Case Based Pediatrics For Medical Students and Residents
Department of Pediatrics, University of Hawaii John A. Burns School of Medicine
Chapter XV.1. Diabetes Mellitus
Greg Y. Uramoto, MD
December 2002

Return to Table of Contents

This is a 9 year old boy who has enjoyed his usual state of good health until his polyuria started 2 months ago. He began to lose weight and reported worsening nocturia over this same period. His appetite increased although lately he has more episodes of stomachaches. Today, he had a noticeably sweet smell to his breath and he was breathing faster than usual so his mother brought him to his pediatrician.

Exam: VS T 37.0, RR 44, P 92, BP110/60, oxygen saturation 100% on room air. His weight was 25 kg (25%tile). He is alert and cooperative. His skin is warm to his wrists and ankles. His oral mucous membranes are tacky. His capillary refill is 3 seconds over his chest. His skin was otherwise normal. His thyroid gland is approximately 1.5 times the normal size. His heart rate is regular. He is slightly tachypneic with clear breath sounds. His reflexes are normal. His abdomen has normal bowel sounds has no tenderness. His genitalia and pubic hair are in Tanner stage I. The rest of the physical examination is unremarkable.

His pediatrician suspects new onset diabetes mellitus. A urine dipstick in the office shows 4+ glucose and 2+ ketones. No other dipstick abnormalities are noted. He is clinically stable. He is hospitalized for further management and treatment. His initial lab studies show Na 132, K3.3, Cl 99, bicarb 11, glucose 380, BUN 21, creatinine 0.4. He is started on an IV fluid infusion and subcutaneous insulin.

Prior to the purification of insulin, type 1 diabetes mellitus was uniformly lethal. Although we have made significant strides in the evaluation and management of diabetes, it remains a significant health problem in the general population. In the pediatric subset of the population, type 1 diabetes mellitus is especially challenging since so many factors need to be balanced. The basics on balancing all of these factors will be covered in this chapter.

In United States, the overall risk of developing type 1 diabetes mellitus is 0.2-0.4 percent in Caucasians. In siblings of patients with type 1 diabetes, the risk is approximately 6%. Children of fathers who have type 1 diabetes mellitus have a 6% risk of developing the problem. Children of mothers with type 1 diabetes mellitus have only a 3% chance of developing the problem. The overall incidence of the disease is 18.2 in 100,000/year among people less than 20 years old. The incidence is much higher in Scandinavian countries than in Asian countries.

The National Diabetes Data Group in 1979 divided the heterogeneous condition of diabetes mellitus into two main groups. Type 1 diabetes mellitus has also been called insulin-dependent (IDDM), juvenile onset, ketosis prone, or brittle diabetes. In this type of diabetes mellitus, islet cells are destroyed by an autoimmune process and insulin that these islet cells produce must be replaced. With our current understanding, type 2 diabetes mellitus is primarily an insulin resistant state with a gradual decrease in beta cell function. It was formally known as non insulin-dependent diabetes (NIDDM), adult-onset diabetes, or stable diabetes. Clinical diabetes mellitus can also result from a large number of pathologic processes. Beta cell destruction due to pancreatitis, cystic fibrosis, or surgery can lead to an insulinopenic state that requires insulin injections. Medications including streptozocin, cyclosporin, and corticosteroids can also lead to clinically high blood sugars.

Approximately 2 percent of the American population have some form of diabetes mellitus. Approximately 85 percent of all patients (adults and children) with diabetes mellitus are categorized as type 2. Since type 2 diabetes mellitus is often very subtle, the number of undiagnosed cases of diabetes mellitus is significant. The other 15 percent of patients with diabetes mellitus nationwide are categorized as type 1. In the pediatric population, type 1 diabetes makes up a larger proportion of the cases. Although our estimates are quite crude, some centers report that approximately 98 percent of their children with diabetes have the Type 1 variety. This estimate will certainly be revised in the future as we recognize more type 2 diabetes in children.

Insulin is the primary hormone that suppresses hepatic glucose production, proteolysis, and lipolysis. It is secreted in a biphasic manner in response to glucose. The first phase of insulin release is followed by a nadir and then by a relatively prolonged second phase of insulin release. Catecholamines, cortisol, growth hormone, glucagon, and gastrointestinal hormones among other hormones modulate the insulin response to glucose.

Due to the portal circulation in the gut, blood draining the islet cells of the pancreas goes to the liver before returning to the heart. This portal circulation exposes the liver to an immediately high concentration of insulin soon after a meal. When treating diabetes with exogenously administered insulin into the systemic circulation, we need to remember that this does not duplicate the physiologic state.

Insulin is an anabolic hormone that increases the transport of glucose into cells. With this transport, insulin will decrease the serum glucose. At the cellular level, glucose does not act alone. In reality, the insulin/glucagon ratio is important for determining the body's response to insulin. A high insulin state will induce glucose uptake and inhibit amino acid release in muscle cells. In the liver, insulin will decrease glucose release and decrease ketone body formation. In fat cells, insulin will increase glucose uptake and decrease lipolysis.

In our most current models, type 1 diabetes mellitus is an autoimmune disease. In our current understanding of the problem, people with type 1 diabetes mellitus have an underlying genetic predisposition to developing diabetes. On top of this predisposition, they are exposed to an environmental insult that triggers the immune response. In this way, not everyone who is genetically susceptible to type 1 diabetes mellitus will develop the problem. The identical twin of the patient with type 1 diabetes mellitus has a 25 to 50 percent risk of developing the problem in their lifetime.

Important genes for transmitting this susceptibility to diabetes include the human leukocyte antigens (HLA) that allow for some of the communication between white blood cells. The HLA-DR and HLA-DQ molecules are especially important. These are antigen-presenting structures that T-cells recognize. The antigens in these presenting molecules are the targets for the immune response. Mutations that lead to defects in the structure of this antigen presenting molecule predisposes to type 1 diabetes mellitus. One of the important residues in the structure of the HLA molecules is at position 57. Homozygosity for aspartic acid at this site confers nearly 100% protection against type 1 diabetes. Conversely, a non-aspartic residue at this spot can lead to a nearly 100 fold increase in the incidence of disease.

On top of this genetic predisposition, an environmental insult is likely to be required for the development of diabetes. The environmental factors are quite varied and we are only now beginning to isolate some of them. Congenital rubella cases provide compelling evidence that some of these environmental triggers are viral proteins. Approximately 20 percent of babies with congenital rubella will develop type 1 diabetes mellitus. Most of these cases do not appear until adolescence. Other viruses such as Coxsackie virus, cytomegalovirus, and hepatitis viruses have been implicated.

Polyuria, polydipsia, weight loss, fatigue, polyphagia, anorexia, deteriorating school performance, failure to thrive, and nocturnal enuresis can occur. Clinical symptoms become apparent when the blood sugar rises above the renal threshold and glycosuria induces an osmotic diuresis. Insulinopenia allows hormone sensitive lipase to cut long fatty-acid chains into two carbon acetate fragments which are converted to ketoacids. Patients will present in varying degrees of decompensation as the serum pH decreases and as the dehydration progresses. New onset type 1 diabetes will frequently present with diabetic ketoacidosis of varying severity.

Secondary enuresis, unexplained weight loss, and polyuria should raise suspicions about diabetes. Testing should include random glucose levels, electrolytes, and ketones. The measurement of hemoglobin A1C and insulin can also be helpful. 90% of children will have elevated anti-insulin, anti-islet cell, or anti-GAD (Glutamic acid dehydrogenase) antibodies. Rarely, an IV or oral glucose tolerance test to evaluate for the degree of insulin-producing capacity may be considered in borderline cases, mostly to confirm type 2 diabetes.

Normal glucose levels should be <126 mg/dl in the fasting state. "Fasting" for this purpose should include no caloric intake for at least 8 hours. A random glucose of >200 mg/dl and elevated ketones in the urine or serum in the presence of classic symptoms of diabetes strongly supports the diagnosis of diabetes.

There is no single test that will definitively differentiate between type 1 and type 2 diabetes. Insulin levels should be high in the presence of high serum glucose levels. At least one of the above antibody tests are usually positive in type 1 diabetes. The clinical course is also helpful in differentiating type 1 and type 2. In the case of type 1 diabetes, the capacity to make insulin will decrease over the course of several months as islet cell destruction advances. In type 2 diabetes, the beta cell function is lost over the course of years to even decades.

Diabetic ketoacidosis (DKA) is a complex subject on its own, beyond the scope of this chapter. DKA occurs in an insulin deficient state when cellular starvation of glucose occurs. Despite hyperglycemia, glucose cannot be transported into many cells in the absence of insulin. Therefore, cellular energy metabolism utilizes lipolysis with resultant organic acid, ketone formation (i.e., ketoacidosis) and visible lipemia (blood samples may appear visibly turbid). Patients classically present with severe dehydration, vomiting, deep respirations (respiratory compensation for metabolic acidosis) and a ketotic odor to the breath. Management begins with IV fluids and insulin (0.1 u/kg) administration. Factitious hyponatremia is present with extreme values of hyperglycemia as seen in DKA. This will correct on its own as the glucose level normalizes. Complications such as cerebral edema may occur, the etiology of which is not fully understood. Fluid administration has not been demonstrated to be the cause of cerebral edema; however caution should be exercised in the rate of administering IV fluids. As a rule of thumb, it may be conservative to infuse normal saline at a rate of LESS than 10 cc/kg/hour (i.e., less than a fluid volume correction of 1% of body's weight per hour). A relatively stable patient can receive fluids at maintenance or 1.5 times maintenance. However, severely dehydrated or patients in shock will need more aggressive fluid administration rates. Cerebral edema occurs because the high extracellular glucose levels result in osmotic gradients which pull water from the cells creating cellular dehydration (similar to how a grape changes into a raisin). At some point, this cellular dehydration results in irreversible cellular injury. A less common phenomenon is known as the hyperosmolar non-ketotic state which is characterized by glucose levels in excess of 1000 mg/dl (DKA glucose levels are typically 300 to 800 mg/dl), little or no ketones (most easily assessed by urine dipstick for ketones), and depressed sensorium or coma. The hyperosmolar non-ketotic state has a substantial mortality rate in the 25% range especially if the patient presents in coma, because cerebral cells are subjected to greater degrees of cellular dehydration and a higher risk of irreversible injury. Why does DKA occur in most patients while the hyperosmolar non-ketotic state occurs in others? The answer is uncertain, but it may have something to do with differences in lipid metabolism between individuals.

Once patients are stabilized on the hospital floor, subcutaneous insulin is currently the standard treatment. This insulin can be delivered with injections or a pump. The back-up system for the pump, however, is injections so everyone should learn the basics of subcutaneous insulin injections.

The key to managing diabetes is to balance the factors that increase the blood sugar with the factors that decrease the blood sugar. The most important of these factors include insulin, diet, and exercise. Parents of children with diabetes need to learn enough about diabetes to take care of their children at home. At the very least, they will need to learn about: insulin injections, the types of insulin, blood glucose monitoring, the influences of diet on blood sugar, the influence of exercise on blood sugar, the influence of illnesses on blood sugar, symptoms of hypoglycemia, and the proper response to high and low blood sugars. Without this basic knowledge, patients cannot be discharged home. Several education sessions are usually needed to cover the large amount of information. With so many important aspects to the treatment of diabetes, a team approach that includes dietitians, counselors, diabetes educators, and doctors usually works best.

The insulin program should be tailored to match the family. The sophistication of the family and the ability of the child to give themselves their own shots are important considerations. Two common insulin programs include the mixed-split program and the multiple daily injection program.

In the mixed split program, two insulin types are mixed together and given in two injections. The two insulin types usually include an intermediate-acting NPH and a short-acting insulin such as Lispro or regular. Each shot is supposed to take care of two different meals so the morning shot will take care of breakfast and lunch. In this way, the child does not need to give a shot in school. This is important if the child is not old enough to give his/her own shots.

The multiple daily injection (MDI) program uses a long-acting insulin like Ultralente or insulin glargine to mimic the physiologic baseline of insulin. On top of this baseline, short acting insulin is given with each meal. This program is fairly flexible and usually leads to better control of the blood sugars.

Insulin doses should be tailored to the patient's needs. One unit per kilogram of insulin per day can be used as a starting dose of insulin in someone who presents in severe DKA. Careful monitoring of the glucose levels is required to adjust the doses on a daily basis while they are in the hospital. Because the insulin shots are not physiologic, we may need to tolerate a high post-prandial sugar in some patients. Some children and infants are continuously post-prandial so their blood sugar control is often quite complex.

The goals of treatment should also be tailored to the family and the patient. Aiming for excessively tight blood sugar control with a complex insulin program will likely fail in children with complex social issues at home. With this in mind long-term management would include getting as many blood sugar levels into a "goal range" as reasonably feasible. A typical goal range for infants and toddlers is between 100 and 200 mg/dl. A typical goal range for 5-11 year old children is between 80 and 180 mg/dl. For older children blood sugars in the normal range are reasonable. Hemoglobin A1C is a measurement of glycosylated hemoglobin. This level reflects the glycemic control over the previous 3 months. Goals for the hemoglobin A1C values should also be tailored to meet the needs of the family and the patient. In general, lower hemoglobin A1C values are desirable, but the incidence of hypoglycemia is important. Hemoglobin A1C values that are less than 8 are often attainable in elementary school children.

To achieve these lower hemoglobin A1 C values, adjusting the insulin doses are mandatory. Most families can learn enough about diabetes to adjust the insulin doses themselves. In the MDI program, trends in the blood sugars that are obtained several hours after the meal-insulin combination can be used to adjust the insulin doses. A consistently high pre-lunch blood sugar, for instance, would imply that the breakfast insulin should be increased. Other factors that influence the blood sugar should also be considered. The insulin dose should be increased if the meal was not excessive and if the patient was not particularly active.

We are still learning about the pathophysiology of type 2 diabetes. There is some debate about whether insulin resistance or decreased insulin release is the initial problem. Both of these problems occur and the effects of the relative insulinopenia can be found in utero. Adults with type 2 diabetes are much more likely to have had an intrauterine growth retardation than the adults without type 2 diabetes. This is not surprising given the importance of insulin in the growth of fetuses. The early stages of type 2 diabetes are characterized by relatively normal fasting glucose levels but elevated post-prandial blood sugars. This occurs since the insulin that is available can eventually lower the blood sugar levels but cannot take care of the glucose load soon after a meal. As the disease progresses, islet cell function slowly declines in type 2 diabetes and the fasting blood sugars will rise as well. The treatment of type 2 diabetes can be exactly the same as type 1 diabetes. The same insulin program with the same adjustment strategies will work very well in even the early phases of type 2 diabetes. When type 2 diabetes, as patients slowly lose their ability to make insulin, they will more closely resemble people with type 1 diabetes and insulin becomes a necessity. Before this loss of islet cell function, oral medications can be used. Theoretically, sulfonylureas, biguanides, and thiazolidinediones can be used in children as they can in adults. Studies that show efficacy and safety in children are not yet available so they must be used with caution.


1. What is a reasonable goal range for infants, children, and teens?

2. Which type of diabetes is primarily an autoimmune problem?

3. The identical twin of a patient with type 1 diabetes has what risk for developing type 1 diabetes?

4. Which antibodies are often present in type 1 diabetes?

5. What is a hemoglobin A1C?

6. In the early phases of type 2 diabetes, is the fasting blood sugar or the postprandial blood sugar elevated?


1. Drash A. Diabetes in the Child. In: Lifshitz F (ed). Pediatric Endocrinology, third edition. 1995, New York: Marcel Dekker, pp. 555-563.

2. Drash A. Management of the Child with Diabetes Mellitus, In: Lifshitz F (ed). Pediatric Endocrinology, third edition. 1995, New York: Marcel Dekker, pp. 617-629.

3. Laron Z, Karp M. Diabetes Mellitus in Children and Adolescents. In: Bertrand J, Rappaport R, Sizonenko PC (eds). Pediatric Endocrinology: Physiology, Pathophysiology, and Clinical Aspects, 2nd edition. 1993, Baltimore: Lippincott, Williams & Wilkins, pp. 597-617.

Answers to questions

1. Infants: 100-200. Children: 80-180. Teenagers: 70-120.

2. Type 1

3. 50 %

4. Anti islet cell, anti insulin, and anti GAD antibodies

5. HgA1C is the combination of hemoglobin and glucose. It is elevated when the glucose levels are high and it is a good marker for diabetes control.

6. Postprandial.

Return to Table of Contents

University of Hawaii Department of Pediatrics Home Page