18 Hypoglycemia
The American Diabetes Association Workgroup on Hypoglycemia set the alert level for hypoglycemia at plasma glucose concentrations ≤70 mg/dL (3.9 mmol/L) in patients with diabetes mellitus. When the plasma glucose concentration is less than this threshold value, actions should be undertaken to prevent clinical/symptomatic hypoglycemia.1 Clinical/symptomatic hypoglycemia is characterized by the Whipple triad: (1) symptoms of hypoglycemia, (2) simultaneous low blood glucose concentration, and (3) relief of symptoms with the administration of glucose. These symptoms may be neurogenic/autonomic or neuroglycopenic (Table 18-1). Symptoms of hypoglycemia are similar in type 1 and type 2 diabetes.2 Elderly patients report fewer neurogenic/autonomic symptoms.3 They and all other patients with “hypoglycemia unawareness” have a sevenfold increased risk of severe hypoglycemia. Episodes of hypoglycemia in these patients tend to be recurrent and unpredictable.4 “Hypoglycemia unawareness” is the loss of autonomic warning symptoms of developing hypoglycemia. Likely pathogenic mechanisms for hypoglycemia unawareness include recurrent exposure to hypoglycemia, leading to increases in brain glucose uptake and possibly reduced β-adrenergic sensitivity.5 Fortunately, scrupulous avoidance of hypoglycemia for a period of weeks to months restores hypoglycemia awareness.6,7
TABLE 18-1 Symptoms of Hypoglycemia
| Neurogenic | Neuroglycopenic |
|---|---|
| THE RESULT OF AN AUTONOMIC RESPONSE | THE RESULT OF BRAIN GLUCOSE DEPRIVATION |
| Blood glucose <55 mg/dL (3.7 mmol/L) | Blood glucose <45 mg/dL (2.5 mmol/L) |
| Cholinergic: hunger, sweating, paresthesias | Cognitive impairment |
| Behavioral change | |
| Adrenergic: tremor, palpitations, anxiety | Psychomotor abnormalities |
| Seizure and coma |
In critically ill patients, however, sedation strongly masks symptoms, so one can only rely on frequent and accurate blood glucose measurements to detect hypoglycemia. The most commonly used definition of hypoglycemia during critical illness is a plasma glucose concentration below 40 mg/dL (2.2 mmol/L) in the absence of symptoms.8–10 Most reflectance blood glucose meters in home and hospital use have poor precision at low levels of blood glucose.11 Capillary blood glucose testing may not be sufficiently reliable to guide management of blood glucose levels in critically ill patients.12 The use of arterial blood samples for glucose measurements is recommended. However, anemia in critically ill patients can result in falsely elevated blood glucose measurements and mask hypoglycemia when using these blood glucose meters. Also, the recently developed continuous interstitial glucose monitoring system13 and the noninvasive GlucoWatch Biographer14 are less effective at detecting low blood glucose levels and can have a delayed response to low blood glucose concentrations. Therefore, for diabetes patients, the laboratory measurement of a low plasma glucose concentration in the presence of appropriate symptoms remains the most reliable way to diagnose severe hypoglycemia. In the ICU, measurements of arterial blood glucose concentration using modern blood gas analyzers approach the accuracy of conventional laboratory methods.8,12
Specific characteristics of the patient can also determine whether hypoglycemia will be symptomatic or increase the risk of hypoglycemia. For example, a precipitous fall from hyperglycemia to euglycemia in a patient with diabetes can produce hypoglycemic symptoms.15 In contrast, hypoglycemia with glucose levels as low as 30 mg/dL (1.7 mmol/L) can occur asymptomatically during fasting in normal women and during pregnancy.16 Some ICU patient populations, such as those with liver or renal failure and septic shock, are at higher risk for hypoglycemia.17 The characteristics of the hypoglycemia itself (absolute level, duration) and its treatment (avoiding overcorrection) also play a significant role (Table 18-2).
TABLE 18-2 Risk Factors Involved in Hypoglycemia
| Hypoglycemia | Patient |
|---|---|
| Level of hypoglycemia | Liver failure |
| Duration | Renal failure |
| (Over)correction of hypoglycemia | Sepsis or shock |
| Reperfusion damage | Prior history of diabetes mellitus |
Incidence of Severe Hypoglycemia
A retrospective study of adults requiring hospitalization indicated that 0.4% of acute medical admissions per year are hypoglycemia related.18 Severe hypoglycemia (i.e., with symptoms severe enough to require assistance) occurs commonly in patients with type 1 diabetes.19 In type 2 diabetes, even with intensive therapy, the risk is probably 100-fold less. Over 6 years of observation in the United Kingdom Prospective Diabetes Study, severe hypoglycemia was reported in 2.4% of patients treated with metformin, 3.3% of those treated with a sulfonylurea, and 11.2% of those treated with insulin.20 As insulin usage among patients with type 2 diabetes increases, it is inevitable that severe hypoglycemia will become more common in daily practice.
With the introduction of tight blood glucose control during ICU stay,8 the incidence of blood glucose values below 40 mg/dL (2.2 mmol/L) has been reported to range from 5.1% to 18.7% of patients, depending on the targeted level of blood glucose control and the patient population under study.8,9 With the use of accurate glycemia measurement methodologies and algorithms that advise frequent blood glucose measurements (i.e., every 1–4 hours), the incidence and impact of these brief episodes of hypoglycemia should be minimized.17
Physiologic Barriers Against Hypoglycemia
The central nervous system (CNS) relies primarily on glucose for the generation of cellular energy. Cells in the CNS have endogenous glucose reserves that are sufficient for only minutes if the supply of glucose from the bloodstream is inadequate. In addition, neurons are unable to synthesize glucose. Finally, the brain cannot use fuels other than glucose during acute hypoglycemia.21 Hence, when the brain is acutely deprived of glucose, serious neurologic dysfunction occurs. Accordingly, the body has several mechanisms to maintain the plasma glucose concentration within the narrow range of 60 to 140 mg/dL (3.3–7.7 mmol/L) in both the fed and fasting states. When glucose use exceeds glucose production, the brain senses decreasing blood glucose levels and activates counterregulatory pathways.22 The glucose threshold for activation of these mechanisms is approximately 67 mg/dL (3.6 mmol/L), but this setpoint can be altered by recent hyperglycemia or antecedent hypoglycemia. As glucose levels decline, the first counterregulatory mechanism activated is the suppression of endogenous insulin secretion.23 Next in the hierarchy of responses is the release of two hormones, glucagon and epinephrine, that antagonize the action of insulin. These hormones activate glycogenolysis and gluconeogenesis and stimulate fatty acid oxidation and protein breakdown to provide substrates for gluconeogenesis. With more severe or prolonged hypoglycemia (>3 hours), increases in growth hormone and cortisol release raise blood glucose levels.
The physiologic responses to hypoglycemia and the glucose threshold at which they occur can be modulated. In type 1 diabetes, the glucagon response to hypoglycemia is lost within 3 years after diagnosis, rendering patients dependent on epinephrine-mediated counterregulation and making them more vulnerable to prolonged episodes of severe hypoglycemia. Exposure to antecedent hypoglycemia diminishes the counterregulatory response to a subsequent episode. The brain adapts to antecedent hypoglycemia by increasing glucose uptake so that a more profound hypoglycemic stimulus is required to trigger sympathoadrenal activation and autonomic symptoms.24 The level of glycemic control also affects counterregulatory thresholds. With strict glycemic control, epinephrine release is not triggered until a lower glucose level is reached.25,26 Conversely, diabetic patients with poor glycemic control can experience hypoglycemic symptoms when the blood glucose concentration decreases to lower values within the normal or even hyperglycemic range.27
Sequelae
Although severe hypoglycemia induces marked cognitive dysfunction, most patients recover rapidly and completely. The effect of repeated severe hypoglycemia on cognitive function in adults is controversial.28,29 Although focal neurologic symptoms secondary to severe hypoglycemia occur occasionally, severe and permanent cognitive impairment is usually the result of protracted hypoglycemia, often in association with excessive alcohol consumption. The neuronal regions that are particularly vulnerable to hypoglycemia are the cerebral cortex, the substantia nigra, the basal ganglia, and the hippocampus.
The long-term neurologic effects of hypoglycemia during critical illness are poorly delineated.17 It appears that brief episodes of hypoglycemia do not cause severe acute brain damage. A recent nested case-control study using more sophisticated neurocognitive tests showed that hypoglycemia mildly aggravated critical illness–induced neurocognitive dysfunction, notably the visuospatial domain.30 This association, however, could not be dissociated from an effect of hyperglycemia or of glucose variability, as the patients who experienced hypoglycemia were also those with more severe hyperglycemia and greater glucose variability.
The overall mortality from severe hypoglycemia is unknown. The mortality rate from alcohol-induced hypoglycemia may be as high as 10% in adults.31 About 2% to 4% of deaths in patients with type 1 diabetes have been attributed to hypoglycemia. Severe hypoglycemia is the cause of unexpected overnight deaths in young diabetic patients.32 It may be explained by the impairment of hormonal responses to hypoglycemia during sleep, resulting in sudden cardiac arrhythmias.
The association of hypoglycemia and mortality during critical illness is very controversial.33–35 Not only is hypoglycemia more frequent in the most severely ill patients (e.g., those with hepatic or renal failure or septic shock), these spontaneous hypoglycemic episodes also more strongly correlate with mortality risk than hypoglycemia induced by intensive insulin therapy. Nevertheless, as a quality-control measure, intensive insulin therapy in the ICU should be implemented with meticulous monitoring of the incidence of hypoglycemic episodes. The importance of careful monitoring of blood glucose concentration is further emphasized by the demonstration of a tight correlation between blood glucose variability and mortality.36
