactive oxygen species.225

1. 1. Glycemic control is directly related to the development and progression of microvascular disease.183 This suggests that excess glucose per se has deleterious effects on tissues. Some of these effects appear to be through binding to collagen by posttranslational nonenzymatic means, depending directly on glucose concentration.

   2. Sorbitol is the alcohol of glucose that is formed from glucose by aldose reductase, the magnitude of the reaction driven by the concentration of blood sugar leading to sorbitol accumulation in a number of tissues, including lens and nerve, which leads to increased permeability and increased pressure and leakage in blood vessels and decreased vessel wall elasticity and microvascular hypertension.226

   3. Intracellular hyperglycemia also increases the synthesis of diacylglycerol, a key activator of protein kinase C, activation of which has multiple effects on gene expression that result in inflammation, increased reactive oxygen species, and vascular dysfunction.225

   4. As described above, diabetes is associated with increased fibrinogen levels and elevations in other antithrombotic factors that may result in increased platelet stickiness, resulting in increased red blood cell aggregation and blockage of small capillaries.227

Nephropathy

Microalbuminuria, defined as urinary albumin excretion rates of 20 to 200 μg/minute or 30 to 300 mg/24 hours, is predictive of renal and cardiovascular disease.228 Higher concentrations are considered to reflect overt diabetic nephropathy. HbA1c level, LDL cholesterol concentration, and body mass index were all risk factors for the development of microalbuminuria in 12.6% of 1,134 men and women with type 1 diabetes aged 15 to 60 years during the 7-year follow-up in the EURODIAB Prospective Complications Study.229

A study of the prevalence and progression of microalbuminuria and clinical proteinuria in 361 children with type 1 diabetes over the course of 12 years found a very low incidence of microalbuminuria. Single episodes of microalbuminuria increased between the ages of 10 and 18 years (30.9% of males and 40.4% of females), but persistent microalbuminuria and progression from microalbuminuria to macroalbuminuria was rare.230

Serum vascular endothelial growth factor concentration may identify patients who are predisposed to develop persistent albuminuria.231 Eleven percent of 101 children and adolescents with diabetes who initially had normal urine albumin excretion developed persistent microalbuminuria during 8 years of follow-up, with an odds ratio of 4.1 for the occurrence of microalbuminuria if vascular endothelial growth factor serum levels were elevated.

Other predictors of persistent albuminuria have included elevated HbA1c and high baseline albumin excretion rates. reported in 12.8% of 339 Danish children and adolescents with type 1 diabetes (T1D) studied over a 6-year period.232 This incidence was higher than in a large French cross-sectional survey of 702 children and adolescents with average diabetes duration of 7.6 years. In this study microalbuminuria, defined as an albumin excretion rate greater than or equal to 15 μg/minute measured in at least 2 of 3 urine collections, increased from 5.1% to 11.6% after 10 years of diabetes duration and completion of puberty; maternal hypertension was a significant risk factor, but patients’ blood pressures were not, and HbA1c had only a marginal effect.233 These findings contrast with other studies in which there is strong association between progression of microalbuminuria and HbA1c values. A 4-year follow-up of 279 patients with T1D found that the rate of progression was 1.3/100 person years in those with HbA1c levels less than 8%, whereas it was 5.1/100 person years with HbA1c levels between 8% and 9%, 4.2 between 9% and 10%, and 6.7/100 person years if HbA1c was greater than 10%. The risk of progression rose steeply between HbA1c values of 7.5% and 8.5%, emphasizing the importance of improved glycemic control.234

The Pittsburgh Epidemiology of Diabetes Complications study is an observational prospective study of 589 patients with onset of T1D before the age of 17 years. Baseline LDL levels greater than or equal to130 mg/dL (3.36 mmol/L), triglycerides greater than or equal to 150 mg/dL (1.7 mmol/L), systolic blood pressure greater than or equal to 130 mm Hg, and diastolic blood pressure greater than or equal to 85 mm Hg conferred relative risks of 2.2, 3.2, 2.3, and 2.5, respectively, for developing nephropathy over 10 years.236

Homocysteine elevation has been identified as a risk factor for premature cardiovascular disease. Children with T1D diagnosed before age 12, with diabetes duration longer than 7 years and who had albumin excretion rates greater than 70 μg/min, had elevated homocysteine levels. This suggested an association of diabetic nephropathy with premature cardiovascular disease.236 Diabetes management improvements over the last 30 years have markedly decreased the incidence of severe nephropathy in adults who were diagnosed with T1D as children.237,238

Annual testing for albuminuria is recommended for children from age 10 who have had diabetes for 5 years.211 The most convenient and reliable way to do this is to measure urinary albumin-to-creatinine ratio in a urine sample. Microalbuminuria in a single specimen may not indicate disease. Repeat testing is required to confirm persistent microalbuminuria. Orthostatic proteinuria is common and must be distinguished from that due to diabetic renal disease. A first-void morning urine should not contain abnormal concentration of albumin, but a late afternoon urine specimen often will if the albuminuria is orthostatic.

Although the Diabetes Care and Complications Trial demonstrated that microalbuminuria and overt nephropathy can be delayed or prevented by intensive diabetes treatment,183,239 the United Kingdom Proscriptive Diabetes Study and other studies have shown that blood pressure control is at least as important as glycemic control in prevention of nephropathy and in decreasing the rate of decline of glomerular filtration rate in adults with T2D.240 Maintaining blood pressure at normal values is of paramount importance in T1D. Loss of the nocturnal decline in blood pressure is often the first sign of hypertension, making ambulatory blood pressure monitoring a useful adjunct in evaluating children with borderline hypertension.

Persistent microalbuminuria should be treated with an angiotensin-converting enzyme (ACE) inhibitor titrated to maintain normal albumin excretion. ACE inhibitors increase afferent vessel dilatation and decrease elevated intraglomerular pressures, resulting in decreased albumin excretion. Although children and adolescents with microalbuminuria have higher blood pressures than those with normal albuminuria,241-244 the beneficial effects of ACE inhibitors are observed whether or not the patient has hypertension.245 Many pediatric diabetologists start therapy with enalapril 5 to 10 mg daily and titrate upward to achieve normal albuminuria and blood pressure.

The role of low-protein diets in treatment of microalbuminuria has been debated for several years; there are no long-term studies in children. Protein restriction slows the rate of decline of glomerular filtration rate246 and decreases albumin excretion rates in adults with microalbuminuria.247 A low-protein diet of 0.6 to 0.8 g/kg body weight per day may be helpful in decreasing albuminuria and slowing the decline in renal function in patients with persistent microalbuminuria. Although high protein intake should be avoided in children with microalbuminuria, protein restriction below 15% of total daily calories is not recommended because of the need for adequate dietary protein in the growing child.

Adolescents should be counseled to avoid smoking, as nicotine has profound vasoconstrictive effects that have been associated with progression of nephropathy in adults with T1D.248

Retinopathy is the most common microvascular disease seen in children and adolescents with T1D. Retinopathy is broadly defined as (1) background retinopathy, with the presence of microaneurysms only or microaneurysms and occasional dot-blot hemorrhages, and (2) proliferative retinopathy, with growth of new vessels, glia, and fibrous tissue. Fluorescein angiography can document severity of retinopathy by showing microaneurysms, areas of retinal ischemia, proliferation of new blood vessels, and permeability of the retinal vessels.

At diabetes diagnosis, there is increased blood flow through the retina that may be associated with leakage of injected fluorescein into the vitreous,249 which is reversed by improved metabolic control. In addition, rapidly improving metabolic control may be associated with an initial worsening of diabetic retinopathywith, however, long-term improvement.221 Anatomic disease results from increased vascular permeability and capillary and arteriolar occlusion within the retina. The increased permeability results in edema and the formation of hard exudates, whereas the occlusion results in formation of microaneurysms, hemorrhages, soft exudates and capillary dropout. The growth of new blood vessels, glia, and fibrous tissue in front of the retina can result in vitreous hemorrhage and retinal detachment from shrinkage of the proliferating tissue.

Although retinopathy has been reported to be present early in the course of childhood diabetes,250 it usually is not recognized before 5 to 10 years of diabetes duration and after the onset of puberty. By that time, 20% to 30% of patients will have background retinopathy, 30% to 50% after 10 to 15 years diabetes duration, and 70% to 80% after 15 years of diabetes.251

In the EURODIAB prospective complications study of 764 patients greater than 15 years of age, retinopathy incidence was 56% after 7 years of follow-up. Risk factors included high albumin excretion rate, elevated cholesterol and triglyceride levels, and increased waist-to-hip circumference ratio. In contrast to other studies, there was no association between retinopathy and blood pressure, cardiovascular disease, or smoking.252

Among 725 African Americans with type 1 diabetes in New Jersey, 3% had macular edema and 20.6% had retinal hard exudates. The severity of retinopathy correlated with proteinuria, higher LDL cholesterol levels, systolic hypertension, poor glycemic control, and duration of disease. Patients with HbA1c values in the highest quartile were 3 times more likely to have retinopathy than those in the lowest quartile. Patients with renal disease were 3 times more likely to have retinopathy and 10 times more likely to have proliferative retinopathy than those without renal disease, and patients in the highest quartile of systolic blood pressure were 3 times more likely to have proliferative retinopathy than patients in the lowest quartile. Systolic blood pressures above 120 mm Hg conferred increased risk of proliferative retinopathy.253

The Pittsburgh Epidemiology of Diabetes Complications found that systolic blood pressure (BP) greater than or equal to 120 mm Hg conferred a relative risk of 1.6 for the development of proliferative retinopathy, and the relative risk was 2.7 if systolic BP was greater than or equal to 130. Diastolic BP greater than 79 mm Hg conferred a relative risk of 1.8, if diastolic BP was greater than or equal to 85 relative risk was 2.4, and if diastolic BP was greater than or equal to 90 mm Hg, the relative risk of proliferative retinopathy was 4.6.236

Homocysteine concentrations were higher in adolescents with proliferative retinopathy, further indicating a link between microvascular complications and later onset of macrovascular disease and implicating a common pathologic mechanism in large and small vessels.237

An ophthalmologic evaluation at the time of diagnosis and after stabilization may be useful to assure that no eye disease or uncorrected visual limitation is present. Annual ophthalmologic evaluation should begin after age 10 and following 3 to 5 years of diabetes.221 Recent studies indicate, however, that biennial screening may be adequate in well-controlled children.254 ACE inhibitors slow the progression of retinopathy in patients with and without hypertension; thus, retinopathy is an indication for ACE inhibition.255 ACE inhibitors are contraindicated in pregnancy.

Retinal photocoagulation has markedly improved the prognosis for vision in patients with proliferative retinopathy.256 The photocoagulation destroys ischemic retinal tissue, thus removing the stimulus for neovascularization. Laser therapy is also used for macular edema by destroying leaking microaneurysms and dilated capillaries. In cases of proliferative retinopathy, panretinal photocoagulation will destroy retinal tissue outside the macula and optic nerve so that the remaining retina is not ischemic and will, therefore, not produce the growth factors that result in proliferative diabetic retinopathy. The panretinal photocoagulation has the adverse effects of compromising color discrimination, visual fields, and night vision. In the case of vitreous hemorrhage or detached retina, vitrectomy can be used to remove the blood from the vitreous or to reattach a retina that is in the process of detaching. Laser photocoagulation should be considered immediately upon diagnosis of proliferative retinopathy to prevent additional visual loss and prolong the period of useful vision.221

Neuropathy

Although symptomatic neuropathy is uncommonly seen in children and adolescents with diabetes, sensory and autonomic motor nerve impairment can be demonstrated in young people. Peripheral neuropathy, including motor and sensory disturbances, as well as autonomic neuropathies (gastrointestinal, cardiovascular, vasomotor instability, and hypoglycemia unawareness) can be seen in children and adolescents with type 1 diabetes. Peripheral sensory function testing should be done at each clinic visit by gently bending a 10-g monofilament against the great toe. Loss of ability to feel the monofilament requires further evaluation.

The pathogenesis of neuropathy in diabetes involves both neural and vascular components. The aldose reductase system in the lens and peripheral nerves is responsible for metabolism of glucose to sorbitol with elevated glucose levels accelerating the synthesis of sorbitol, which is not freely diffusible, exerting an osmolar affect within the nerve and causing swelling and eventual destruction. Myoinositol, a component of neuronal cell membranes involved in the control of neural transmission, is decreased in animals and humans with diabetes. The basement membrane of the Schwann cell is thickened in diabetes, and the vasa nervorum show capillary and arteriolar wall thickening as the result of basement membrane accumulation. This can be considered another microvascular complication of diabetes. Numerous thrombosed vessels, axonal loss, and a characteristic segmental demyelination are demonstrated on biopsy of affected nerves.

Symptoms of peripheral neuropathy include bilateral numbness and paresthesias, especially pain and burning in the lower extremities, which is much worse at night. A decrease in vibratory sense is the usual first clinical sign of neuropathy, followed by loss of ankle jerks and later by loss of pinprick sensation in a stocking distribution.257 The Pittsburgh Epidemiology of Diabetes Complications study found that 3% of 65 children under 18 years of age had neuropathy based on history and physical examination.258 Subclinical neuropathy with decreased motor nerve conduction velocity and sensory change is much more common, however, in as many as 50% to 72% of adolescents.259

Neuropathy has been correlated with hyperlipidemia, smoking,259 limited joint mobility,260 and albuminuria.261 Among 1172 patients followed for an average of 7.3 years, neuropathy, determined by clinical evaluation, quantitative sensory testing, and autonomic function tests, developed in 23.5% of patients. Risk markers were HbA1c, duration of diabetes, and modifiable cardiovascular risk factors such as LDL cholesterol, triglycerides, body mass index, smoking, and blood pressure.256

The Pittsburgh Epidemiology of Diabetes Complications (EDC) study found the relative risk of developing peripheral neuropathy to be 2.2 if LDL cholesterol was greater than 129 mg/dL and 1.5 if LDL cholesterol was greater than 99 mg/dL but less than 130 mg/dL. Similarly, triglyceride levels greater than 140 mg/dL conferred a 1.5 relative risk. A systolic blood pressure greater than 130 mm Hg was associated with a relative risk of 4.0 and a diastolic blood pressure greater than 85 mm Hg with a 2.0 relative risk.236

Improved metabolic control may result in resolution of sensory and motor nerve dysfunction as well as autonomic symptoms. The Diabetes Care and Complications Trial showed that intensive diabetes therapy decreased clinical neuropathy by 35% to 90%.183 Peripheral motor and sensory nerve conduction velocities were significantly faster after 5 years of intensive therapy compared to conventional treatment in the adolescent cohort.261

Manifestations of autonomic neuropathy include gastroparesis, cardiovascular reflex loss, and hypoglycemic unawareness. Diabetic gastroparesis is associated with decreased gastric motility and delayed emptying time. Affected patients often complain of bloating and feelings of satiety following intake of small amounts of food, and they may have anorexia and weight loss. Diagnosis is by study of gastric emptying. Gastroparesis also results from hyperglycemia, so improved metabolic control is important for management.262,263 Prokinetic agents including metoclopramide, domperidone, tegaserod, and erythromycin may alleviate symptoms.

Autonomic neuropathy of the cardiovascular system affects heart rate, typically with fixed heart rate, and defects in central and peripheral vascular dynamics. There is hypersensitivity to catecholamines and an inability to increase cardiac output with hypotension on standing or with exercise. Affected patients will often have a lower blood pressure when sitting or standing than when supine. Cardiovascular autonomic neuropathy also increases the risk for painless myocardial ischemia and fatal cardiovascular events.264

Hypoglycemia unawareness is the loss of the catecholamine-induced symptoms of sweating, tachycardia, nausea, anxiety, and tremulousness that normally accompany severe drops in blood glucose concentrations.

Macrovascular Disease and Hyperlipidemia

Diabetes confers a two- to fourfold increased risk for cardiovascular disease.265-268 Factors that further independently increase cardiovascular risk in people with diabetes include smoking, hypertension, dyslipidemia, renal dysfunction, and hyperglycemia.269

The Pathological Determinants of Atherosclerosis in Youth evaluated risk factors for coronary heart disease in approximately 3000 young people who died between the ages of 15 and 34. Pathologic findings of fatty streaks and lipid-laden plaques were correlated to blood lipid values, smoking, hypertension, and diabetes. There was a strong association of extent of disease with smoking, as measured by serum thiocyanate concentrations. All the aortas and about half of the right coronary arteries in the 15- to 19-year age group already had lesions, with 7% of the aortas and 12% of the right coronary arteries having raised lesions or advanced lesions of atherosclerosis. The percentage of intimal surface involved with lesions in both the aorta and right coronary artery was positively associated with VLDL and LDL cholesterol, with a 5% increase in surface involvement with each 1 standard deviation increase of LDL and VLDL cholesterol levels. Conversely, a 1 standard deviation increase in HDL cholesterol was associated with a 3% decrease in intimal surface involvement. HbA1c concentration was positively associated with more extensive and advanced atherosclerosis in the aorta and right coronary artery, primarily in people between the ages of 25 and 34 years of age. The prevalence of raised lesions involving 5% or more of the intimal surface was twice as great in both the aorta and right coronary artery of people with hypertension throughout the entire 15- to 34-year age span.270

The increased cardiovascular risk in type 1 diabetes (T1D) may be due to a number of factors, including endothelial dysfunction, increased arterial stiffness, and loss of endothelial integrity.267,271,272 The endothelium controls vascular tone through the production of nitric oxide. Hyperglycemia increases reactive oxygen species that directly reduces the bioavailability of nitric oxide, resulting in endothelial dysfunction.273 Endothelial dysfunction is a precursor to atherosclerosis, and its measurement can be used to predict cardiovascular events.

Endothelin-1, an indicator of endothelial damage, is increased in patients with T1D and hyperlipidemia and those with T1D who had vascular complications had significantly higher endothelin-1 levels than did patients without complications. The patients with complications also had significantly elevated apolipoprotein B levels. Patients without microvascular or macrovascular disease had levels similar to those of controls. Thus, it is possible that the susceptibility to the development of atherosclerosis might be attributed to the relationship between elevated lipid levels and endothelin-1.274

The Heart Outcome Prevention Evaluation (HOPE) trial involving 5545 individuals aged 55 years or more with a history of cardiovascular disease or with diabetes mellitus and at least 1 cardiovascular risk factor, observed that microalbuminuria increased the relative risk of major cardiovascular events (relative risk 1.83), all causes of death (relative risk 2.09), and hospitalization for congestive heart failure (relative risk 3.23). The use of the angiotensin-converting enzyme (ACE) inhibitor ramipril significantly reduced the risk by 25% for myocardial infarction, 37% for cardiovascular deaths, 33% for stroke, and 24% for all causes of mortality. ACE inhibition also reduced the risk of overt nephropathy by 22% and dialysis by 15%.275 Ramipril works not only by decreasing blood pressure, but also has antithrombotic effects, reducing collagen-induced platelet aggregation by 18% and ADP induced platelet aggregation by 39%.276 ACE inhibitors also have effects on endothelium and fibrinolysis that are beneficial in preventing plaque rupture and subsequent thrombosis, 2 key events in the acute formation and progression of atherosclerotic disease.276

As previously noted, the Pittsburgh EDC demonstrated a strong correlation between lipid levels, blood pressure, and risk of cardiovascular disease in their 10-year study of patients with T1D onset before age 18 years. Tight glycemic control may decrease cardiovascular event rates in patients with T1D and may provide the same “metabolic memory” effect demonstrated by the Diabetes Care and Complications Trial (DCCT) follow-up studies for microvascular complications. Mean HbA1c levels 1 year after conclusion of the DCCT were nearly identical between the former intensive and former conventional groups. While the Epidemiology of Diabetes Interventions and Complications cohort is still relatively young and the total event rates are still quite small, the intensive insulin therapy group demonstrated lower cardiovascular event rates than the conventional therapy group.277

Treatment of Hyperlipidemia

The American Diabetes Association (ADA) recommendation is that adults with diabetes maintain LDL cholesterol less than 100 mg/dL, triglycerides less than 150 mg/dL, and high-density lipoprotein (HDL) greater than 40 mg/dL.278 Even lower LDL levels may further decrease cardiovascular disease in high-risk patients.279 The ADA clinical practice recommendations now suggest an LDL goal of less than 70 mg/dL in high-risk patients with diabetes and cardiovascular disease.278 That fatty streaks and atherosclerotic plaques are present during childhood indicates that the process begins early, suggesting that it is reasonable to apply adult intervention goals in children who have T1D.

If LDL cholesterol is borderline (100–129 mg/dL; 2.6–3.4 mmol), or elevated (⩾ 130 mg/dL; 3.4 mmol), dietary intervention to decrease total and saturated fat in the diet should be initiated and a repeat lipid profile obtained in 6 months. If LDL cholesterol remains elevated after 6 months of attempting to improve blood glucose control and diet, pharmacotherapy is warranted. Statins are as safe and effective in children as in adults and should be the first pharmacologic intervention.183,280,281 The usual starting dose of atorvastatin is 10 mg, after normal liver function is verified. Fasting lipid levels should be rechecked 6 weeks after starting therapy and the dose should be titrated upward at 10-mg increments to achieve the target level of LDL cholesterol. Adolescent females should be warned of the teratogenic potential of statins. Cholesterol absorption inhibition can be used as adjunct therapy, adding ezetimibe 10 mg daily if the LDL cholesterol goal is not reached with 20 mg atorvastatin. Triglyceride elevation should also be vigorously treated. Isolated hypertriglyceridemia may be seen in some patients unable to achieve optimal blood glucose levels. In those instances, fenofibrate (48 mg daily), should be considered. Tryglycerides should be rechecked 6 weeks after initiating therapy and the dose titrated upward to achieve normal triglyceridemia.

Hypertension

The United Kingdom Prospective Diabetes Study showed that above a systolic blood pressure of 110 mm Hg, the higher the blood pressure, the greater the risk of cardiovascular disease.282 The Hypertension Optimal Therapy study compared the outcomes of maintaining diastolic blood pressure at goals of less than or equal to 90 mm Hg, less than or equal to 85 mm Hg, or less than or equal to 80 mm Hg. In the 15,001 patients with diabetes, the group randomized to maintain a diastolic blood pressure less than or equal to 80 mm Hg had half the risk of myocardial infarction, stroke, or death due to any cardiovascular event than did the group randomized to a blood pressure less than or equal to 90 mm Hg. This differs from the results seen in the group without diabetes, in which the risk reduction was only 10% for the lower compared to the higher blood pressure readings, indicating that diabetes conferred an additional risk for cardiovascular disease.283

Treatment of Hypertension

Children with high-normal blood pressure, defined as average systolic or diastolic blood pressure greater than the 90th percentile for age, sex, and height, should reduce salt intake and increase exercise (30–60 minutes of daily exercise that raises heart rate). If blood pressure has not improved after 3 to 6 months of lifestyle interventions, pharmacologic treatment is indicated. Hypertension, defined as average systolic or diastolic blood pressure consistently greater than the 95th percentile for age, sex, and height on 3 separate occasions, should be treated with angiotensin-converting enzyme inhibitors.278,280

Classification of Non-Type 1 (Autoimmune Mediated) Diabetes

Diabetes classification was reviewed at the beginning of this chapter. Non-type 1 (autoimmune-mediated) forms of diabetes include atypical diabetes mellitus of African Americans (aka “Flatbush” diabetes), monogenic forms (MODY), mitochondrial mutations, and genetic defects in insulin action. A vexation in the classification has been the finding of diabetes-specific autoantibody positivity in 15% to 40% of individuals with what appears to be typical type 2 diabetes (T2D), including many who were not receiving insulin 1 year after diagnosis.32-34 This may represent autoimmune type 1 diabetes (T1D) in overweight or obese individuals with underlying insulin resistance or an inflammatory response to antigen exposure caused by apoptosis of β cells.34 Antibody-positive young adults with a T2D phenotype are less overweight and younger than antibody-negative patients, and HbA1c concentrations are significantly higher in antibody-positive compared to antibody-negative young adults with T2D. Furthermore, β-cell function is significantly less in adults who are antibody positive, with the most dramatic difference being in the younger age group (25–34 years), with more rapid development of insulin dependence, usually by 3 years after diagnosis. The presence of islet cell antibodies and glutamic acid dehydrogenase antibodies in adults with phenotypic T2D has been referred to as latent autoimmune diabetes of adults, which is inappropriate, as neither the autoimmunity nor the diabetes is latent.34

T2D in children and adolescents is becoming an increasingly important public health concern throughout the world.8,28,284-301 In North America, T2D comprises 30% of all newly diagnosed diabetes in patients 10 to 20 years of age.302 Not surprisingly, this epidemic is closely associated with the increased prevalence of obesity among youth of all ethnic backgrounds, as increased visceral adipose tissue produces adipokines that increase insulin resistance.303 As in adult T2D, children with T2D typically have the insulin-resistance/metabolic syndrome, which includes abdominal obesity, disturbed glucose regulation, dyslipidemia, and hypertension.304

Because of the relatively recent recognition of the problem in this age group, newly diagnosed patients with T2D may be misclassified as having T1D, especially if they are ketotic. Risk factors for cardiovascular disease are already present at the time of diagnosis of T2D, making immediate and continuing normalization of blood glucose levels, blood pressure, and lipids important.304

Epidemiology

Beginning around 1990, pediatric endocrinologists noticed a sharp increase in the numbers of patients with T2D.28,286,305-307 This was predominantly, but not exclusively, seen in the African American and Mexican Hispanic-American population, suggesting ethnic differences in background insulin sensitivity. Indeed, in Asian youth and adults, T2D occurs without overall obesity, but increase in intraabdominal fat accumulation.308 In studies from Cincinnati, Arkansas, and Texas, African Americans accounted for 70% to 75% of pediatric diabetes.286,306,309 It is estimated that one third of Mexican American children and youth with diabetes in Southern California and over two thirds of those in South Texas have T2D mellitus.310,311

The first recognition of a substantial prevalence of T2D in children and adolescents was in 1977 among 15- to 24-year-old Pima Indians, with 9 out of 1000 (0.9%) having diabetes associated with obesity and early onset of microvascular and cardiovascular complications.291 This was in a population in which 50% of adults have T2D. Half of these youngsters had ketoacidosis. By the 1990s, 5% of 15 to 19 year olds and 2.2% of 10 to 14 year olds (previously 0%) had T2D.312 Among First Nations people in Canada, the frequency of T2Dmellitus in children and youth is comparable to that of T1Dmellitus in those of European ancestry.313

Reports from throughout the United States indicate that the percentage of patients with new-onset diabetes who have T2D increased from 2% to 4% in 1994 to 20% to 50% by 2000.301The percentage of patients diagnosed with T2D had previously been stable. Knowles reported approximately 3.5% of patients in a Cincinnati-based diabetes clinic having T2D in 1971305; a similar proportion (2–4%) was noted from the same clinic between 1982 and 1992, with a sharp increase to 16% by 1994.314 In Native North Americans, there have been 4 to 6 times as many females as males affected, but in the African American and Mexican American groups with T2D mellitus, the sex ratio has been far less skewed, varying from 1.7 females for every male in African Americans to nearly 1:1 in Mexican Americans.315

A study of 682 children age 5 to 19 years old diagnosed between 1994 and 1998 at the 3 university-based diabetes centers in Florida found that 14% of the patients had T2D. Whereas 47% of T1D mellitus patients were female, 63% of T2D patients were female.28 In contrast to the studies from Arkansas and Cincinnati, African Americans comprised only 46% of those with T2D, while 22% were Hispanic and the rest non-Hispanic whites. The risk for developing T2D was 3 times greater for African American youngsters and 3.5 times greater for Hispanics than for whites. During the initial year of the study, 8.7% of newly diagnosed patients were eventually classified as having T2D, whereas in the last year of the study 19% were thus classified, indicating a twofold increase in the proportion of new diabetes patients in this age group having T2D over the relatively brief period of 5 years.28

The recognition of an epidemic of T2D is not unique to North America (eTable 544.2). In Japan 80% of all new cases of diabetes in children and adolescents in 2000 were T2D. Annual urine testing for glucose of schoolchildren in the Tokyo area has been carried out since 1975, followed by oral glucose tolerance testing for those who have glucosuria. In the primary school age group, the incidence of T2D increased from 0.2 per 100,000 in 1976 to 2 per 100,000 in 1995, while in the 12- to 15-year age group, there has been a doubling of incidence of T2D from 7.2 to 13.9/100,000, paralleling increasing obesity rates.294 In contrast, the incidence of T1D was 1.5 cases per 100,000 in 1976 and did not change appreciably over the next 20 years. Other reports indicate that increasing rates of T2D with concurrent increases in obesity are also being observed in children in Thailand, China, India, New Zealand, Australia, and throughout Europe.301 The prevalence of T2D in England is estimated at 0.2%, and this problem is expected to be more prevalent with the epidemic of childhood obesity in that country.316

A constant in the emergence of type 2 diabetes (T2D) in young patients has been the association with increasing rates of obesity. The US National Health and Nutrition Examination Survey conducted between 1988 and 1994 found that 20% of children 12 to 17 years of age had body mass indexes above the 85th percentile for age, the definition of overweight, and that depending on ethnicity, 8% to 17% were obese, with a body mass index greater than the 95th percentile. Not only was there a doubling of the frequency of childhood obesity since 1980, but the severity of obesity was greater (see Chapter 32).317

Etiology

The recent increase in T2D prevalence in young patients has been so rapid that it can only be explained by changes in the environment. The most important of these is the increasing prevalence of obesity.318-324 Nonetheless, not all or even a majority of obese youngsters develop T2D, emphasizing the importance of underlying genetic predisposition. The thrifty genotype hypothesis, advanced nearly 40 years ago and recently updated,325,326 explains the insulin resistance and relative β-cell insufficiency that is associated with the development of obesity and T2D as an adaptation for conserving energy and surviving famine. Until the modern era of ready availability of high-calorie foods, such a genotype would have had great survival advantage.

Numerous studies have demonstrated that both insulin resistance and limited pancreatic β-cell capability to respond to the increased insulin requirements associated with obesity are the basic abnormalities in the development of T2D.327-333

Insulin resistance affects glucose, lipid, and protein metabolism. Normal glycemic control requires the sensing of glucose concentration by the cells, the synthesis and release of insulin, the coupling of insulin to its receptors, and postreceptor activation of specific signal transduction pathways. This results in increased glucose uptake by muscles, fat, and liver tissue, with decreased glucose production by the liver. In T2D, peripheral insulin resistance in muscle and fat tissue, taken together with decreased pancreatic insulin secretion and increased hepatic glucose output, result in hyperglycemia (Fig. 544-3).334

Low Birth Weight

The association of lower birth weight, smaller head circumference, and thinness at birth, with later development of insulin resistance and impaired glucose tolerance or T2D suggests in utero programming that limits β-cell capacity and induces insulin resistance in peripheral tissues.335,336 The maternal undernutrition that led to low birth weight is thought to impair development of the endocrine pancreas. The effect of fetal undernutrition on adult glucose tolerance has been confirmed in large studies in Sweden and the United States.337,338 Glucose intolerance was also found in adults who were the offspring of mothers who had starved during the last trimester of pregnancy during the Dutch famine near the end of World War II.339

Three studies of young subjects from high-risk populations substantiate the reports of the effect of fetal nutrition on the risk for development of the insulin-resistance syndrome (T2D, hypertension, dyslipidemia) in adulthood. Current weight correlated with birth weight among 3061 Pima Indians aged 5 to 29 years, and 2-hour glucose concentrations had a U-shaped relationship with birth weight in those greater than 10 years old, regardless of current weight. Thus, higher blood glucose levels were present in those who had both high and low birth weights, irrespective of their current weight. The 2272 subjects without diabetes had negative correlations between birth weight and insulin resistance, determined by measurement of insulin concentrations at baseline and 2 hours following a glucose load. These findings support the hypothesis of a survival advantage for insulin resistance in low–birth weight babies.341

In a study of 477 8-year-old Indian children, insulin resistance variables and plasma total and LDL cholesterol concentrations were strongly related to low birth weight and high-fat mass at 8 years of age. Lower birth weight was associated with calculated insulin resistance, elevated systolic BP, fasting plasma insulin and proinsulin concentrations, plasma lipids, and glucose and insulin concentrations 30 minutes after glucose.342 Low birth weight children of European ancestry and African American children (n = 139) studied when they were 4 to 14 years old had significant differences in the effect of low birth weight on visceral fat mass as measured by dual energy x-ray absorptiometry and computed tomography, fasting insulin, acute insulin response to intravenous glucose, -cell function, and HDL cholesterol concentrations, indicative of the genetic differences suggested by the thrifty genotype hypothesis.343 These findings are consistent with the higher prevalence of T2D in African Americans.

Early Childhood Nutrition

Infant and child feeding practices appear to play a role in the development of insulin resistance later in life. An association between high protein intake in infancy and later obesity has been suggested.344 Breast-feeding results in a more appropriate caloric intake at a critical stage in development than does bottle feeding, which is more likely to be associated with overfeeding and obesity. Glucose tolerance testing was carried out in 720 Pima Indians aged 10 to 39 years, including 325 who had been exclusively bottle-fed as infants, 144 who were exclusively breast-fed, and the rest partially breast-fed for the first 2 months of life. Those exclusively breast-fed had significantly lower rates of type 2 diabetes (T2D) mellitus than those exclusively bottle-fed for each age decade, with an odds ratio for T2D in exclusively breast-fed individuals of 0.41.345 Furthermore, prolonged breast-feeding markedly reduced the risk of overweight in nearly 10,000 5- to 6-year-old German children; 3.8% of those exclusively breast-fed for 2 months were overweight versus 0.8% of those breast-fed for longer than 12 months.346 This finding may be due to the lower insulin responses and the lower energy and protein intake in breast-fed infants versus bottle-fed infants. The frequent overweight of the bottle-fed infant may contribute to insulin resistance and obesity in adolescence and young adulthood. The SEARCH diabetes case-control study in the United States analyzed breast-feeding histories of 80 youth with T2D aged 10 to 21 years matched with 167 nondiabetic control participants. Using various methods of analysis, breast-feeding conferred an odds ratio of 0.26 to 0.43 for the development of T2D, but when current body mass index z scores were added to the model, the odds ratio was attenuated to 0.82. The conclusion was that breast-feeding is protective for the later development of T2D, but the effect is mediated by weight status in childhood.347

The thrifty phenotype hypothesis states that poor nutrition in fetal and early infant life would be detrimental to the development and function of the cells and insulin-sensitive tissues, primarily muscle, leading to insulin resistance. With obesity in later life, T2D would develop. This hypothesis may explain the effect of fetal nutrition on later glucose tolerance and other manifestations of insulin insensitivity, such as hypertension and increased cardiovascular problems.335-337 These findings could also be interpreted as a reflection of the thrifty genotype hypothesis, that defective insulin action in utero results in decreased fetal growth and obesity-induced impaired glucose tolerance in later childhood or adulthood.335,336 Genetic factors affecting birth weight and glucose metabolism are of interest in this regard. The polymorphism at the VNTR (variable number of tandem repeats) locus of the insulin chain is associated with decreased body length, weight, and head circumference at birth.348 Decreased birth weight has been associated with heterozygosity for a mutation in the glucokinase gene.349 Finally, the glucose transporter 4 expression is impaired in young adults with insulin resistance who had undernutrition in utero.350

The Role of Maternal Diabetes

The influence of the diabetic intrauterine environment on the risk of T2D mellitus in children was first appreciated from studies in the Pima Indian population; the prevalence of diabetes in the offspring of Pima women with diabetes during pregnancy is significantly greater than in nondiabetic mothers or those who develop diabetes after delivery.351 In another study of the effect of maternal diabetes, fetal β-cell function was assessed in 88 pregnancies with pregestational or gestational diabetes by measuring amniotic fluid insulin (AFI) concentration at 32 to 38 weeks gestation and performing oral glucose tolerance tests annually in the offspring from 18 months of age. Only 1 of 27 adolescents with normal AFI had impaired glucose tolerance, in contrast to one third of those with elevated AFI.352 These studies suggest a generation-to-generation accumulation of risk for T2D that further increases the public health concern of the epidemic of this disease in young persons.352-354

The Role of Puberty

In all reports of T2D in childhood, the mean age at diagnosis is approximately 13.5 years, corresponding to the time of peak adolescent growth and development. Puberty is associated with relative insulin resistance, reflected in a twofold to threefold increase in the peak insulin response to oral or intravenous glucose355; insulin-mediated glucose disposal is approximately 30% lower in adolescents than in prepubertal children or young adults.356 The physiologic insulin resistance of puberty is of no consequence in the presence of adequate β-cell function. The cause of this physiologic resistance is likely the transitory increased activity of the GH-IGF axis, which coincides with the physiologic insulin resistance of adolescence.8

The Role of Obesity

The insulin resistance associated with obesity is the fundamental problem in T2D in children and adolescents, as it is in adults. Total obesity is not as important as location of adipose tissue in causing insulin resistance.357 Visceral fat is more metabolically active than subcutaneous fat and produces adipokines that cause insulin resistance, thus predisposing to T2D. This would explain the relative frequency of nonobese T2D in Asia.358 That obese children have hyperinsulinism has been known for over 30 years.359,360 Obese children have approximately 40% lower insulin-stimulated glucose metabolism compared to the nonobese.357 African American 5- to 10-year-old children, especially girls, have reduced insulin sensitivity, and this correlates with increases in blood pressure, triglycerides, subcutaneous fat, percentage of total body fat, and stage of sexual maturation.361 The amount of visceral fat in obese adolescents correlates directly with basal and glucose-stimulated insulin levels and inversely with insulin sensitivity.357 Insulin-stimulated glucose metabolism decreases while fasting insulinemia increases with increasing body mass index.356

The Role of Race, Gender, and Family History

There is a racial difference in the insulin responses to various stimuli that parallels the ethnic/racial differential in T2D frequency. Greater insulin responses to oral glucose are seen in African American children and adolescents than in European American children after adjustments are made for weight, age, ponderal (obesity) index, and pubertal stage. This is indicative of compensated insulin resistance in the African American youngsters.362-364 Both prepubertal and pubertal African American children have higher fasting and stimulated insulin concentrations during glucose clamp studies than do European American youngsters.364 Lipolysis is also significantly less in African American children than in European American children, suggesting an energy conservation phenotype that would have survival value in times of famine, but be detrimental with excess nutrition (thrifty genotype).365,366

The SEARCH for Diabetes in Youth population-based study found the proportion of physician-diagnosed diabetes being T2D among 10 to 19 year olds to vary greatly by ethnicity in the United States: 6% for non-Hispanic whites, 22% for Hispanics, 33% for blacks, 40% for Asians/Pacific Islanders, and 76% for Native Americans.367 In Asian Indians, T2D accounts for 27% of diabetes diagnosed at or under 16 years of age.358

A noninterventional cohort study showed that girls are more insulin resistant than boys as early as at 5 years of age.367 The study also found that girls have 26% more subcutaneous fat than boys, which may contribute to their development of insulin resistance. Thus, sex-linked genes may contribute to the development of T2D and the insulin resistance/metabolic syndrome in the current food-rich environment.

Prepubertal children who have a family history of T2D have lower insulin-stimulated glucose disposal and nonoxidative glucose disposal than do those without such a family history, indicating that family history of T2D is a risk factor for insulin resistance in children, as it is in adults.368

Genetic Considerations

The evidence that T2D is a genetic disease includes the family clustering and segregation analyses that indicate that siblings of affected individuals have 3.5 times the general population risk of developing T2D; studies of monozygotic twins that indicate a concordance of 80% to 100%, greater than twice the concordance for T1D; and the previously noted variations in insulin sensitivity and frequency of T2D by ethnicity. With rare exception, T2D in children and adolescents, as in adults, is polygenic.

eFigure 544.3 illustrates possible candidate genes in the β cell-target cell interaction. Over 20 loci have been linked to or associated with T2D in adults, the most important being NIDDM type 1 (NIDDM1), described among Mexican American sibships in Starr County, Texas. This county, which is 97% Mexican American, has the highest disease-specific diabetes mortality in Texas. The gene pool is 31% Native American. Examined were 474 autosomal markers and 16 X-linked markers in 170 affected sibships, involving 300 affected siblings and 78 unaffected siblings. Identified as linked to T2D was the NIDDM1 site on chromosome 2, accounting for approximately 30% of family clustering—equal in importance, therefore, to the linkage with HLA in T1D mellitus.369 This linkage was not identified in other populations, including non-Hispanic people of European ancestry, Japanese, French, Sardinian, and Finnish.370 The Calpain gene in the NIDDM1 region of chromosome 2 was subsequently found to be associated with T2D in several of these populations, linked to different loci.371

Calpains are calcium-activated neutral proteases that are ubiquitously expressed from fetal life through adulthood, functioning in signaling, proliferation, differentiation, and insulin-induced downregulation of IRS-1. Multiple polymorphisms of the gene encoding Calpain-10, encoded by the CAPN10 gene within the NIDDM1 region have been found to be associated with T2D. The highest risk combination of polymorphisms gives odds ratios of 2.8 to 3.6 in Mexican Americans, 2.6 in Finns, and 5.0 in Germans.371 Nondiabetic Pima Indians who are homozygous for a common polymorphism of CAPN10 have reduced insulin-mediated glucose turnover as the result of decreased glucose oxidation rates.372

Clinical Features

The most striking difference between type 1 diabetes (T1D) and type 2 diabetes (T2D) is that in T2D, hyperglycemia/diabetes is 1 of many manifestations of the insulin-resistance syndrome (also known as the diabesity syndrome, syndrome X, or the metabolic syndrome), whereas T1D is a disorder of insulin deficiency and is not usually associated with the comorbidities seen with insulin resistance early in the course of the disease (Fig. 544-4).

Pediatric endocrinologists in the United States report that 100% of the patients identified as having T2D will have a body mass index greater than 85% and 90% to 95% will have acanthosis nigricans, a cutaneous manifestion of insulin resistance characterized by velvety, hyperpigmented skin, most often in intertriginous areas.373 In contrast to T1D, at presentation of T2D, most children will have glycosuria, with mild or absent polyuria and polydipsia, and little to no recent weight loss. Although diabetic ketoacidosis (DKA) is not as common at presentation of T2D as with T1D severe insulinopenia due to β-cell dysfunction from glucose toxicity can result in DKA in patients with T2D. The hyperglycemic hyperosmolar state is a serious acute presentation that is almost exclusive to T2D.86,374,375 While a family history of T2D may be helpful in distinguishing T1D from T2D, as 74% to 100% of children with T2D have a first- or second-degree relative with T2D, T2D is sufficiently frequent in the population, especially among ethnic minorities, that such a history will be frequently obtained among those with T1D.8 Occasionally, diagnosis will only be clarified by observing the clinical course of the patient and may be helped by testing for the presence of diabetes-specific autoantibodies to islet cells, glutamic acid decarboxylase, insulin, and tyrosine dehydrogenase. Diabetes-specific autoimmunity has been described with typical T2D, however, as noted above. Measurement of C-peptide levels to distinguish T1D from T2D may not be helpful because C-peptide levels may be low in children with T2D at diagnosis due to the glucose toxicity to the β cells, and in T1D they may be adequate during the early recovery (honeymoon) phase.

Manifestations of the insulin-resistance syndrome (IRS) include hypertension, endothelial dysfunction, atherosclerosis, dyslipidemia, a prothrombotic environment, impaired glucose tolerance, acanthosis nigricans, hyperuricemia, polycystic ovary disease, and obesity, the core abnormality (Fig. 544-4).37 IRS increases the risk for development of cardiovascular disease. Individuals who have the components of IRS in youth track them into adulthood. The prevalence of IRS is 4% of all adolescents and 30% of overweight adolescents in the United States.376 The fact that IRS is being identified in younger individuals indicates that the development of cardiovascular disease will also occur earlier in adulthood.

Dyslipidemia

Lipoprotein abnormalities noted in T2D are associated with increased cardiovascular disease and include hypertriglyceridemia, elevated very low-density lipoprotein (VLDL), elevated LDL cholesterol, elevated lipoprotein (a), decreased HDL cholesterol, increased small dense LDL particles, decreased lipoprotein lipase activity, increased lipoprotein glycation, and increased lipoprotein oxidation.384

This dyslipidemia reflects the reduction of insulin effect in adipose tissue, which is to store triglyceride and suppress hormone-sensitive lipase, the enzyme that breaks down triglycerides to release free fatty acids. Insulin also provides glucose to the fat cells for forming glycerol, the triglyceride backbone. With insulin resistance, there is abnormal breakdown of triglyceride and release of free fatty acids and glycerol, the latter contributing to gluconeogenesis. The free fatty acids cause insulin resistance in muscle tissue and in the liver are reconverted to triglyceride, driving the production of LDL (the lipoprotein carrier of triglycerides), and the other dyslipidemic changes follow. The reduction of insulin effect on lipoprotein lipase contributes to the elevated LDL levels. The hyperinsulinemia of the insulin-resistance state further drives the synthesis of fatty acids from glucose in the liver.385

Hypertension

Hypertension is estimated to account for 35% to 75% of diabetes complications involving both the microvasculature and macrovasculature.386 Diabetes or impaired glucose tolerance doubles the risk of developing hypertension.387 There is evidence for a genetic predisposition to hypertension and T2D that is related to the angiotensin-converting enzyme genotype.388 Hypertension in T2D is multifactorial, due to volume expansion and increased vascular resistance, with reduced nitric oxide–mediated vasodilatation and increased activity of the renin-angiotensin system.384

Endothelial Dysfunction

Endothelial dysfunction, 1 of the earliest signs of increased risk for cardiovascular disease, develops in the milieu of cardiovascular risk factors such as obesity, hypertension, dyslipidemia, insulin resistance, and T2D and is predictive of cardiovascular events.389 Endothelial dysfunction is present in obese children and is related to the amount of obesity as well as to the degree of insulin resistance.390

Several cytokines produced by adipose tissue (adipokines) have been shown to be associated with endothelial dysfunction, including tumor necrosis factor α (TNF-α), interleukin-6, plasminogen activator inhibitor-1, leptin, and resistin.392 These adipokines are proinflammatory, and in vitro they have activating effects on the endothelium and contribute to the induction of thrombosis. The adipokine adiponectin, which is underexpressed in individuals with obesity or T2D mellitus, decreases endothelial inflammation, inhibits vascular smooth muscle proliferation, and suppresses macrophage-to-foam cell transformation.393 Levels of adiponectin correlate inversely with other markers of inflammation, and low levels have been found to be associated with increased cardiovascular risk.

Acanthosis Nigricans

Acanthosis nigricans (AN), an indicator of insulin resistance, is prominent in genetic insulin-resistance syndromes that are not associated with obesity. The frequency of AN in obese adolescents varies greatly by ethnicity; approximately 90% in Native Americans, approximately 50% in African Americans, approximately 15% in Hispanic Americans, and less than 5% in those of non-Hispanic European ancestry. As an indicator of hyperinsulinism, AN also varies by ethnicity inversely to its frequency of association with obesity.394 In a study of 139 overweight 6 and 10 year olds, AN was present in 50% of African Americans and 8.2% of whites. Half of those with fasting hyperinsulinemia did not have AN and AN was not considered a reliable marker for hyperinsulinemia in overweight children, despite its presence in a child with diabetes being an indicator of T2D.395

Prevention

The concept of prevention of T2D mellitus arises naturally from the appreciation of its environmental causation. Primary prevention is intervention to prevent the development of obesity or to correct obesity before the development of other features of insulin-resistance syndrome. Prevention of obesity is a societal challenge. The importance of primary prevention of obesity lies in the recognition that preclinical impairment of glucose tolerance as the result of obesity-induced insulin resistance conveys cardiovascular risk and that a substantial number of severely obese adolescents have impaired glucose tolerance.304

From the individual physician’s standpoint, the major question is Who should be tested?8 Considerations of testing for T2D mellitus in children began with the assumption that this will be done in obese individuals, making the determination of obesity the screening test. A consensus panel of the American Diabetes Association8 recommended screening testing based upon the following criteria.

Overweight is defined as body mass index greater than 85th percentile for age and sex, weight for height greater than 85th percentile, or weight greater than 120% of ideal for height, plus any 2 of the following risk factors:

• Family history of type 2 diabetes in first- or second-degree relative
• Race/ethnicity (American Indian, African American, Hispanic, Asian/Pacific Islander)
• Signs of insulin resistance or conditions associated with insulin resistance (acanthosis nigricans, hypertension, dyslipidemia, polycystic ovarian syndrome [PCOS])

If these criteria are fulfilled, fasting plasma glucose testing is recommended every 2 years starting at age 10, or at the onsert of puberty, if puberty occurs at a younger age.

These criteria were not databased, which is why the consensus panel provided a disclaimer that they should not replace individual clinical judgment. One might also take issue with the preference for the lower-cost and more convenient fasting plasma glucose measurement, because 2-hour postprandial plasma glucose increases earlier in the course of the development of type 2 diabetes (T2D), making it a more sensitive measure. Fasting hyperglycemia occurs later, and therefore is highly specific, but is relatively insensitive as a diagnostic test. Higher fasting blood glucose levels in the normoglycemic range (ie, >87 mg/dL [4.83 mmol/L]), however, have been shown to be a risk factor for T2D in 26- to 45-year-old men.397 Random plasma glucose concentration can be measured in those with food intake shortly before testing, with a glucose concentration greater than or equal to 7.8 mmol/L (140 mg/dL) serving as an indication for further testing.

Recognition that obesity is associated with the rapid increase in incidence of T2D in children has provided justification for primary prevention efforts for a number of years, but the trend toward increasing obesity persists.398-404 The sine qua non for prevention of T2D is lifestyle modification: decreased caloric intake and increased exercise. Diabetes prevention programs have to be designed with an understanding of the health beliefs and behaviors of the community. For example, 1 survey of Native American youth with family members having diabetes did not relate the complications of retinopathy or amputation to the diabetes. Over half of the youth thought that diabetes was caused by bad blood, and greater than one third attributed it to general weakness.405-424

Fewer than 5% of people who attempt diet and exercise modifications to lose weight actually lose a substantial amount of weight and maintain that weight loss for a long period of time, while greater than 90% regain their weight within 1 year.425 Exercise alone, with no or minimal weight loss, can reverse T2D. This has been documented in a 1-week summer camp program for Native American youth with T2D, who were able to achieve normoglycemia after 5 days of increased physical activity. Unfortunately, the behaviors of camp were not continued, with most of the youngsters reverting to poor glycemic control at home.410

Due to the difficulties of making enduring lifestyle changes, many patients seek and physicians attempt pharmacologic intervention to achieve weight loss and prevent the onset of T2D. Orlistat is currently approved for the reduction of body weight in children, but is associated with flatulence and fecal soilage, making it an unpopular choice.426 Metformin has been studied in 29 black and white adolescents aged 12 to 19 years with body mass index (BMI) greater than 30 kg/m2, a family history of T2D, elevated fasting insulin concentrations (> 15 U/mL), and normal fasting glucose and HbA1c levels.427 Subjects were randomized to receive metformin 500 mg twice daily or placebo for 6 months. Controls had an increase in BMI of 0.23 SD and metformin-treated subjects a decrease of 0.12 SD, a significant difference. There was also a significant decrease in fasting glucose concentrations and insulin levels in metformin-treated youngsters versus controls. Transient abdominal discomfort or diarrhea occurred in 40% of those taking metformin. This study, too, suffers from small numbers, short follow-up, and a select study population. In adults, lifestyle intervention was more effective than metformin in preventing progression from impaired glucose tolerance to T2D.421 Surgery is the last option for patients who have significant obesity-related morbidities that have failed lifestyle modification and medication (see Chapter 32).434,435 A randomized controlled trial of gastric banding versus conventional treatment for recent onset T2D in a small population of obese individuals from Australia achieved a 73% remission rate that correlated with weight loss and lower baseline HbA1c, without serious complications.436

Treatment of Type 2 Diabetes

The treatment challenges of type 2 diabetes (T2D) in children and adolescents differ greatly from those of type 1 diabetes (T1D), due largely to the nature of the disease and those most likely affected. While T1D is distributed throughout the population, T2D disproportionately affects families with fewer resources, paralleling the distribution of obesity in the population. Whereas T1D occurs throughout childhood, predominantly during the time when parental influence predominates, T2D affects mostly those in adolescence or beyond, when peer influence is most important. There is also a large difference in family experience, with only approximately 5% of families with a child with T1D having affected family members, while 90% or more of youth with T2D have family experience. These family members have typically failed to control their weight and glycemia, developing complications, and creating an aura of despair and futility. Treatment priorities are also different between T1D and T2D. Extensive lifestyle modification, beyond insulin administration and glucose monitoring, is only required by those patients with T1D who are overweight and inactive; however, the emphasis is on lifestyle modification in all those with T2D and secondarily on glucose monitoring and hypoglycemic medication. Finally, technologic innovation has revolutionized management of T1D with improved insulin purity and delivery, self- self-monitoring of blood glucose, and the development of insulin analogs, with an artificial pancreas on the horizon along with the possibility of islet cell replacement. Technological developments have, however, been the underlying cause of the problem of T2D, with advances in home entertainment systems, labor-saving devices, transportation, and food preparation making calorically dense food increasingly available, desirable, and inexpensive.

The goals of treatment for T2D are to promote weight loss, normalize glycemia and HbA1c, control or prevent hypertension and hyperlipidemia, increase exercise capability, and reduce acanthosis nigricans. Treatment is more important than might be indicated by the level of glycemia in some patients because of the multitude of cardiovascular risk factors associated with insulin resistance. Just as in adults with newly diagnosed T2D, young patients may already have evidence of complications, reflecting a prolonged period of impaired glucose tolerance. Among 100 Pima Indian children and adolescents with T2D mellitus, 7% had high cholesterol (⩾ 200 mg/dL), 18% had hypertension (BP ⩾ 140/90, now considered far too high a cutoff), and 22% had microalbuminuria (albumin/creatinine ⩾ 30) at the time of diagnosis. Ten years later, while still in their 20s, they had mean HbA1c of 12%, indicative of poor control, 60% had microalbuminuria, and 17% macroalbuminuria (albumin/creatinine ⩾ 300).437 Japanese investigators have described a high risk of renal failure in those who developed T2D mellitus under 30 years of age.438 It is likely that reduction in the risk of complications may require more stringent glycemic control in the insulin-resistance state of T2D than is required in T1D, with diligent attention to comorbidities. In the United Kingdom Prospective Diabetes Study there was a 25% decrease in the risk of microvascular complications when the average HbA1c decreased from 7.9% to 7.0%.439 Reduction in BP to below 144/82 resulted in a more dramatic decrease of 37% in the risk for microvascular disease, 44% decrease in stroke occurrence, and 36% decrease in heart failure.440

Changing Behavior

The motivation required to promote lifestyle changes usually requires a treatment team, including a psychologist or social worker. One of the simplest changes to make is to eliminate the frequent consumption of sugar-containing drinks, including high-caloric soft drinks, sweetened tea, and juices, and substituting water, diet soft drinks, and artificial sweeteners for tea or Kool-Aid®.441 Consumption of sugar-sweetened beverages is strongly associated with an increased risk of childhood obesity and T2D. Daily exercise should be documented in an attempt to break the vicious cycle of increased weight producing increased torpor, resulting in decreased activity and increased weight. The most effective single method for doing this is turning off the TV and other electronic screen interactions. A relatively small reduction in weight, accomplished by an increase in activity, can restore euglycemia and decrease hyperinsulinemia, as discussed above. The effectiveness of diet as a means of controlling T2D in children was demonstrated in a short-term study using a very low-calorie ketogenic diet during which insulin or metformin treatment could be stopped and normal blood glucose concentrations were maintained.442

Hypoglycemic Agents

Pharmacologic therapy either decreases insulin resistance, increases insulin secretion, slows postprandial glucose absorption, or, in the case of insulin injection, supplements inadequate secretion of insulin (Table 544-8).

Biguanides

The biguanides act on insulin receptors in the liver, reducing hepatic glucose production, and in muscle and fat tissue, enhancing insulin-stimulated glucose uptake. They have an anorexic effect that can promote weight loss, and as noted above, metformin has been used for this purpose. It has also been used to reduce acanthosis and ovarian hyperandrogenism. Long-term use is associated with a 1% to 2% reduction in HbA1c, but a high rate of side effects, including transient abdominal pain, diarrhea, and nausea, limits compliance in adolescents.443

Metformin must not be given to patients with renal impairment, to those who have hepatic disease or cardiac or respiratory insufficiency, or to those who are undergoing radiographic contrast studies, because of the risk of lactic acidosis. Metformin is the only oral hypoglycemic agent currently approved for use in children. US Food and Drug Administration (FDA) approval was based on a study of 80 previously untreated patients age 8 to 16 years who were randomized to receive metformin or placebo in a multicenter study. Dosage began with 500 mg twice daily and increased to 2000 mg per day over 2 weeks. Rescue criteria resulted in few placebo cases remaining by 16 weeks of the study. At 4 months or longer, the mean fasting glucose change from baseline was a decrease of 44 mg/dL with metformin and an increase of 20 mg/dL with placebo. The adjusted mean HbA1c with metformin was 7.5% and with placebo 8.6%. With metformin, there was no weight gain, and modest decrease in some patients, and lipid profiles improved. No serious adverse events were recorded.444 Review of 29 trials done using metformin compared to placebo, insulin, or other hypoglycemic agents in adults with T2D mellitus showed that metformin not only normalizes blood sugars but also decreases cholesterol levels and lessens hypertension. Metformin may cause gastrointestinal upset, thus decreasing compliance, but the extended-release formulation has fewer side effects and tends to be better tolerated by patients.445 The extended-release formulation of metformin is given as a once-a-day dose, usually in the evening. Dosing is begun at 500 mg per day and can be increased to a maximum of 2000 mg per day. When metformin is not effective, it is important to do an in-depth history of medication intake, including refill history from the pharmacy, which may demonstrate that the medication is not being taken. It is also must be recognized that metformin may normalize ovulatory abnormalities in girls with polycystic ovarian syndrome (PCOS) or ovarian hyperandrogenism, increasing pregnancy risk.

Sulfonylurea and Meglitinide/Repaglinide

This group of drugs increases insulin secretion and is thus most useful when there is residual β-cell function. The sulfonylureas bind to specific receptors on the K+/ATP channel complex, while meglitinide and repaglinide bind to a separate site on the complex. Activating ATP or binding by these drugs causes K+ channels to close, with resultant membrane depolarization, allowing calcium influx and insulin release. While ATP-binding sites equilibrate rapidly, sulfonylurea sites do so slowly, with prolongation of binding explaining the sustained effects of traditional sulfonylureas. Meglitinide has an intermediate equilibration and binding duration, explaining its use for more rapid stimulation of insulin secretion and need for premeal dosing.446 The major adverse effects of the sulfonylureas have been hypoglycemia, which, as noted above, can be prolonged, and weight gain, which is particularly troublesome for adolescent patients. Glimepramide, a third-generation sulfonylurea, has been compared to metformin in a pediatric T2D trial, with comparable safety and efficacy.447

Glucosidase Inhibitors

α-Glucosidase inhibitors such as acarbose and miglitol reduce the absorption of carbohydrates in the upper small intestine by the inhibition of oligosaccharide breakdown, delaying absorption in the lower small intestine. This results in reduction in postprandial glycemia. Long-term use of glucosidase inhibitors is associated with a reduction in HbA1c of 0.5% to 1%.448 Flatulence associated with the use of these agents makes them unacceptable to young patients.

Thiazolidinediones

These drugs, also referred to as glitazones, act directly on muscle, adipose tissue, and liver to increase insulin sensitivity and are therefore considered specific agents for the manifold problems of the insulin-sensitivity syndrome. They bind to nuclear proteins, activating peroxisome proliferator activator receptors, orphan steroid receptors that are particularly abundant in adipocytes, increasing formation of proteins involved in the nuclear-based actions of insulin. These include cell growth, adipose cell differentiation, regulation of insulin receptor activity, and glucose transport into cells. Long-term treatment with thiazolidinediones has been associated with a reduction in HbA1c of 0.5% to 1.3%.449 Side effects have included edema, weight gain, and anemia. The original member of this group of drugs, troglitazone, was associated with liver enzyme elevations in approximately 1% of those taking the drug, with mortality in some who had existing liver problems, resulting in its withdrawal from the market. The newer glitazones, rosiglitazone and pioglitazone, have not been shown to have significant hepatotoxicity. The binding of thiazolidinediones to peroxisome proliferator activator receptor-γ is ubiquitous and includes arterial walls that contain muscle, affecting the growth of muscle cells and their migration in response to growth factors.450 These drugs also improve lipid profiles, decreasing LDL cholesterol and triglycerides and increasing HDL cholesterol. These effects on vascular muscle and lipids could be important for the reduction of macrovascular disease associated with T2D mellitus.451 Rosiglitazone use in pediatric T2D was compared to metformin in a 24-week double-blind study with 195 patients; reduction in HbA1c was comparable in both groups, and there were no safety problems with the rosiglitazone. However, there was weight gain, as in adults taking thiazolidinediones.452

Insulin

The development of T2D is an indicator of limited -cell function, estimated to be approximately 50% by the time of diagnosis in adults, most of whom will require insulin by 6 or 7 years later.439 Despite the insulin resistance, relatively small doses of supplemental insulin, a few units, may be sufficient to maintain euglycemia. Early normalization of blood sugars in patients with new-onset T2D, using intensive insulin therapy, results in lower HbA1c and improvement of insulin area under the curve on oral glucose tolerance testing 1 year after diagnosis. Long-acting insulin analog without peak effects, such as insulin glargine, may be especially useful for T2D, alone or in combination with pre-meal meglitinide, particularly in those individuals who are unwilling to take metformin.453 Hypoglycemia has not been as common a side effect in T2D as in T1D, but weight gain is an important adverse effect. All patients with T2D eventually require treatment with insulin because of ongoing loss of β-cell function.439 In adults with inadequately controlled T2D mellitus from multiple centers in Europe and Australia, a single daily injection of insulin glargine, a long-acting (24 hours) analog with little or no peak action, was almost as effective as a thrice-daily rapid insulin analog.454

Incretins

Incretins are gut-derived factors that potentiate insulin secretion following the oral ingestion of nutrients. In the mid-1960s it was found that the insulin secretory response to oral glucose was stronger than the response to intravenous glucose, and, therefore, it was recognized that that gut-derived factors contribute to postprandial insulin response (“the incretin effect”).455 Glucagon-like peptide-1 (GLP-1) is a product of the glucagon gene and is secreted by the L-cells of the small and large intestine within 5 to 10 minutes following meal ingestion. GLP-1 exerts its effects by binding and activating receptors that are found on the cells as well as many other tissues. GLP-1 enhances glucose-stimulated insulin release, is able to restore first-phase insulin release, decreases glucagon secretion, and slows gastric emptying.456 These effects of GLP-1 occur only when glucose levels rise above fasting concentrations. GLP-1 has been investigated as a treatment for obesity because of its effects on decreasing fluid and food intake. Additionally, GLP-1 has been shown in rodents to delay or prevent the loss of β-cell mass, making it an important treatment consideration for the progressive disease of T2D.457 The 2 most well-studied incretins are GLP-1 and glucose-dependent insulinotropic polypeptide (also known as gastric inhibitory polypeptide). A GLP-1 mimetic, exenatide (ByettaTM), is being used for the treatment of T2D in adults.458 It is given as a twice-daily injection. Side effects include nausea, vomiting, diarrhea, dyspepsia, jitteriness, dizziness, headache, and hypoglycemia, particularly when given with sulfonylurea. Safety and efficacy have not been established in pediatric patients, but it is likely that the need for twice-daily injection and the side effect profile will make this an unacceptable treatment for children with T2D.

Dipeptidyl Peptidase IV Inhibitors

Dipeptidyl peptidase IV is an enzyme that breaks down GLP-1, and its inhibition increases circulating concentrations of GLP-1. This agent has the advantage over GLP-1 mimetics of being effective orally, with metformin or thiazolidenedione once daily. Because it does not affect gastric emptying or satiety, it does not have the salutary effect on weight loss of GLP-1.459 In a placebo-controlled 1-year trial, Dipeptidyl peptidase IV–inhibition treatment improved β-cell function, insulin secretion, and glucose sensitivity, with significant decreases in HbA1c.460

Amylin (Pramlintide)

Amylin peptide is cosecreted with insulin from the cells in response to food ingestion. It suppresses glucagon release, slows gastric emptying, and reduces food intake, thus lowering blood glucose. It is used as an injection before meals, and insulin dose must be reduced a recommended 50% to avoid hypoglycemia. Treatment results in mild improvement in HbA1c and modest weight loss or stabilization.459

Treatment Recommendations

Treatment decisions are based on symptoms, severity of hyperglycemia, the presence or absence of ketosis/ketoacidosis (diabetic ketoacidosis), or severity of dehydration. Signs of dehydration may be less obvious in the obese. Symptomatic youngsters with type 2 diabetes (T2D) may have ketoacidosis at diabetes onset; the management is not different from that for type 1 diabetes (T1D).

Hyperglycemic Hyperosmolar State

Adolescents with undiagnosed T2D are at particular risk for the hyperglycemic hyperosmolar state that results from diuresis induced by high glucose levels and severe hyperosmolar dehydration. Hyperglycemic hyperosmolar state (HHS) carries a high mortality.373,461-467 Eight obese African American youth thought to have died from diabetic ketoacidosis (DKA) at onset of T1D actually had previously unrecognized T2D; 7 of them met the criteria for HHS and not DKA.461,463 As with DKA, hypokalemia can ensue with correction of acidosis. Hypokalemia resulting from failure to replace or from administration of bicarbonate was responsible for arrhythmia and cardiac arrest in 3 patients and may have contributed, along with hypophosphatemia, in 2 others. One patient developed cerebral edema and another had pancreatitis and renal failure. Fourtner et al.464 reported 1 death at onset of T2D in their series, as a result of multisystem organ failure from HHS, while another patient survived rhabdomyolysis, acute renal failure, and pancreatitis. Another report described 6 male adolescent patients with hyperthermia and rhabdomyolysis at diabetes onset. Of the 5 who were obese, 3 met the criteria for HHS, and 2 of these died, 1 with intractable arrhythmia and the other with multisystem failure (cardiac, pulmonary edema, renal).462 A high index of suspicion is needed for T2D with dehydration in a sick obese adolescent, and rehydration needs to be vigorous, as these patients die of persistent shock, with continuing hypernatremia, hypotension, renal failure, and multisystem organ failure due to hypoperfusion.

Outpatient Management

A decision tree for outpatient management is given in Figure 544-5. Metformin should be the first oral agent used, because it is associated with similar HbA1c reductions to sulfonylureas and thiazolidinediones, with minimal risk of hypoglycemia and without weight gain. Furthermore, there may be greater effect on reducing LDL cholesterol and triglyceride levels than with a sulfonylurea. With failure of monotherapy using metformin over a 3- to 6-month period, sulfonylurea, meglitinide, or insulin can be added. Thiazolidinediones may be used in older adolescents. Combination formulations may result in better compliance. It is important to counsel adolescents with T2D about sexuality and pregnancy and to provide contraceptive advice, as necessary, not only because metformin and thiazolidinediones may restore normal periods467 but because oral agents should not be used during pregnancy. Although routine self-monitoring of blood glucose may not be needed as frequently as with T1D, frequent monitoring is needed during periods of acute illness, during dosage adjustment, or with symptoms that indicate hyper- or hypoglycemia. It is also necessary to monitor for asymptomatic hypoglycemia in those individuals who are taking insulin or sulfonylureas. The frequency of routine self-monitoring of blood glucose needs to be individualized, but should include a combination of fasting and postprandial measurements. Assessments of HbA1c concentration should be done at least twice a year and, if metabolic control is unsatisfactory and requires treatment adjustment, every 3 months. The involvement of a dietitian with skill in the management of nutritional problems in children with diabetes is essential; the dietary recommendations need to be culturally appropriate, sensitive to family style and resources, and need to be understood by all caregivers.

Treating Comorbidities

Monitoring lipids and urinary albumin excretion and examining the retina in those with T2D should begin at the time of diagnosis.8 This contrasts with the recommendations for T1D where regular monitoring for complications is not recommended until adolescence and 3 to 5 years after the diagnosis of diabetes.

Lipid levels and urine albumin excretion should be measured shortly after diagnosis and annually, or more often if there is abnormality and treatment effects need monitoring. Exercise, weight loss, and glycemic control may be sufficient to correct hyperlipidemia. Recommendations should be for the reduced-fat diet consistent with step 2 American Heart Association guidelines. Lipid-lowering medications should be added if lipids do not improve satisfactorily after 3 to 6 months of dietary and diabetes control efforts. HMG-CoA reductase inhibitors (statins) are the most commonly used lipid-lowering agents in pediatric patients; they are contraindicated in pregnancy or if there is risk of pregnancy.469

Hypertension is an independent risk factor for the development of albuminuria, retinopathy, and cardiovascular disease in T2D in adults. Its importance is emphasized by the experience of the United Kingdom Prospective Diabetes Study, in which hypertension control was more important than blood glucose control in reducing the risk of cardiovascular disease.440 Blood pressure should be measured at diagnosis and at least quarterly and compared to standards appropriate for age and height percentiles.470 Elevations must be treated aggressively if there is persistent elevation above the usual percentile for the child or above the 90th percentile for either systolic or diastolic pressure. Angiotensin-converting enzyme inhibitors are the initial drug of choice.170 As with type 1 diabetes, many physicians use angiotensin-converting enzyme inhibitors prophylactically in patients with type 2 diabetes.