Insulin is a 51-amino acid peptide hormone composed of an alpha- and a beta-chain linked by disulfide bonds. Insulin is synthesized in the pancreatic beta cell as a single polypeptide called pre-proinsulin, which contains both chains. As pre-proinsulin assumes its secondary and tertiary structures, the small peptide linking the alpha- and beta-chains (termed C-peptide) is cleaved. C-peptide is released from the beta cell along with insulin, and can be measured in the serum as a marker of endogenous insulin production.

In recent years, a number of new insulin preparations have significantly improved the flexibility of insulin regimens. Insulin preparations are generally classified according to their rapidity of onset and duration of action (Table 15.2). In general, the shorter acting an insulin preparation, the earlier its onset of action, the earlier and more pronounced its peak effect, and the shorter its duration of action [11–13]. Rapid–acting insulin analogues are convenient for meal coverage and correction of hyperglycemia and are increasingly used in this role in place of regular insulin, which has the same structure as the native insulin molecule and is considered short–acting. The sole intermediate–acting insulin is neutral protamine Hagedorn (NPH), which remains in wide use because its duration of action allows for convenient twice-daily dosing. Long–acting insulin analogues provide a nearly constant serum level of insulin; for this reason, they are a common choice to provide a basal level of insulin throughout the day, and have largely replaced older long-acting insulin preparations such as Lente and Ultralente.

If more than one insulin are given at the same time, they are often mixed and given as a single injection (with the exception of glargine and detemir, which cannot be mixed with other insulins). Premixed insulin preparations in fixed proportions are also available that may provide a simpler option for patients who have difficulty with several injections per day.

In a normal individual, the pancreas produces insulin at a low basal rate equivalent to about 0.01–0.02 units/kg/h [14]. Basal insulin is required even in the fasting state to supply the metabolic needs of tissues and to prevent unregulated hepatic production of glucose and ketones. When food is consumed, the ensuing rise in blood glucose triggers a sharp increase in insulin production that allows for disposition of ingested glucose. In type 1 diabetes, the goal of insulin therapy is to maintain the blood glucose as close to normal as possible by approximating the native pattern of basal and prandial insulin secretion. Thus, all insulin regimens consist of three components:

In the outpatient setting, insulin is administered subcutaneously, either as multiple daily injections, or by continuous infusion using an insulin pump.

A multiple daily injection regimen is one in which the patient administers a series of injections over the course of the day, using insulin syringes or an insulin pen. Such regimens follow two general patterns:
This consists of fixed doses of insulin given at specified times of day. Basal insulin coverage is provided by NPH given twice daily (generally at breakfast and at dinner or bedtime). The morning dose of NPH also provides prandial coverage of lunch, since NPH has its peak effect 2–4 h after administration. Rapid- or short-acting insulin is given at breakfast and dinner for coverage of these meals. Correction doses for hyperglycemia are often incorporated by replacing fixed mealtime doses with a sliding scale of rapid-acting insulin, in which the insulin dose varies based on the blood glucose. A split-mixed regimen offers predictable dosage and timing of injections, fewer daily injections than other regimens, and usually does not require an injection at lunch, making it useful for young children with a consistent daily schedule or who would prefer not to receive an injection at school. Disadvantages include lack of flexibility; the fact that snacks are generally not covered with insulin; and that a peaking insulin (NPH) is given at bedtime, putting the patient at risk of nocturnal hypoglycemia. A simplification of this regimen uses a premixed insulin (which includes both NPH and a rapid- or short-acting insulin) given twice daily, at breakfast and dinner.
This is more flexible and more closely mirrors physiologic insulin secretion. Basal insulin is provided by one or two daily doses of a long-acting insulin analogue, which provide a relatively constant insulin level. Prandial doses of a rapid-acting insulin analogue are given for any meal or snack based on its content: the patient determines the amount of carbohydrate to be consumed and calculates the appropriate dose of insulin using an insulin:carbohydrate ratio (e.g., 1 unit of insulin per 25 g of carbohydrate). To correct hyperglycemia, rapid-acting insulin analogue is given based on a correction factor (or sensitivity factor) and a target blood glucose. For example, a patient might receive a correction factor of 1 unit of insulin per 75 mg/dL that the blood glucose exceeds the target of 120 mg/dL. Since insulin sensitivity varies over the course of the day, a patient may have different insulin:carbohydrate ratios or correction factors at different times of day. Advantages of a basal-bolus regimen include closer approximation of physiologic insulin production and increased flexibility in meal timing and content, but at the expense of increased complexity and a greater number of injections per day.

Because of its precision and programmability, an insulin pump allows the most accurate approximation of physiologic insulin production of any currently available insulin regimen. It allows precise dosing and adjustment of the basal rate, insulin:carbohydrate ratio, and correction factor, all of which may vary by time of day, with exercise, or during illness. A pump provides great flexibility in meal and activity patterns, avoids the discomfort of multiple injections, and reduces the risk of hypoglycemia. In addition, “closed-loop” systems are being developed that will link an insulin pump to a continuous blood glucose monitor, with computer algorithms that mimic normal pancreatic function by adjusting insulin administration based on blood glucose. However, like any device, insulin pumps can experience mechanical, electrical, or software malfunctions that may cause failure of insulin delivery. Because rapid-acting insulin analogue has a short duration of action, interruption of insulin delivery by the pump can result in diabetic ketoacidosis within hours. Therefore, a patient on an insulin pump and her caregivers must check blood glucose frequently and remain constantly alert for the possibility of pump malfunction.

Patients with type 1 diabetes must monitor their blood glucose frequently to achieve optimal diabetes control. At a minimum, blood glucose should be measured upon awakening in the morning, before each meal, and at bedtime. More frequent monitoring is necessary during periods of stress, illness, or under other circumstances that may precipitate either hyper- or hypoglycemia. Blood glucose monitoring is most commonly performed with a glucometer using capillary whole blood obtained by fingerstick. However, glucometers have limited precision, need frequent calibration to ensure accuracy, and depend on good user technique. Therefore, any blood glucose value measured by glucometer that is inconsistent with the clinical scenario should be repeated after addressing potential sources of error. In the inpatient setting, a laboratory measurement is useful if uncertainty remains.

The first line of therapy in type 2 diabetes is lifestyle modification. Insulin resistance is closely related to obesity and can be markedly improved with sufficient weight loss. In addition to contributing to weight loss, increasing exercise has an independent effect on improving blood glucose and insulin sensitivity. Improving diet and exercise alone can be quite effective in improving HbA1c [18]. However, lifestyle modifications can be difficult to sustain and compliance is often poor, and many patients are unable to achieve adequate control of their diabetes by these means alone.

In pediatric patients with type 2 diabetes who do not respond adequately to lifestyle modification, metformin is frequently the first medication that is initiated. Metformin’s mechanism is action is not fully understood, but it appears to improve insulin sensitivity. It increases peripheral glucose uptake to some degree, but its effect is most pronounced in the liver, where increased insulin action leads to decreased gluconeogenesis. Metformin is given orally once or twice daily, and extended release formulations are available. The most common adverse effects of metformin are gastrointestinal symptoms, including nausea and diarrhea, which can often be avoiding by starting at a low dose and titrating up slowly.

The most worrisome potential adverse effect of metformin is lactic acidosis, which can be life-threatening. Because metformin is exclusively excreted by the kidney, the risk of lactic acidosis is increased in situations of reduced kidney function such as dehydration, shock, or renal insufficiency. Although lactic acidosis was described with an earlier related drug, phenformin, it is not clear whether metformin carries the same risk. A large meta-analysis failed to show any increased risk of lactic acidosis of metformin over other treatments for hyperglycemia in the ambulatory setting [19]. However, given the potential severity of this side effect, most experts continue to recommend caution around anesthesia and surgery, which may decrease renal perfusion and increase the risk of lactic acidosis. Therefore, metformin should be stopped 24 h prior to an elective procedure, and can be restarted postoperatively once adequate renal function has been ensured. Metformin should also be held prior to the administration of radiographic contrast that may impair kidney function.

These two classes of oral medications enhance endogenous insulin production by directly stimulating insulin release from the pancreatic beta cell. Meglitinides (rapeglinide, nateglinide) are short-acting and a given immediately before each meal. Sulfonylureas (glipizide, glyburide, gliclazide, glimepiride) have various durations of action, and may be taken one or more times daily. In part because of the adverse effect of hypoglycemia, sulfonylureas are used infrequently in children. Hypoglycemia is more common with longer acting formulations and in the setting of other risk factors such as decreased caloric intake, stress, or illness. Therefore, these medications should be stopped on the day of an elective surgical procedure.

Only two agents in this class, pioglitazone and rosiglitazone, are currently available routinely. Neither is approved for use in children, but they are used occasionally in older adolescents who do not tolerate metformin. Thiazolidinediones improve insulin sensitivity by interacting with peroxisome proliferator-activated receptor gamma (PPAR-γ) to increase glucose utilization and decrease glucose production. In adults, these agents are about as effective as metformin as monotherapy, but they are more expensive and have more associated adverse effects: in particular, rosiglitazone has been associated with an increased risk of cardiovascular events. Both agents rarely cause hepatitis, particularly in the setting of other risk factors. More commonly, thiazolidinediones cause weight gain, as well as fluid retention that can lead to peripheral edema and heart failure. These medications should be stopped on the day of an elective surgical procedure.

Patients with type 2 diabetes who are unable to achieve adequate control with lifestyle modifications and oral medication require insulin. However, practices vary as to the precise criteria for starting insulin therapy in pediatric patients with type 2 diabetes: persistent elevation of HbA1c, fasting or postprandial hyperglycemia, ketosis, or symptoms of diabetes despite oral therapy are all potential reasons to initiate insulin. The principles of insulin therapy are similar to those in type 1 diabetes, though there are important differences. First, due to their insulin resistance, individuals with type 2 diabetes generally require substantially higher doses of insulin than those with type 1 diabetes. Second, because patients with type 2 diabetes are not absolutely insulin deficient, they generally do not require continuous treatment with insulin to prevent ketoacidosis. The same types of insulin regimens are employed in type 2 diabetes as in type 1, with the exception that insulin pumps are infrequently used. Many patients achieve adequate control with only a basal dose of long-acting insulin analogue. Others use a correction factor (or sliding scale) of rapid-acting insulin analogue to correct hyperglycemia. Split-mixed insulin regimens for type 2 diabetes often consist of two daily injections of a premixed insulin preparation given before breakfast and dinner.

DKA is a severe complication of diabetes that can be life-threatening if not detected and treated urgently. DKA is defined by the presence of the following:

Insulin suppresses glucose production by the liver (glycogenolysis and gluconeogenesis) and oxidation of fatty acids to ketones. When insulin action is insufficient in uncontrolled diabetes, these processes proceed unchecked. The liver produces excess glucose, and tissue uptake of glucose is reduced, leading to hyperglycemia. At the same time, the accumulation of ketones—which are organic acids—leads to metabolic acidosis. Dehydration is an invariable finding in DKA and an important factor in its pathogenesis. Factors contributing to dehydration include osmotic diuresis due to glycosuria, nausea and vomiting due to ketonemia, and respiratory losses due to compensation for metabolic acidosis. Dehydration decreases the glomerular filtration rate, which further worsens both hyperglycemia and ketonemia due to decreased excretion. As dehydration worsens, decreased tissue perfusion may lead to lactic acidosis that can further exacerbate metabolic acidosis.

The primary treatments for DKA are rehydration and insulin administration. During the therapeutic course these are titrated, along with appropriate electrolyte infusions, to steadily close the anion gap, resolve ketosis, and normalize blood glucose and electrolytes. Surgical intervention during DKA should be avoided if at all possible.

It has long been recognized that patients with diabetes may develop a syndrome of severe hyperglycemia, hyperosmolality, and dehydration, but without the severe ketosis and metabolic acidosis that characterize diabetic ketoacidosis. This hyperosmolar hyperglycemia syndrome (HHS) is defined by:

HHS is more common in obese patients, most of whom have type 2 diabetes, but it can occur in type 1 diabetes and has been reported in cystic fibrosis-related diabetes. HHS and DKA are thought to lie on a spectrum of deficient insulin action. While DKA occurs in the setting of severe insulin deficiency, HHS is thought to occur when insulin action is reduced to the point that it can suppress ketogenesis but not control hyperglycemia, which leads to osmotic diuresis and dehydration. As dehydration worsens, renal perfusion declines and less glucose is excreted, causing progressive hyperglycemia and hyperosmolality. Polydipsia may exacerbate hyperglycemia if the patient drinks glucose-containing beverages such as juice or soda.

Patients with HHS are profoundly dehydrated, but because hyperosmolality helps preserve intravascular volume, the true degree of dehydration may be difficult to assess clinically. Confusion, lethargy, and coma are frequent present when hyperosmolality is severe. Acidosis is uncommon (10–30%), and if present is due to lactic acidosis from decreased perfusion rather to ketoacidosis. Death occurs in up to 40% of cases and is usually caused by multiorgan failure due to hypovolemic shock, emphasizing the importance of early and aggressive volume resuscitation. Other complications include renal failure, rhabdomyolysis, hyperthermia, pancreatitis, hypokalemia, and hypophosphatemia. Treatment should take place in an intensive care setting and consists of fluid resuscitation with isotonic saline (at least 40 mL/kg initially), with administration of insulin only after the blood glucose has ceased to fall with further hydration [20].

Hypoglycemia (blood glucose < 70 mg/dL or 3.9 mmol/L) is the most common acute complication of diabetes, and is generally caused by diabetes therapy rather than by the disease itself. Although hypoglycemia is occasionally caused by frank overdose of insulin or an oral insulin secretagogue, in the vast majority of cases hypoglycemia is due to a mismatch between insulin dose and carbohydrate intake. The most common precipitant is a decrease in carbohydrate intake without a corresponding adjustment of insulin doing, such as during a prolonged fast or a gastrointestinal illness. In the hospital setting, patients who are unable to eat in the perioperative period and do not receive sufficient carbohydrate by another route (e.g., intravenous dextrose) are at risk for hypoglycemia if their insulin dosing is not reduced accordingly.

Uncontrolled diabetes predisposes to infection in a number of ways. First, hyperglycemia impairs neutrophil chemotaxis and function. In patients with microvascular disease, decreased perfusion to injured areas may further impair the immune response as well as delay wound healing. With respect to surgical site infections, hyperglycemia may also interfere with collagen structure to impair wound healing and tensile strength. Historically, adults with diabetes have had a 10-fold higher rate of postoperative wound infections than those without diabetes [21]. Adults with a preoperative HbA1c above 7% have an overall twofold increased risk of postoperative infection, including wound infection, pneumonia, sepsis, and urinary tract infection [22]. Therefore, the degree of preexisting glycemic control has important implications for surgical planning in patients with diabetes. Postoperative glycemic control also affects the risk of infection (see Sect. 8).

Sustained hyperglycemia causes abnormal glycation of cellular proteins and metabolic alterations that damage endothelial cells. In patients with uncontrolled diabetes, progressive endothelial injury leads to a variety of vascular complications. Microvascular disease in the retina, glomerulus, and vasa nervorum manifests as diabetic retinopathy, nephropathy, and neuropathy, respectively. Changes in larger vessels lead to macrovascular complications such as atherosclerosis, myocardial ischemia, and stroke. These risks are further magnified by dyslipidemia and hypertension, which are more common in diabetic patients. Once established, the combination of peripheral neuropathy and vasculopathy increases the risk of lower extremity infections. Autonomic neuropathy can cause delayed gastic emptying and gastroparesis.