Hypoglycemia and Hypoglycemic Syndromes

Published on 28/03/2015 by admin

Filed under Endocrinology, Diabetes and Metabolism

Last modified 28/03/2015

Print this page

rate 1 star rate 2 star rate 3 star rate 4 star rate 5 star
Your rating: none, Average: 0 (0 votes)

This article have been viewed 6491 times

Chapter 21

Hypoglycemia and Hypoglycemic Syndromes


Hypoglycemia, a condition of low plasma glucose, has many varied causes, which are discussed below. The definition of hypoglycemia is often debated, but in this chapter, the glucose values adopted by the hypoglycemia working panel of the American Diabetes Association1 will be used. Hypoglycemia will be defined as any plasma glucose value below 70 mg/dL (3.9 mmol/L). Severe hypoglycemia is reserved for occasions when the plasma glucose is very low (usually below 50 mg/dL, or 2.9 mmol/L) and is accompanied by significant neurologic deficits.

Hypoglycemia is associated with significant morbidity and can be fatal.2,3,9 Hypoglycemia occurring in a nondiabetic individual always warrants attention and should be investigated (Table 21-1). Before one embarks on a comprehensive workup for possible hypoglycemia, certain pitfalls need to be considered. Whole blood glucose values are 10% to 15% lower than plasma glucose levels. Thus, the lower limit of normal for a whole blood glucose value would be about 60 mg/dL (3.3 mmol/L). Keeping this in mind, it is important to determine whether the glucose value is blood or plasma. This will also apply to glucose meters that can provide blood or plasma glucose values. Another important consideration is that mixed venous blood samples can be dramatically lower than arterial glucose values. This will vary depending upon the insulin sensitivity of the individual and the prevailing insulinemia. Thus, mixed venous glucose levels could be 25 to 30 mg/dL (≈1.5 mmol/L) lower than arterial levels in lean healthy individuals during conditions of physiologic hyperinsulinemia. Therefore, hypoglycemic values obtained from mixed venous blood during a 2- to 3-hour oral glucose tolerance test should be interpreted with caution. Similarly, mixed venous glucose measurements obtained during periods of non–steady state (e.g., after a meal, after exercise) can significantly underestimate arterial glucose values.4 In fact, very low glucose levels of between 30 and 50 mg/dL (1.7 to 2.8 mmol/L) have been measured in healthy adults after prolonged exercise.5

Artifactual hypoglycemia can also occur if a blood glucose sample is not collected in a tube containing fluoride and/or oxalate to inhibit glycolysis. Without appropriate sample collection, glucose values can decrease by 10 to 20 mg/dL (≈0.5 to 1.0 mmol/L) per hour at room temperature.6 In addition, even if glucose samples are collected in tubes containing glycolytic inhibitors, artifactually low readings can be obtained if the sample contains large quantities of blood cells,7 if the sample remains unmeasured for many hours, or if the sample is heavily lipemic with triglycerides.8

Physiology of Hypoglycemia

During typical physiologic conditions, the brain requires a constant and adequate supply of glucose. Under normal postabsorptive conditions (e.g., an overnight fast), the brain accounts for ≈65% of whole body glucose uptake. Following feeding, the amount of glucose taken up by the brain can increase, but insulin does not influence brain glucose kinetics in a similar fashion to other organs such as liver or muscle. Although the brain was classically considered an insulin-insensitive organ, recent work has challenged this concept.10 Several studies have elegantly demonstrated that insulin can regulate appetite and feeding mechanisms in rodent models.10 Additionally, insulin administration into areas of the hypothalamus has been demonstrated to regulate hepatic glucose output.11 Furthermore, studies in dogs12 and mice13 have demonstrated direct CNS effects of insulin to amplify autonomic nervous system (ANS) counterregulatory responses to hypoglycemia. Thus, accumulating data indicate that insulin can have direct metabolic effects on certain areas of the brain.

Although typically dependent upon glucose as a fuel, the brain can adapt and utilize other substrates. Thus during periods of fasting, ketone bodies, lactate, and alanine can be used as alternative brain fuels.14 Several studies have demonstrated, experimentally, that high levels of alternative substrates can be infused during acute hypoglycemia with a concomitant reduction in neuroendocrine and ANS responses. This indicates that the brain has the capacity to switch from glucose to alternative substrates in a matter of hours. However, it should be noted that the concentrations of substrate infused experimentally are far higher than levels observed during most physiologic conditions (the exception is levels of ketone bodies that occur during prolonged fasting).

During hypoglycemia, brain glucose uptake falls. The exact glycemic value for the start of decreased blood-to-brain glucose transport is debated but is thought to be around 3.6 to 3.8 mmol/L in humans. As hypoglycemia deepens (−3 mmol/L), blood-to-brain glucose transport becomes rate limiting for brain glucose metabolism. Glycolytic derived lactate and a small amount of stored astrocytic glycogen can provide a short duration of fuel supply. Recent work has estimated that stored glycogen could provide the brain oxidative fuel for about 20 minutes15; based on an estimate that blood-to-brain glucose transport could provide up to 90% of the brain’s oxidative requirements during moderate hypoglycemia,16 it can be seen that the remaining 10% of energy requirements obtained from lactate and glycogen would last only ≈200 minutes, thus emphasizing the need for a continuous and adequate supply of glucose to the brain from the circulation.

Physiologic Responses to Hypoglycemia

As plasma glucose falls, a well-orchestrated response of multiple physiologic mechanisms is activated. The initial defense is a reduction in endogenous insulin secretion. This occurs as plasma glucose levels fall to below 80 mg/dL (4.4 mmol/L). The reduction in endogenous insulin in response to falling plasma glucose is often overlooked in the hierarchy of defenses against hypoglycemia. Data from the Diabetes Control and Complications Trial (DCCT) demonstrate that the presence of even small amounts of C-peptide (i.e., endogenous insulin) are protective against severe hypoglycemia.17 Similarly, experience from recent islet cell transplantations clearly demonstrates that the ability to modulate endogenous insulin levels is also protective against episodes of hypoglycemia.18 As glucose levels continue to fall at or around 70 mg/dL (3.9 mmol/L), a coordinated release of anti-insulin (counterregulatory hormones) occurs. Epinephrine (from the adrenal medulla), glucagon (pancreatic α cells), norepinephrine (sympathetic nerve endings and adrenal medulla), growth hormone (anterior pituitary), and cortisol (adrenal cortex) all have been demonstrated to have protective metabolic effects during acute or prolonged hypoglycemia. It should be noted that the release of neuroendocrine counterregulatory hormones and the inhibition of endogenous insulin secretion occur before a healthy adult can feel any symptoms of hypoglycemia. If plasma glucose continues to fall, at ≈60 mg/dL (3.3 mmol), a series of autonomic (sometimes called neurogenic) signs and symptoms is activated. Autonomic warning responses to hypoglycemia include adrenergic and cholinergic symptoms. Adrenergic symptoms include palpitations, tremor, dry mouth, warmth, and anxiety.19 Cholinergic symptoms include sweating, hunger, and paresthesias.2022 Signs of adrenergic activation include sweating and pallor. If the glucose level continues to fall, neuroglycopenic symptoms are activated at ≈50 to 55 mg/dL (2.8 to 3.1 mmol/L). These include blurred vision, drowsiness, slurred speech, confusion, and difficulty in concentrating. Defects in cognitive function are also apparent at this level of glycemia. If plasma glucose continues to fall, individuals can become drowsy, enter into a coma, and suffer seizures. Alternatively, individuals can become aggressive, which can be difficult to control and can be distinct from their usual personality. If severe hypoglycemia is prolonged, life-threatening events such as arrhythmias, myocardial infarction, and stroke, can occur.23,24 Long-term cognitive damage and even death can occur if very severe hypoglycemia continues for longer than a few hours.25 Although the above represents typical responses to falling glucose levels, it should be appreciated that many patients have idiosyncratic neurologic and symptomatic responses to hypoglycemia that fall outside the classical description. Thus, it is worthwhile to measure the glucose level in anyone who presents with neurologic deficits and/or strange and uncharacteristic behavior. It also should be noted that the typical physiologic responses to hypoglycemia can be modified by a number of factors, including antecedent hypoglycemia, long duration diabetes, age, gender, pregnancy, autonomic neuropathy, and use of certain drugs (Table 21-2). These altered physiologic and pathophysiologic responses are discussed in detail in the following sections.

Neurohumoral Regulation During Hypoglycemia


Insulin is the principal physiologic factor that lowers plasma glucose. Insulin is secreted primarily in response to glucose, but amino acids, nonesterified fatty acids, β2-adrenergic stimulation, and acetylcholine can also activate secretion of the hormone. Insulin secretion can be inhibited by hypoglycemia, insulin itself, somatostatin, and α2-adrenergic activity.75

Insulin is released into the portal vein following secretion from pancreatic β cells. As the result of hepatic extraction, portal vein levels of insulin are approximately twofold higher than peripheral levels. Insulin lowers plasma glucose through several different mechanisms. Endogenous glucose production is inhibited directly first by suppressing hepatic glycogenolysis and then by restraining hepatic and renal gluconeogenesis. Simultaneously, insulin increases glucose uptake into insulin-sensitive tissues (primarily muscle but also liver and adipose tissue). As soon as the rate of exit of glucose exceeds the rate of entry of the substrate into the circulation, then the plasma glucose level starts to fall. Additionally, insulin can suppress glucose production indirectly by restraining lipolysis. (This restricts the flow of glycerol, which is an important gluconeogenic precursor, and of nonesterified fatty acids [NEFA], which provides energy for gluconeogenesis.) Insulin also restricts proteolysis, which suppresses the flow of amino acids as gluconeogenic precursors.

After a typical overnight fast, the liver produces at least 85% of the glucose that enters the circulation. At this point, about 50% to 66% of glucose production is driven by glycogenolysis, and the rest is gluconeogenesis. If the fast were to continue, then most glucose production would be driven by gluconeogenesis, and glycogen stores would be depleted. After 48 hours of fasting, gluconeogenesis is responsible for 80% of glucose output. By 72 hours, gluconeogenesis becomes the only source of glucose production in the body.


Glucagon is released from the α cells in the islet of Langerhans. Similar to insulin, a number of physiologic factors can regulate secretion. These include hypoglycemia, amino acids, catecholamines (epinephrine and norepinephrine via β2-adrenergic mechanisms), and free fatty acids.26 Inhibitors of glucagon release include insulin and somatostatin. The regulation of glucagon release during hypoglycemia in humans is still undecided. Hypoglycemia, per se, can stimulate glucagon release in humans with cervical transection, individuals with a transplanted pancreas, and in vitro pancreas preparations.27 These data would point to the fact that direct α cell sensing of hypoglycemia would be the mechanism responsible for glucagon release. However, convincing data demonstrate that autonomic input (both sympathetic and parasympathetic) into the pancreas also can result in glucagon secretion.27 More recently, a third hypothesis has been proposed, which argues that a reduction in islet cell insulin levels is the mechanistic trigger for glucagon release during hypoglycemia.28

Glucagon rapidly increases glucose production by the liver over a period of 10 to 15 minutes. The initial stimulus for glucose output is provided by an increase in hepatic glycogenolysis. If hypoglycemia continues, glucagon can stimulate hepatic gluconeogenesis, but only if three carbon precursors such as glycerol, lactate, and amino acids are present. Glucagon’s physiologic actions are restricted almost exclusively to the liver.

Epinephrine and Norepinephrine

Epinephrine (adrenaline) is released from the adrenal medulla. Similar to glucagon, the hormone can act rapidly to increase hepatic glucose output via elevations in hepatic glycogenolysis. If hypoglycemia continues and three-carbon precursors are present, epinephrine will stimulate gluconeogenesis in the liver and kidneys.29 Unlike glucagon, epinephrine has important effects on peripheral tissues. Epinephrine can restrict insulin-stimulated glucose uptake in skeletal muscle, which, when quantified in terms of maintaining glucose in the circulation, is greater than the contribution made by any increase in endogenous glucose production. This latter property of epinephrine is especially important in the defense against hypoglycemia that typically is encountered in clinical practice. Unlike the model of rapid hypoglycemia produced during an insulin tolerance test, the usual clinical course involves a slower decline into hypoglycemia and a more protracted duration of low glycemia, which, during the night, can last up to several hours. In the acute induction of hypoglycemia caused by a large bolus of rapid acting insulin, it is the activation of endogenous glucose production that is the primary physiologic defense against hypoglycemia, but in a model of more prolonged hypoglycemia, the restriction to insulin-mediated glucose uptake is paramount.30 Epinephrine’s important metabolic effects, which are mediated via β2-adrenoreceptors, also include stimulation of lipolysis to provide substrate (glycerol) and energy (NEFA) for gluconeogenesis. Additional effects on muscle provide lactate, pyruvate, and amino acids for gluconeogenic precursors.

Norepinephrine has similar metabolic actions to those of epinephrine. However, because 90% of norepinephrine is taken up at the level of sympathetic clefts, and 90% of catecholamine is taken up by the gut and then the liver, the increase in circulating norepinephrine during hypoglycemia is relatively modest (about 50% as compared with the 30-fold elevations that can occur with epinephrine). Therefore, quantifying the effects of norepinephrine at a tissue level in humans is problematic. However, recent work in conscious dogs has demonstrated that a 2.5-fold greater infusion of norepinephrine as compared with epinephrine is needed to produce similar increases in hepatic glucose production.31

Syndromes of Disordered Counterregulatory Responses During Hypoglycemia

Almost immediately after the discovery of the hormone in the early 1920s, hypoglycemia was recognized as an unpleasant and dangerous side effect of insulin therapy. More than 90% of all patients with type 1 diabetes mellitus (DM) have suffered an episode of hypoglycemia. Typically, a patient with type 1 DM will experience 2 to 10 episodes of glucose below 70 mg/dL (3.9 mmol/L) per week and may have glycemic values below 50 mg/dL (2.9 mmol/L) at least 10 times each week. Episodes of major hypoglycemia requiring resuscitative measures from an additional person or hypoglycemic events resulting in seizures or coma have been reported to occur about once every other year to up to three times a year.17 The duration of type 1 DM plays an important role in increasing the frequency of severe hypoglycemia. Recent data from the UK hypoglycemia study group33 demonstrate that the frequency of severe hypoglycemia increases from 110 per 100 patient-years after less than 5 years of insulin therapy to 320 episodes per hundred patient-years after insulin therapy for longer than 15 years.

The risk and frequency of hypoglycemia are considered to be much lower in type 2 DM as compared with type 1 DM. This is undoubtedly true in patients with type 2 DM treated with monotherapy or combination therapy of metformin, a thiazolidineodione, a DPP-4 inhibitor, and/or a GLP-1 agonist. However, the incidence of hypoglycemia increases sharply in longer duration (insulin-deficient) type 2 DM. Recent clinical trials investigating insulin replacement strategies in type 2 DM report hypoglycemia in more than 70% of participants.34 Furthermore, the UK hypoglycemia study group reported that patients with type 2 DM receiving insulin for longer than 5 years had rates of severe hypoglycemia of about 70 episodes per 100 patient-years.33 This rate is slightly higher than the rate of 60 episodes per 100 patient-years reported for patients with type 1 DM in the intensive group of the DCCT.17 Other studies examining the frequency of severe hypoglycemia in type 2 DM have reported rates of one-third or similar to those occurring in type 1 DM. Furthermore, large recent multicenter studies investigating the effects of glucose control on complications of diabetes in type 2 DM have reported a significant incidence and prevalence of hypoglycemia.35,36 Thus, severe hypoglycemia continues to be a significant clinical problem for patients with type 1 DM and longer duration type 2 DM (Fig. 21-1).

Death from hypoglycemia can be considered on the one hand to be rare when one considers the very large number of episodes of hypoglycemia that occur in clinical practice. However, mortality does occur during severe hypoglycemia in both type 1 and type 2 DM. Studies investigating the cause of death in patients with type 1 DM have reported that 2% to 10% may have died because of hypoglycemia.2,3 Similar death rates from hypoglycemia have been reported in patients with type 2 DM. The specific mechanism responsible for hypoglycemia-induced death is not currently understood. Possible suggested causes include cardiac arrhythmias, thrombotic events, and brain death. However, it should be noted that in primates, several hours of profoundly low glucose <1 mmol/L is required to induce irreversible brain damage and death.37

Hypoglycemia-Associated Neuroendocrine and Autonomic Failure

As discussed earlier, the four primary physiologic defenses against falling plasma glucose include inhibition of endogenous insulin secretion, release of glucagon and epinephrine, and symptomatic cues to ingest carbohydrate (Fig. 21-2). Unfortunately in patients with diabetes, all of these physiologic defenses can become defective and/or deficient. In type 1 DM and in long duration type 2 DM, the individual becomes critically insulin deficient, and thus the first physiologic line of defense (modulation of endogenous insulin) becomes lost. After varying periods of duration of disease (≈5 years), the ability of individuals with type 1 DM to release glucagon in response to hypoglycemia is lost (Fig. 21-3). This defect also occurs to a similar extent in long duration type 2 DM (Fig. 21-4).38 The mechanism for this finding is currently under intense investigation in humans. Hypotheses include lack of β cell turnoff, a possible specific autonomic nervous system dysfunction, and another as yet unidentified local signaling defect at the level of the α cell. The defect involved in releasing glucagon in type 1 DM is specific for hypoglycemia as the α cells in these individuals are present in normal number and size. In fact underscoring this point, glucagon responses to other metabolic stressors in type 1 DM such as exercise or amino acid infusions are preserved.39 The fact that glucagon responses are preserved during exercise in type 1 DM is interesting in that exercise is also typically associated with a physiologic “β cell switch off.” Whatever the mechanism, unfortunately after a few years’ duration, individuals with type 1 DM lose two of the four primary defenses against falling blood glucose. This leaves a functioning ANS (sympathoadrenal and sympathetic nervous systems) to serve as the primary defense against hypoglycemia in type 1 DM. In some individuals with type 1 DM, the ability to secrete epinephrine in response to hypoglycemia is preserved and can compensate for the lack of glucagon release.40 However, it is now clear that epinephrine responses to hypoglycemia are significantly reduced in patients with type 1 DM with intensive metabolic control.41 This epinephrine deficiency has been determined to be due to previous episodes of hypoglycemia42 and is separate from the syndrome of classic diabetic autonomic neuropathy that can occur after many years of suboptimal glycemic control. Models of repeated antecedent hypoglycemia have been demonstrated to produce acute reductions (30% to 50%) in epinephrine, pancreatic polypeptide (a marker of parasympathetic nervous system activity), and muscle sympathetic nerve activity (a direct marker of sympathetic nerve system activation) in individuals with type 1 DM, in those with type 2 DM, and in nondiabetic individuals.43 Additionally, recent (within 24 hours) antecedent hypoglycemia has been found to blunt a wide spectrum of neuroendocrine responses such as glucagon, growth hormone, ACTH, and cortisol during subsequent hypoglycemia44 in healthy and diabetic men. Confirming the role of antecedent hypoglycemia in causing blunted counterregulatory responses are a number of studies that prospectively investigated the effects of avoiding hypoglycemia in type 1 DM and following successful removal of an insulinoma.4548 In all cases, there were initial blunted, neuroendocrine, ANS, and symptomatic responses to hypoglycemia. However, when patients with type 1 DM were restudied several months later, all showed improved counterregulatory responses to hypoglycemia, and patients with previous insulinomas had counterregulatory defenses restored to normal. Symptomatic responses, the fourth critical physiologic counterregulatory response, were significantly improved in all studies following a period of hypoglycemia avoidance.4548 The blunting effects of antecedent hypoglycemia on subsequent counterregulatory responses have been termed by Cryer as “hypoglycemia-associated autonomic failure” (HAAF).

Following identification of this syndrome, a great deal of work has been performed in both animal and human models to further elucidate the mechanisms responsible for and the characteristics of hypoglycemia-associated counterregulatory failure. However, it should be appreciated that HAAF does not occur only in type 1 DM. Work from two independent laboratories has determined that this syndrome also occurs in type 2 DM.49,50 Segel and colleagues clearly demonstrated that moderate antecedent hypoglycemia of 50 mg/dL (2.8 mmol/L) can blunt ANS responses to subsequent hypoglycemia in moderately controlled (HbA1c 8.4%) individuals with type 2 DM.49 More recently, Davis and coworkers have demonstrated that even milder antecedent hypoglycemia of only 60 mg/dL (3.3 mmol/L) can blunt ANS responses to subsequent hypoglycemia in patients with type 2 DM with suboptimal (HbA1c ≈10.0%) or intensive glycemic control (HbA1c ≈6.7%).50

The great challenge in determining the mechanisms responsible for HAAF is explaining the simultaneous reduction in ANS-neuroendocrine responses and the change in glycemic thresholds that activate the physiologic defenses against falling plasma glucose. As described earlier, the usual physiologic thresholds for activation of ANS-neuroendocrine and symptom responses during hypoglycemia occur in the plasma glucose range between 50 and 80 mg/dL. In individuals with chronic hyperglycemia, symptoms of hypoglycemia can occur at plasma glucose levels between 90 and 140 mg/dL, depending upon the severity of the prevailing hyperglycemia. Simultaneous measurements of ANS-neuroendocrine hormones are in fact elevated during these symptoms, thus the individuals are experiencing the condition of “relative hypoglycemia.” This syndrome occurs because the thresholds for activation of ANS-neuroendocrine responses have been pushed to a higher plasma glucose level by the chronic hyperglycemia. On the other hand, individuals with intensive glucose control and multiple episodes of hypoglycemia often find that the activation of physiologic responses to hypoglycemia is pushed to a lower plasma glucose level. This dangerous condition, called hypoglycemic unawareness, results in inability of patients to recognize a falling plasma glucose until the value is below 50 mg/dL (2.9 mmol/L). In some individuals, a falling plasma glucose level is not recognized at plasma glucose levels of 30 mg/dL. This reduces the interval between first recognition of hypoglycemia and the onset of serious sequelae (such as coma or seizure). Thus, thresholds for the activation of physiologic defenses against hypoglycemia are labile and can change rapidly. The duration and depth of antecedent hypoglycemia required to induce HAAF have been characterized. Repeated episodes or relatively mild (3.9 mmol/L, or 70 mg/dL) and only brief durations (15 to 20 minutes) of hypoglycemia can independently blunt counterregulatory responses to subsequent hypoglycemia.51 However, one prolonged episode (2 hours) of moderate hypoglycemia (50 mg/dL, or 2.9 mmol/L) is sufficient to induce HAAF within a few hours on the same day.52

Numerous mechanisms responsible for the syndrome of HAAF have been proposed over recent years, with data both supporting and at times contrary to any given hypothesis. Boyle et al. proposed that repeated hypoglycemia increased cerebral glucose uptake in both healthy individuals and patients with type 1 DM, thereby reducing the stimulus for neuroendocrine counterregulatory responses during subsequent hypoglycemia.53 This finding was later challenged by work reporting no increase in blood-to-brain glucose transport following antecedent hypoglycemia.54 Other mechanisms that have been proposed include activation of the hypothalamic-pituitary-adrenal axis,44,55 increases in neurotransmitters such as GABA, and changes in hypothalamic fuel sensors such as glucokinase or AMP kinase (increases and decreases, respectively).56 Additionally, experimental evidence demonstrates that alcohol and opioids can downregulate subsequent ANS and neuroendocrine responses to hypoglycemia.57 Other physiologic mechanisms also have been found to cause forms of HAAF. These include sleep, exercise, and gender (Fig. 21-5).43,58 The association between exercise and hypoglycemia in type 1 DM has been both problematic and perplexing. Hypoglycemia can occur during, 1 to 2 hours after, or up to 21 hours after exercise. Traditionally, the explanation for this phenomenon was either a relative or absolute excess of subcutaneously injected insulin (due to an increase in insulin sensitivity following exercise) and/or incomplete glycogen repletion following exercise. Although these factors are important contributors to exercise-associated hypoglycemia, they cannot explain some of the profound episodes of hypoglycemia that occur during or after exercise. Studies from our own laboratory and others have demonstrated that exercise and hypoglycemia could reciprocally blunt subsequent ANS responses to either stress (Fig. 21-6).59,60 Thus, exercise blunts ANS responses (by 30% to 50%) to subsequent hypoglycemia, and vice versa. This feed-forward vicious cycle of blunted ANS responses between exercise and hypoglycemia can occur after only a few hours and persists for at least 24 hours following either stress (Fig. 21-7).39 Appreciation that deficient counterregulatory responses are also involved in the pathogenesis of exercise-related hypoglycemia explains why this phenomenon can occur many hours after exercise and provides a platform for therapeutic intervention. In general, in an individual with type 1 DM who is experiencing exercise-related hypoglycemia, it is appropriate to consider reducing both basal and mealtime insulin doses. Additionally, slightly raising glycemic targets and ensuring adequate carbohydrate repletion of glycogen stores are useful recommendations.

Gender can also play a large role in modulating ANS and neuroendocrine responses during hypoglycemia.58 In premenopausal, nondiabetic, and type 1 DM women, moderate hypoglycemia of ≈50 mg/dL (2.9 mmol/L) produces 30% to 50% reduced ANS responses compared with that seen in age- and body mass index (BMI)-matched men.58 However, when postmenopausal women, not on estrogen replacement, were compared with postmenopausal women receiving estrogen and age- and BMI-matched men, it was found that the large sexual dimorphism in ANS (epinephrine, muscle sympathetic nerve activity) and neuroendocrine (glucagon, growth hormone) responses was no longer present in estrogen-deficient women. It therefore would appear that in healthy humans, estrogen is a major mechanism responsible for sexual dimorphic counterregulatory responses during hypoglycemia.61 It should also be mentioned that women appear to be more resistant than men to the blunting effects of prior hypoglycemia. Thus, antecedent hypoglycemia can have up to a twofold greater suppressive effect on subsequent counterregulatory responses in men than in women.62 The mechanism for this intriguing finding is as yet unknown. Many studies have focused on the blunting effects of antecedent hypoglycemia on neuroendocrine and symptom responses during subsequent hypoglycemia. However, an additional component in the spectrum of deficient counterregulatory responses deserves mention. Several laboratories have identified that adrenergic receptors and particularly epinephrine action are downregulated by intensive glucose control and prior hypoglycemia.63,64 These reduced metabolic (lipolytic, endogenous glucose production, glycogenolysis) and cardiovascular responses contribute to defective counterregulatory defenses against a falling plasma glucose in patients with diabetes. Thus, it should be noted that strategies aimed at increasing epinephrine levels during hypoglycemia will be only partially successful if tissue resistance or adrenoreceptor downregulation to the action of the hormone is present. What is clear from the above wealth of data is that multiple mechanisms can downregulate ANS responses to hypoglycemia, and that this complex model presents numerous targets for therapeutic interventions to stimulate and restore counterregulatory responses during hypoglycemia in patients with diabetes.

Strategies to Improve Counterregulatory Responses During Hypoglycemia

Parallel with studies investigating the mechanisms responsible for HAAF, a number of laboratories have been exploring strategies for improving ANS and neuroendocrine responses during hypoglycemia. These have included preclinical approaches in rodents through to interventions in humans with type 1 DM. As mentioned above, hypothalamic kinases can act as important fuel sensors. Recent work in rats has shown that methods to reduce AMP-activated protein kinase (AMPK) in the ventral medial nucleus of the hypothalamus reduce epinephrine and glucagon responses during hypoglycemia, whereas increases in AMP kinase action can increase responses of these key neuroendocrine hormones during hypoglycemia.65 Other approaches to increase neuroendocrine responses during hypoglycemia in man have included amino acid infusions (glucagon) and use of caffeine (epinephrine and symptom responses). Terbutaline before bed has been demonstrated recently to increase plasma glucose levels during the night and to prevent nocturnal hypoglycemia.66 Recently, peroxisome-proliferator–activated receptor-γ (PPAR-γ) agonists, which are known to activate AMP kinase and fructose, which under certain conditions can inhibit glucokinase,67 have been demonstrated to increase counterregulatory responses in both healthy men and those with type 1 DM. Although at first glance these latter studies may seem unrelated, both may have mechanisms working through hypothalamic fuel sensing. Most recently, studies in conscious rats and healthy and type 1 DM men have highlighted the possible role of serotonergic transmission in modulating counterregulatory responses during hypoglycemia. Prolonged (i.e., weeks) use of two different selective serotonin reuptake inhibitors (sertraline and fluoxetine) has led to dramatic (30% to 60%) increases in ANS (epinephrine) responses during hypoglycemia.68

Drug-Induced Hypoglycemia

By far the most common cause of drug-induced hypoglycemia is insulin followed by sulfonylurea, meglitinides, and benzoic acid derivatives (i.e., oral insulin-producing agents). Hypoglycemia induced by oral insulin secretagogues is much less frequent than that caused by insulin but in certain instances can still be common. For example, hypoglycemia can occur in up to ≈35% of patients receiving glyburide or repaglinide. Hypoglycemia rates are less in agents with glucose-dependent insulin secretion (i.e., stimulation of insulin release is reduced during periods of hypoglycemia). Thus, the percentage of patients who experience hypoglycemia when receiving glimepiride, glipizide XL, or nateglinide is lower and is in the range of 5% to 10%. However, combination of even glucose-dependent insulin secretagogues with agents such as insulin or glucagon-like peptide-1 agonists can result in a much higher frequency of hypoglycemia (≥35%). Generally, newer sulfonylureas produce less hypoglycemia than do older first-generation sulfonylureas such as chlorpropamide. Severe hypoglycemia remains relatively uncommon with oral insulin secretagogues at a rate of ≈1.5 per 100 patient-years.33 Alcohol might be a more common cause of severe hypoglycemia in the United States than sulfonylureas.69 Alcohol can cause hypoglycemia in overnight fasted normal volunteers,70,71 with plasma glucose values as low as 5 mg/dL (0.3 mmol/L)72 and mortality rates ranging from 10% in adults to 25% in children.119 In a series of deaths caused by hypoglycemia, alcohol was the most common causative agent.74 The most common situation is a glycogen-depleted state, such as occurs in an individual who drinks after a considerable fast, or who drinks and then fasts. In the latter situation, blood alcohol levels can be low or undetectable.

Alcohol induces hypoglycemia by inhibiting gluconeogenesis72; as little as 50 g might be sufficient.73,76 Its mechanism of action is complex, with evidence of impaired counterregulatory hormone responses71 and impaired uptake of gluconeogenic precursors,77 but the predominantly accepted mechanism is its inhibition of the gluconeogenic process stemming from an increased reduced nicotinamide adenine dinucleotide (NADH)/NAD ratio as a result of the oxidation of alcohol to acetaldehyde and acetate, thus reducing the ability of the liver and kidney to oxidize lactate and glutamate to pyruvate and α-ketoglutarate, respectively.7880 Although plasma insulin levels are appropriately suppressed in this condition, because of this inhibition of gluconeogenesis, glucagon and catecholamines are ineffective in increasing glucose release and raising plasma glucose levels.81 Thus, in a patient with suspected alcohol-induced hypoglycemia, oral or intravenous glucose is the treatment of choice.

Only about 10% of reported cases of drug-induced hypoglycemia have occurred without concomitant insulin, sulfonylurea, or alcohol.82 Of these, propranolol,83 sulfonamides,84 and salicylates169 have been reported most frequently. Propranolol and other nonselective β-blockers decrease the ability of the liver and kidney to increase their release of glucose,85,86 enhance peripheral insulin sensitivity,87 and mask symptoms of impending hypoglycemia. The adverse effects of β-adrenergic β-blockers are mediated through β2-receptors. Recent studies indicate that β1-selective blockers do not present an increased risk for severe hypoglycemia and therefore should not be considered as being contraindicated in diabetic patients.87,88

Salicylates can act by inhibiting hepatic glucose release and increasing insulin secretion, although their exact mechanism remains to be determined. Sulfonamides probably act by stimulating insulin release in a manner similar to that of sulfonylureas. Angiotensin-converting enzyme inhibitors89and pentamidine90 are associated more frequently with hypoglycemia, as their use increases in diabetic patients and those with AIDS, respectively. Angiotensin-converting enzyme inhibitors can increase tissue insulin sensitivity91 and can decrease the degradation of bradykinin, which has certain insulin-mimetic actions.92 Pentamidine is cytotoxic to pancreatic β cells, and hypoglycemia occurs with the release of insulin from degenerating cells, often with subsequent permanent diabetes mellitus.93 Many of the drugs listed in Table 21-2 have been reported to cause hypoglycemia only in association with the use of antidiabetic medications or have been the subject of isolated case reports, and their etiologic significance remains to be established. However, their use in a patient with otherwise unexplained hypoglycemia should be discontinued whenever possible.

Treatment and Strategies to Reduce Hypoglycemia

Clinical Strategies to Reduce Hypoglycemia

Over the last generation, interest has increased in replacing insulin in the most physiologic manner for patients with diabetes. In the 1980s, the introduction of recombinant human insulin reduced the formation of antibodies and provided more predictable pharmacokinetic profiles. The next decade produced analogue insulins that initially were designed to provide a quicker onset and a shorter duration of action. These insulins (lispro, aspart, glulisine) were designed to reproduce more closely the typical physiologic prandial spikes of insulin observed following meals. The second wave produced long-acting basal types of insulin (glargine, detemir) designed to mimic the background constitutive release of the hormone that regulates nocturnal and interprandial glycemia. Studies in type 1 DM have demonstrated that hypoglycemia (particularly nocturnal) can be reduced when short-acting analogues versus regular (soluble) insulin are used.94,96 Similarly, long-acting analogues have been demonstrated to reduce hypoglycemia by 20% to 33% in patients with type 2 DM when compared with NPH-based regimens.34 Thus, current recommendations are to use analogue-based insulin replacement regimens whenever possible.

Insulin pump development began in the 1970s and over the past 20 years has become a major method of insulin replacement in the United States. Although widely acknowledged for its usefulness, the cost of treatment (pump and supplies) has restricted a more widespread acceptance of this technology. Nevertheless, studies in children94 and pregnant women have demonstrated a reduction in hypoglycemia when compared with multiple daily insulin injection regimens.

Most recently, continuous, real-time glucose monitoring has been introduced into clinical practice. Currently, a variety of devices can be worn in combination with an insulin pump or independently. In a multicenter, randomized, controlled trial, the use of a continuous glucose monitor has been shown to reduce both glycosylated hemoglobin (HbA1c) and, as a secondary end point, the incidence of severe hypoglycemia in type 1 DM adults and children but not adolescents95 when compared with conventional self-blood glucose monitoring. Currently, another large randomized study is under way that will test the hypothesis that a continuous glucose sensor–driven insulin pump replacement approach will provide better HbA1c and less hypoglycemia than multiple daily injections of insulin and traditional self-blood glucose monitoring in both children and adults with type 1 DM.

Pancreas Transplantation and Hypoglycemia

Pancreas transplantation has been performed in patients with type 1 DM for over 25 years. Typically, patients have experienced episodes of severe hypoglycemia, and this is one of the major criteria for consideration of pancreas transplantation. Generally, rates of hypoglycemia improve dramatically in the first year after transplantation. However, ≈30% of patients report episodes of minor hypoglycemia. Most (but not all) studies also demonstrate that counterregulatory defenses are improved after pancreatic transplantation.9799 Most notably, glucagon responses to hypoglycemia increase, accompanied at an early stage by some improvement in epinephrine and symptomatic responses. Improvements in glucagon responses appear to be persistent and have been documented in long-standing pancreas transplantation recipients of up to 19 years post transplant.100 Although the improvement in glucagon responses to hypoglycemia appears to be durable after long-term pancreatic transplantation, it is clear that improvements in epinephrine and symptomatic responses to hypoglycemia are not sustained.100 Along similar lines, it has been reported that physiologic insulin and glucagon responses during exercise are maintained for the most part following pancreas transplantation.101 Robertson and colleagues have investigated whether living donors of pancreas segments have normal counterregulatory responses to hypoglycemia.102 This has particular relevance, as ≈25% of donors can develop diabetes within 1 year after the procedure. It is interesting to note that despite deficient insulin and glucagon responses to glucose or arginine infusion, it was found that glucagon responses during hypoglycemia were preserved among donors.

A history of recurrent severe hypoglycemic events is also a major indication for islet cell transplantation. Although the number of transplanted individuals who are insulin independent is low (≈10%) up to 5 years after transplantation,103 it generally is reported that major episodes of hypoglycemia are significantly improved following islet cell transplantation.104 This finding appears to be due to the fact that there is some restoration of the ability to modulate insulin levels during hypoglycemia, as no reports have described significant increases in glucagons during hypoglycemia following islet cell transplantation.105 These findings underscore the importance of modulating insulin in the defense against hypoglycemia, but they also pose the intriguing question as to why there is an absent glucagon response. One plausible explanation is that the intrahepatic site for the islet transplant is the cause of the deficient glucagon response to hypoglycemia.106 In a recent series of elegant experiments in rats, Zhou et al.106 demonstrated that in the normal postprandial state, glucagon responses to hypoglycemia were absent in intrahepatially transplanted islets. However, when the animals were fasted for 48 hours and thus intrahepatic glucose flux was reduced, normal glucagon responses to hypoglycemia were obtained in the islet-transplanted animals.106 Supporting the theory that placing islets in the liver may be cause for the absent glucagon response during hypoglycemia are data indicating that alternate sites for islet cell transplantation away from the liver may produce improved glucagon responses to hypoglycemia.107

Hypoglycemia and Gastric Bypass Surgery

Accompanying the rapid increase in obesity is an increase in the number of bariatric surgical procedures performed. Recently, accumulating reports are demonstrating an increase in severe postprandial hypoglycemia following gastric bypass surgery.108110 Although the disorder was initially thought to represent a version of the adult nesidioblastosis syndrome, subsequent analysis of the pancreata have revealed potentially alternative pathologic causes.111 It is interesting to that that there is a strong female preponderance, and this condition can develop months or even years after bypass surgery is performed. Treatment for hypoglycemia has proved to be problematic. Neither acarbose given to try to alleviate dumping syndrome symptoms by slowing glucose absorption nor somatostatin given to inhibit endogenous insulin secretion has been successful.112 Consequently, pancreatic resection was needed to reduce the occurrence of hypoglycemia in the six patients originally reported with this syndrome.108 Subsequent reports have highlighted that patients with this syndrome have exaggerated glucagon-like peptide-1 (GLP-1) and insulin responses to a mixed meal.113 In late 2007, Vella and Service provided an update on the Mayo Clinic’s experience regarding hypoglycemia following gastric bypass surgery.112 They reported that 43 patients required an ≈60% gradient-guided pancreatic surgical resection to alleviate hypoglycemia. Pathologic inspection revealed that most of the pancreata had islet hypertrophy with nesidioblastosis, although some also had one or more insulinomas. It appears that this syndrome of hypoglycemia is specifically associated with bypass surgery, as a recent report of gastric banding to induce weight loss did not report any significant hypoglycemia over a 2-year postprocedure follow-up period.114

Neonatal Hypoglycemia

Plasma glucose levels below 70 mg/dL (3.9 mmol/L) commonly occur in newborn infants. It has been estimated that up to 50% of neonates can have plasma glucose levels below 50 mg/dL (2.9 mmol/L) following short-term fasting (up to 8 hours) after birth.115 Over the first 72 hours of life, gluconeogenic pathways mature, and the risk of hypoglycemia in normal infants is removed. However, in addition to the above physiologic causes of neonatal hypoglycemia, numerous pathologic conditions can cause hypoglycemia (Table 21-3).

The most common causes of persistent neonatal hypoglycemia are the congenital hyperinsulinism syndromes. These syndromes have been estimated to occur at a rate of about 1:30,000 births worldwide. However, in certain areas (e.g., Finland, the Arabian Peninsula), these conditions have been reported to occur at much higher rates (≈1:2,100).116,117 Over the past 50 years, congenital hyperinsulinism syndromes have been given a number of titles (e.g., nesidioblastosis, islet dysregulation syndrome, persistent hyperinsulinemic hypoglycemia of infancy). Recently, it has been discovered that nesiodioblastosis can be a normal feature of the pancreas in the neonatal period and is histologically different from the pathologic processes that cause congenital hyperinsulinism. The pathophysiology and the molecular and genetic bases for congenital hyperinsulinism were reviewed recently by De Leon and Stanley.115 Of these, the most frequently occurring involve loss-of-function mutations in the pancreatic β-cell K-ATP channel, Kir6.2 and Sur-1 receptors.

The clinical presentation of infants with congenital hyperinsulinism may be varied but typically involves prolonged and severe hypoglycemia with lethargy, seizures, and apnea, and babies are often large for gestational dates. Typically, hypoglycemia occurs in the fasting or postprandial state, and these infants require large rates of glucose infusion >10 mg/kg/min to prevent severe hypoglycemia. In the glutamate dehydrogenase version of congenital hyperinsulinism, hypoglycemia can occur during both the fasting and the postprandial state. Hypoglycemia in this condition can be precipitated by a protein meal and is characteristically accompanied by elevated ammonia levels. Diagnosis of glutamate dehydrogenase hyperinsulinism typically occurs when the child is several months of age.118

Diagnostic Investigation of an Infant With Persistent Hypoglycemia

Because of the wide range of conditions that can cause hypoglycemia in an infant, a variety of blood and urine tests can be helpful in identifying metabolic and endocrine causes of the disorder (Table 21-4).

It is surprising that plasma insulin levels often are not strikingly elevated in the congenital hyperinsulinism syndromes. Similar to insulinomas, plasma insulin levels often are modestly increased but are inappropriately raised in relation to the prevailing hypoglycemia. However, measurement of free fatty acids and β-hydroxybutyrate levels reveals dramatic suppression and evidence of significant insulin action. An increase in plasma glucose greater than 30 mg/dL following glucagon is also supportive of the diagnosis. Infants with the GDH-M1 form of congenital hyperinsulinism exhibit increased responses to leucine118 and thus can be distinguished from those with the KATP channel forms of hyperinsulinism. Additionally, genetic testing is now widely available for identification of four of the five genes associated with congenital hyperinsulinism. Initially, treatment consists of diazoxide, which is a KATP channel agonist and has a suppressive affect on insulin secretion. If diazoxide does not produce a clinical response, octreotide has been used to suppress insulin release. However, in children who are unresponsive to medical therapy, the next treatment is surgery. This can comprise a local resection or a near total (98%) pancreatectomy, depending upon the presence of focal or diffuse disease.115,118 The application of 18F-DOPA positron emission tomography (PET) scans to differentiate focal from diffuse pancreatic involvement in the congenital hyperinsulinism syndromes has provided a clinical breakthrough in the management of these children.119

Mitochondrial Fatty Acid Oxidation Disorders

Eleven different mitochondrial fatty acid oxidation disorders120 can present with fasting (hypoketotic) or postprandial hypoglycemia. Hepatic oxidation of fatty acids produces energy and acetyl-CoA, both of which are essential for gluconeogenesis. Additionally, fatty acid oxidation is required for hepatic ketone production. However, because of specific defects in mitochondrial fatty acid oxidation, gluconeogenesis is reduced, and this results in hypoglycemia and a lack of ketone bodies. The combination creates a scenario whereby the brain becomes starved of glucose and its primary alternative substrate (ketones), which creates conditions for severe neurologic consequences. Plasma membrane carnitine transporter deficiency is an autosomal recessive disorder that occurs in ≈1:40,000 infants. Low plasma carnitine and acyl carnitine are indicative of this disorder, and treatment with dietary carnitine prevents hypoglycemia. Carnitine palmitoyl transferase I and II (CPT I and CPT II) deficiencies can present as severe, life-threatening events that require frequent feeding and supplementation with dietary medium-chain triglycerides and l-carnitine (CPT II deficiency). Unfortunately, the most severe form of CPT II deficiency can present soon after birth and can be fatal. Medium-chain acetyl-CoA dehydrogenase (MCAD) deficiency is the most common fatty acid oxidation defect and is inherited as an autosomal recessive trait. Hypoglycemia usually occurs during fasting and conditions of stress such as viral illness.121,122 The presentation can be serious with vomiting, apnea, coma, encephalopathy, and death. It has been estimated that if left undiagnosed, up to 25% of children may die during the presenting event. Blood tests demonstrate elevated medium-chain acylcarnitine. As with other mitochondrial fatty acid oxidation disorders, the enzyme deficiency can be demonstrated diagnostically in fibroblasts. Treatment consists of avoidance of fasting and precipitation of attacks that could lead to neurologic defects and even death. Thus, frequent meals, bedtime snacks, uncooked cornstarch (to avoid nocturnal hypoglycemia), and carnitine supplementation have been used successfully in these children.

Very long chain (VLCAD) and short-chain acyl-CoA dehydrogenase (SCAD) deficiencies can present with serious, sometimes fatal, episodes of hypoglycemia and metabolic acidosis. These defects are also associated with cardiomyopathy (VLCAD), hepatomegaly, and neurologic defects (SCAD). Treatment involves avoidance of fasting and maintenance of glucose levels during stress, with frequent meals enriched with carbohydrate and medium-chain triglycerides. Carnitine supplementation in these conditions is not recommended (VLCAD) or has limited applicability (SCAD).

Long-chain 3-hydroxyacyl-CoA dehydrogenase (LCHAD) and mitochondrial trifunctional protein (MTP) deficiency are also autosomal recessive disorders. Both conditions can present with severe hypoglycemia, encephalopathy, neurologic complications, and death. Both conditions can result in peripheral neuropathy, and LCHAD deficiency can result in pigmentary retinopathy and blindness. Treatment is similar to that provided for the above fatty acid oxidation defects acyl-CoA dehydrogenase and carnitine/acylcarnitine translocase (CACT). Long-chain ketothilase (LCKAT) and 2-4-dienoyl-CoA reductase deficiency are very rare conditions that have been described recently.120

Sepsis, Trauma, and Burns

Buy Membership for Endocrinology, Diabetes and Metabolism Category to continue reading. Learn more here