Individuals with normal glucose tolerance maintain their blood glucose concentration between 60 and 120 mg per dL. Maintenance of glucose within this narrow range is controlled by the degree of tissue insulinization (Fig. 100.1) [18]. This is a function of the amount of insulin available and the responsiveness of target tissues. After eating, blood glucose concentration rises but remains within the normal range as a result of increased insulin secretion. Insulin first promotes the transport of glucose into cells and the repletion of glycogen and protein stores. It then mediates the storage of excess glucose as triglyceride. When absorption of nutrients is complete, the concentrations of all metabolites and hormones return to basal levels.

In the fasting state, two mechanisms keep blood glucose concentration in the normal range, glycogenolysis and gluconeogenesis. Initially, hepatic glycogen is mobilized. If fasting persists longer than 12 to 18 hours, peripheral tissues begin to use free fatty acids for fuel, thereby sparing glucose. A low level of circulating insulin is permissive to the lipolytic release of these fatty acids. At the same time, gluconeogenesis supplies glucose for obligate glycolytic tissues, most notably the central nervous system.

If starvation continues for more than 72 hours, the brain begins to use ketone bodies as an alternative fuel, further sparing glucose utilization [18]. At this stage, a progressive decrease in hepatic gluconeogenesis occurs as a consequence of decreased amino acid release in the periphery. As starvation continues, lactate, pyruvate, and glycerol become the main gluconeogenic precursors in place of amino acids. At all times, a low level of circulating insulin regulates the rate of lipolysis, glucose transport, and gluconeogenesis. Healthy humans are always insulinized to an appropriate degree.

Major surgery and critical illness are physiologically stressful events that provoke complex metabolic responses. Tissue hypoxia and hypoxemia adversely affect normal oxidative phosphorylation, and counterregulatory hormones are secreted. These hormones include epinephrine, norepinephrine, cortisol, growth hormone, glucagon, and various cytokines (e.g., tumor necrosis factor-α). They raise blood glucose concentration, mobilize alternative fuels, and increase peripheral resistance to the effects of insulin. In the ICU, their effects may be further amplified by the concurrent administration of exogenous vasopressors, glucocorticoids, and other drugs that can affect intermediary metabolism.

Stress-induced changes in metabolism normally lead to increased insulin release. This in turn enhances peripheral glucose utilization and inhibits alternative fuel mobilization. In this way, the body resists stress without losing control of the biochemical machinery. In patients with decreased insulin reserves (i.e., diabetes mellitus), failure of this feedback loop produces hyperglycemia and related metabolic complications. To preclude these complications, careful management of insulin, fluid, and electrolytes is necessary.

Diabetes is not one disease but rather a family of syndromes that have in common hyperglycemia resulting from inadequate insulinization. These syndromes vary with respect to
genetics, pathophysiology, and appropriate treatment modalities [19]. Table 100.1 outlines the American Diabetes Association’s (ADA’s) classification system.

All acutely ill patients should have their blood glucose level measured at entry into the ICU and at regular intervals throughout their stay. In the outpatient setting, diabetes is diagnosed by a fasting blood glucose level greater than 126 mg per dL or a glucose level greater than or equal to 200 mg per dL measured 2 hours after a 75-g oral-glucose tolerance test. It is required that this be a persistent condition confirmed by repeating the test on another day. A formal diagnosis of new onset diabetes should be made tentatively during the stress of an ICU admission as hyperglycemia may subsequently resolve. Hyperglycemic ICU patients with no prior history of diabetes should be evaluated for persistence of impaired glucose tolerance after recovery.

The majority of seriously ill patients with hyperglycemia do not have a preexisting diagnosis of diabetes. In one study of 1,200 subjects treated in a medical ICU, 70% of individuals at some time experienced a plasma glucose concentration of more than 215 mg per dL, and only about 17% of these patients had a prior history of diabetes [25].

Whenever the glucose concentration in any patient is greater than about 250 mg per dL, actual or impending ketoacidosis and hyperosmolality must be excluded. Ketoacidosis can be diagnosed on the basis of history, physical findings, and the presence of an anion gap acidosis and ketonemia. Osmolarity can be measured by the laboratory or calculated from the serum concentrations of glucose, blood urea nitrogen, sodium, and potassium (Table 100.1; Chapter 101). Hyperosmolar states in the setting of diabetes are usually associated with severe dehydration, obtundation, and extreme hyperglycemia. Diabetic ketoacidosis and hyperosmolar coma require urgent treatment (see Chapter 101). In this chapter, we describe the goals, methods, and pitfalls of treating intercurrent diabetes mellitus in the ICU when neither ketoacidosis nor hyperosmolar coma is the primary disease process.

Physicians caring for patients with diabetes in an ICU should attempt to determine the type of diabetes, its duration, the presence of diabetic complications, and the degree of previous glycemic control. Patients with type 1 diabetes require insulin treatment at all times; those with type 2 diabetes may or may not require insulin. Patients with diabetes that is secondary to some other disorder (Table 100.2) require diagnosis and treatment of the precipitating factors.

Long-standing diabetes is associated with complications that tend to be worse in patients with either type 1 [26] or type 2 [27] disease that is poorly controlled. These sequelae of diabetes complicate the management of critical illness. Diabetes is a leading cause of cardiovascular and peripheral vascular disease. Assessments of both cardiac function and peripheral circulation are advisable for all patients with diabetes. Diabetic neuropathy can affect the autonomic nervous system, with implications for management of blood pressure, heart rate, and voiding. Autonomic neuropathy should be suspected in patients with an abnormal pupillary response to light or absence of heart rate response (R-R interval change in the electrocardiogram) during Valsalva maneuver. Assessment of kidney function should include a urinalysis for protein; albuminuria precedes abnormalities in blood urea nitrogen and creatinine levels. Diabetic eye disease is not a contraindication to anticoagulation, but its severity should be documented before instituting therapy.

A history of poor control should alert the clinician to other potential problems. Poorly controlled diabetes may imply poor nutrition. This has important implications for resistance to infection and wound healing; nutritional assessment and vitamin repletion may be required. Thiamine in particular is a critical cofactor in carbohydrate metabolism, and patients with uncontrolled diabetes may be thiamine deficient. Occult infections to which individuals with diabetes are particularly susceptible include osteomyelitis, cellulitis, tuberculosis, cholecystitis, gingivitis, sinusitis, cystitis, and pyelonephritis [11,12,28,29]. Patients with type 2 diabetes are frequently hyperlipidemic and may develop pancreatitis when poorly controlled [30].

Systems are now available for accurately measuring a patient’s blood glucose levels at the point of care and with near-immediate results using hospital-grade bedside machines, and should be on hand in all emergency departments, ICUs, operating suites, and recovery rooms [31,32]. It should be noted that glucose meter measurements can be influenced by hematocrit, creatinine concentration, plasma protein concentration, and PO₂, all of which can be very abnormal in ICU patients [33,34,35]. Test strips based on glucose dehydrogenase-pyrroloquinoline quinone (GDH-PQQ) methodology can falsely elevate blood glucose readings in patients receiving maltose or icodextrin by more than 100 mg per dL [36]. Maltose is used in a number of biological preparations and peritoneal dialysis solutions may contain icodextrin. Extremely elevated blood glucose concentrations may be outside the range accurately measured by the bedside monitor and should be verified with a serum sample sent to the laboratory [37]. In general, however, therapy should not be delayed by waiting for confirmatory results of laboratory glucose concentration.