Posteado por: delacruz1955 | noviembre 16, 2008

17. Diabetes mellitus, 10-Julio-2015, Profesores

Goldman: Cecil Textbook of Medicine, 22nd ed., Copyright © 2004 W. B. Saunders Company


Robert S. Sherwin

Diabetes mellitus is a chronic disorder characterized by the impaired metabolism of glucose and other energy-yielding fuels as well as by the late development of vascular and neuropathic complications. Diabetes comprises of a group of disorders involving distinct pathogenic mechanisms, for which hyperglycemia is the common denominator. Regardless of its cause, the disease is associated with a common hormonal defect, namely, insulin deficiency, which may be total, partial, or relative when viewed in the context of coexisting insulin resistance. Lack of insulin effect plays a primary role in the metabolic derangements linked to diabetes, and hyperglycemia in turn plays an important role in disease-related complications.
In 1998, the United States Centers for Disease Control and Prevention estimated that 16 million Americans (or nearly 6% of the U.S. population) fulfilled the diagnostic criteria for diabetes mellitus; more than one third of these cases were thought to be undiagnosed. The number of affected patients continues to rise as the 21st century begins, with current estimates exceeding 800,000 new cases per year. Diabetes is the fourth most common reason for patient contact with an American physician, accounting for approximately 12% of U.S. health care dollars and total annual costs exceeding 100 billion dollars. Worldwide, diabetes affects more than 135 million people; this figure is projected to reach 300 million cases by 2025. Unfortunately, the rate of growth of diabetes is largest in developing nations, where barriers exist to proper diagnosis and treatment.
Diabetes is a leading cause of both mortality and early disability; in the United States, it is the leading cause of blindness among working-age adults, of end-stage renal disease, and of nontraumatic limb amputations. Diabetes increases the risk of cardiac, cerebral, and peripheral vascular disease two- to seven-fold and in the obstetric setting is a major contributor to neonatal morbidity and mortality. On the bright side, a growing body of evidence suggests that most (if not all) of the debilitating complications of diabetes ( Fig. 242-1 ) can be prevented or delayed by the prospective treatment of hyperglycemia and other cardiovascular risk factors. When treating diabetes, the timing of


Figure 242-1 Long-term complications of diabetes mellitus. (From Forbes CD, Jackson WF: Color Atlas and Text of Clinical Medicine, 3rd ed. London, Mosby, 2003, with permission.)
therapy is crucial; clinical outcomes depend critically on early recognition and treatment of the disease.
The newly revised American Diabetes Association (ADA) classification scheme for diabetes mellitus is summarized in Table 242-1 . Clinical diabetes is divided into four general subclasses, including (1) type 1, primarily caused by autoimmune pancreatic ß-cell destruction and characterized by absolute insulin deficiency; (2) type 2, characterized by insulin resistance and relative insulin deficiency; (3) “other” specific types of diabetes, associated with identifiable clinical conditions or syndromes; and (4) gestational diabetes mellitus. In addition to these clinical categories, two “risk conditions”—impaired glucose tolerance (IGT) and impaired fasting glucose (IFG)—have been defined to describe metabolic states in between normal glucose homeostasis and overt diabetes. Both IGT and IFG significantly increase the future risk of developing diabetes mellitus, and in many cases are part of the disease’s natural history. It should also be noted here that patients with any form of diabetes may require insulin therapy; for this reason, the previously used terms “insulin-dependent” (for type 1) and “non-insulin-dependent” (for type 2) diabetes have been eliminated.
Patients with type 1 diabetes mellitus have little or no insulin secretory capacity and depend on exogenous insulin to prevent metabolic decompensation and death. Classically, symptoms appear abruptly (i.e., over days or weeks) in previously healthy, nonobese children or young adults; in older patients, however, the disease may manifest more gradually. At the time of initial evaluation, most type 1 diabetic patients are ill and symptomatic, most commonly presenting with polyuria, polydipsia, polyphagia, and weight loss; such patients may also present with ketoacidosis. Type 1 diabetes is believed to have a prolonged asymptomatic preclinical phase (often lasting years), during which pancreatic ß cells are gradually destroyed by an autoimmune attack influenced by HLA and other

I. Type 1 diabetes, formerly known as insulin-dependent diabetes mellitus (IDDM) or “juvenile-onset diabetes” (primarily due to ß-cell destruction, usually leading to absolute insulin deficiency)
A. Immune mediated
B. Idiopathic
II. Type 2 diabetes, formerly known as non-insulin-dependent diabetes (NIDDM) or “adult-onset diabetes” (may range from predominantly insulin resistance with relative insulin deficiency to predominantly secretory defect with insulin resistance)
III. Other specific types
A. Genetic defects of ß-cell function (e.g., maturity-onset diabetes of the young [MODY] types 1–6 and point mutations in mitochondrial DNA)
B. Genetic defects in insulin action (e.g., type A insulin resistance, leprechaunism, Rabson-Mendenhall syndrome, lipoatrophic diabetes)
C. Disease of the exocrine pancreas (e.g., pancreatitis, trauma, pancreatectomy, neoplasia, cystic fibrosis, hemochromatosis, fibrocalculous pancreatopathy)
D. Endocrinopathies (e.g., acromegaly, Cushing’s syndrome, hyperthyroidism, pheochromocytoma, glucagonoma, somatostinoma, aldosteronoma)
E. Drug- or chemical-induced (e.g., vacor, pentamidine, nicotinic acid, glucocorticoids, thyroid hormone, diazoxide, ß-adrenergic agonists, thiazides, phenytoin, a-interferon)
F. Infections (e.g., congenital rubella, cytomegalovirus)
G. Uncommon forms of immune-mediated diabetes (e.g., “stiff-man” syndrome, anti-insulin receptor antibodies)
H. Other genetic syndromes (e.g., Down sydrome, Klinefelter’s syndrome, Turner’s syndrome, Wolfram’s sydrome, Friedrich’s ataxia, Huntington’s disease, Laurence-Moon-Biedl syndrome, myotonic dystrophy, porphyria, Prader-Willi syndrome)
IV. Gestational diabetes mellitus
I. Impaired fasting glucose (IFG)
II. Impaired glucose tolerance (IGT)

genetic factors, as well as by the environment ( Fig. 242-2 ). In some patients, an acute illness may speed the transition from the preclinical phase to clinical disease. Initially, most type 1 patients require high-dose insulin therapy to restore a disordered metabolism. A so-called “honeymoon period” (lasting weeks or months) may follow, however, during which small doses of insulin are needed due to partial recovery of ß-cell function and reversal of the insulin resistance caused by acute illness. Thereafter, insulin secretory capacity is gradually lost; this process may take several years. That type 1 diabetes is an autoimmune disease is supported by its association with specific immune response (HLA) genes and by the presence of autoantibodies to islet cells and their constituents (e.g., insulin, glutamic acid decarboxylase). Type 1 diabetes accounts for less than 10% of cases of diabetes in the United States.
Type 2 accounts for over 90% of cases of clinical diabetes. Patients with type 2 disease retain some endogenous insulin secretory capacity; however, their insulin levels are low relative to their ambient glucose levels and magnitude of insulin resistance. Type 2 patients are not dependent on insulin for immediate survival, and ketosis rarely develops, except under conditions of great physical stress. Nevertheless, many of these patients do require insulin therapy for proper glycemic control. Although found with increasing frequency in adolescents, type 2 diabetes is usually associated with advancing age; most cases are diagnosed after age 45. Type 2 diabetes has a high rate of genetic penetrance unrelated to HLA genes and is associated with a high-fat diet, obesity, and/or a lack of physical activity. The clinical features of type 2 diabetes can be quite insidious; classic symptoms may be mild (fatigue, weakness, dizziness, blurred vision, and other nonspecific complaints may dominate the clinical picture) or may be tolerated for many years before a patient seeks medical attention. Moreover, if the degree of hyperglycemia is insufficient to produce symptoms, the diagnosis may be made only after the development of vascular or neuropathic complications.
This category encompasses a wide variety of diabetic syndromes attributed to a specific disease, drug,


Figure 242-2 A summary of the sequence of events that lead to pancreatic ß-cell loss, and ultimately to the clinical appearance of type 1 diabetes. DKA = diabetic ketoacidosis.
or condition (see Table 242-1 ). Categories include genetic defects of ß-cell function or insulin action, diseases of the exocrine pancreas, endocrinopathies, drug- or chemical-induced diabetes, infections, and other immune-mediated and genetic syndromes associated with diabetes mellitus.
Maturity-onset diabetes of the young (MODY), formerly classified as a subtype of type 2 diabetes, has now been more accurately described as a consequence of genetic research. Clinically, patients with MODY generally present in adolescence or young adulthood; unlike patients with classic type 2 diabetes, they are usually nonobese, normotensive, and normolipidemic at the time of diagnosis. MODY is a heterogeneous disorder encompassing several monogenic defects of ß-cell function, with autosomal dominant inheritance and penetrance exceeding 80%. Mutations at several genetic loci have been identified. The most common form—MODY type 3—is associated with a mutation of hepatocyte nuclear factor 1a (HNF-1a), a gene transcription factor encoded on chromosome 12. MODY type 2 patients share a mutation in the gene encoding glucokinase, the key enzyme responsible for the phosphorylation of glucose within the ß cell and the liver. A variety of glucokinase mutations have been identified in different families, each capable of interfering with the transduction of the glucose signal to the ß cell. Other described forms of MODY are shown in Table 242-2 ; the existence of additional forms of MODY

ß-Cell mass, insulin secretion
OHA, insulin
Glucose phosphorylation
Diet and exercise
ß-Cell mass, insulin secretion
OHA, insulin
IPF-1 (PDX-1)
ß-Cell development and function
OHA, insulin
ß-Cell mass, insulin secretion
Neuro D1 (BETA2)
ß-Cell development and function
HNF = hepatocyte nuclear factor; IPF = insulin promoter factor; Neuro D1 = neurogenic differentiation factor 1; OHA = oral hypoglycemic agent.
Adapted from Fajans SS, Bell GI, Polonsky KS: Molecular mechanisms and clinical pathophysiology of maturity-onset diabetes of the young. N Engl J Med 2001;345:971–980.

is suggested by the presence of patient clusters with similar clinical findings whose genetic basis for disease remains unknown.
Severe illness (e.g., burns, trauma, sepsis) can provoke stress hyperglycemia as a result of the hypersecretion of insulin antagonistic hormones (e.g., catecholamines, cortisol). Although this may represent the unmasking of underlying diabetes, the metabolic disturbance may be self-limited and should therefore not be formally classified as diabetes until the precipitating illness has resolved. It should also be emphasized that while most patients can be readily classified on clinical grounds, a small subgroup of patients are difficult to classify because they display features common to both type 1 and type 2 diabetes. Such patients are classically nonobese, with reduced insulin secretory capacity but little tendency for ketosis. Many of these “in-between” patients initially respond to oral agents; however, nearly all of them will eventually require insulin therapy. Many of these patients appear to have a slowly evolving form of type 1 diabetes; others defy easy categorization.
The term gestational diabetes mellitus (GDM) describes a condition in women with impaired glucose tolerance that appears or is first detected during pregnancy. Women with known diabetes prior to conception are not classified as having gestational diabetes. GDM usually appears in the second or third trimester, when pregnancy-associated insulin antagonistic hormones reach their peak. After delivery, glucose tolerance generally (but not always) reverts to normal. Within 5 to 10 years, however, type 2 diabetes develops in nearly one half of women with prior GDM; occasionally, pregnancy can precipitate type 1 diabetes as well. As a whole, GDM occurs in about 4% of U.S. pregnancies, producing approximately 135,000 cases per year; local prevalence rates may rise as high as 14% in high-risk populations. Although patients with GDM generally present with mild, asymptomatic hyperglycemia, rigorous treatment is indicated to protect against hyperglycemia-associated fetal morbidity and mortality. Insulin is often required.
Diagnosis and Screening
The diagnosis of diabetes mellitus is straightforward when classic symptoms of polyuria, polydipsia, and unexplained weight loss are present. In these cases, a random plasma glucose measurement of 200 mg/dL or greater is sufficient to clinch the diagnosis; confirmatory testing is unwarranted and may delay treatment. Although glycosuria is strongly suggestive of diabetes, urine test results should never be used exclusively to diagnose diabetes, since an altered renal threshold for glucose can produce similar findings. If suspected diabetes is not confirmed through random glucose determination, additional diagnostic testing should be performed.
An 8-hour (overnight) fasting plasma glucose measurement is most convenient; diabetes is established if fasting glucose levels are 126 mg/dL or greater on two separate occasions. Alternatively, a 75 g oral glucose tolerance test (oGTT) may be employed. The oGTT should be performed after an overnight fast, using a glucose load containing 75 g of anhydrous glucose dissolved in water; 2-hour postload glucose levels of 200 mg/dL or greater confirm the presence of diabetes. An important note about the oGTT: while able to detect diabetes in its earliest stage, this test should be performed under controlled conditions to ensure its accuracy. Common factors that nonspecifically

deteriorate the oGTT include (1) carbohydrate restriction (<150 g for 3 days), (2) bed rest or severe inactivity, (3) medical or surgical stress, (4) drugs (e.g., thiazides, ß-blockers, glucocorticoids, or phenytoin), (5) smoking, and (6) anxiety from repeated needlesticks. As a result, the oGTT should not be performed in acutely ill patients, and patients taking the oGTT should ideally stop smoking and consume a liberal carbohydrate diet for at least 3 days prior to testing. The current American Diabetes Association criteria for the diagnosis of diabetes mellitus are shown below; in the absence of unequivocal hyperglycemia with acute metabolic decompensation, each criterion used should be confirmed by repeat testing on a separate occasion.

Classic symptoms of diabetes (polyuria, polydipsia, and unexplained weight loss) PLUS random glucose concentration of 200 mg/dL or greater (=11.1 mmol/L) OR

Fasting (=8-hour) plasma glucose concentration of 126 mg/dL or greater (= 7.0 mmol/L) OR

2-hour postload glucose concentration of 200 mg/dL or greater (=11.1 mmol/L) during a 75 g oGTT
In recent years, increasing emphasis has been placed on two “risk categories” for diabetes, IFG and IGT. Since both conditions are associated with an increased risk of developing diabetes and subsequent vascular disease, all patients with IFG or IGT should be treated with diet and exercise and should be screened annually for progression to diabetes. The recent report of the NIH-funded Diabetes Prevention Program as well as studies from Finland and China have demonstrated that modest changes in life style sharply reduced the development of type 2 diabetes in patients with IGT. As detailed earlier, diabetes mellitus is established if fasting glucose levels are 126 mg/dL or greater; however, a fasting glucose concentration of 109 mg/dL, not 125 mg/dL, has been designated as the upper limit of normal. While somewhat arbitrary, this level was chosen because it approximates the level above which acute-phase insulin secretion is suppressed in response to intravenous glucose. More importantly, fasting glucose levels above 109 mg/dL are associated with an increased risk of developing diabetes. Patients with fasting glucose levels between 110 and 125 mg/dL are classified as having IFG ( Table 242-3 ). Because individuals with IFG may exhibit severe postprandial hyperglycemia, a 75 g oGTT should be performed in all such patients to rule out diabetes. During the 75 g oGTT, 2-hour postload glucose concentrations of 200 mg/dL or greater are diagnostic of diabetes, whereas patients with levels between 140 and 199 mg/dL are defined as having IGT. Table 242-3 summarizes the diagnosis of IFG, IGT, and overt diabetes mellitus.
Because patients with diabetes may harbor the disease for many years before symptoms are appreciated, the ADA has endorsed the screening of “high-risk” individuals at 3-year intervals ( Table 242-4 ). By current ADA criteria, “high-risk” patients include those with a personal history of IFG, IGT, GDM, obesity, hypertension, or dyslipidemia. Patients in high-risk ethnic groups and patients with first-degree relatives with diabetes also qualify for screening. In most cases, a fasting plasma glucose level is the screening test of choice; however, the oGTT has the distinct advantage of detecting patients with IGT.



<140 mg/dL
140–199 mg/dL
>22 mg/dL
<110 mg/dL
110–125 mg/dL
=126 mg/dL
DM = diabetes mellitus; IFG = impaired fasting glucose; IGT = impaired glucose tolerance; oGTT = oral glucose tolerance test.

*These diagnostic categories are based on the combined results of a fasting plasma glucose level and a 2-hour, 75-g oral glucose tolerance test. Note that a confirmed random plasma glucose level of =200 mg/dL in the appropriate clinical setting is diagnostic of diabetes and precludes the need for further testing.

1. Testing for diabetes should be considered in all individuals at age 45 years and older and, if results are normal, it should be repeated at 3-year intervals.
2. Testing should be considered at a younger age or be carried out more frequently in individuals who
• Are obese (>120% desirable body weight or a body mass index >27)
• Have a first-degree relative with diabetes
• Are members of a high-risk ethnic population (e.g., African-American, Hispanic American, Native American, Asian American, Pacific Islander)
• Have delivered a baby weighing >9 pounds or have been diagnosed with gestational diabetes mellitus
• Have systemic hypertension (blood pressure >140/90)
• Have a high-density lipoprotein cholesterol level <35 mg/dL and/or a triglyceride level >250 mg/dL
• On previous testing, had impaired glucose tolerance or impaired fasting glucose
Adapted from Report of the expert committee on the diagnosis and classification of diabetes mellitus. Diabetes Care 2000;23(Suppl 1):S4–S19.

*A fasting plasma glucose (FPG) or an oral glucose tolerance test (OGTT) can be used for diagnosis. In most clinical settings, the FPG is preferred because of ease of administration, convenience, acceptability to patients, and lower cost.

Since even mild glucose elevations can have serious adverse effects on a developing fetus, an aggressive screening approach is recommended during pregnancy. Women with a high clinical risk of gestational diabetes (personal history of GDM, obesity, glycosuria, or a strong family history of diabetes) should undergo screening as soon as possible after conception; in these patients, screening prior to pregnancy is preferred if possible. At 24 to 28 weeks of gestation, screening is recommended for all pregnant women, except those in the lowest risk category who meet all of the following clinical characteristics:
• Age less than 25 years
• Weight normal before pregnancy
• Member of an ethnic group with a low risk of GDM (e.g., European)
• No known diabetes in first-degree relatives
• No history of abnormal glucose tolerance
• No history of poor obstetric outcome
In pregnant women, a casual plasma glucose level of 200 mg/dL or greater or a confirmed fasting plasma glucose level of 126 mg/dL or greater establishes the diagnosis of GDM and precludes the need for a glucose challenge. In the absence of obvious hyperglycemia, a screening 1-hour 50 g oGTT should be performed between 24 and 28 weeks of gestation. If the fasting glucose level is 105 mg/dL or greater or the 1-hour postload value is 140 mg/dL or greater, a diagnostic 100 g oGTT is indicated. Gestational diabetes is then diagnosed if two or more values equal or exceed the upper limits of normal: fasting, 95 mg/dL; 1-hour, 180 mg/dL; 2-hour, 155 mg/dL; and 3-hour, 140 mg/dL. To save time and effort, proceeding directly to the 100 g diagnostic oGTT is an acceptable alternative.
Prevalence rates for type 1 diabetes are relatively accurate, since these patients invariably become symptomatic; current estimates for the United States hover between 0.3 and 0.4%. Type 1 diabetes is more prevalent in Finland, Scandinavia, and Scotland, less prevalent in Southern Europe and the Middle East, and uncommon in Asian nations. The annual incidence appears to have risen in the last half-century, which could imply the introduction of an unidentified environmental factor. Prevalence rates are strikingly different among ethnic groups living in the same geographic region, likely due to genetic differences in susceptibility to the disease.
Recent recognition that type 1 diabetes has a protracted preclinical phase has shed new light on some epidemiologic characteristics of the disease. Type 1 diabetes has an increased incidence in the winter months and may be associated with specific viral epidemics. These

observations may in part be explained by the superimposition of illness-provoked insulin resistance in patients with marginal ß-cell function. Similarly, the common appearance of type 1 diabetes during puberty may also be attributed to insulin resistance; even under normal circumstances, puberty is accompanied by impaired insulin-stimulated glucose metabolism. New methods for tracking islet-directed autoimmunity have led to a reappraisal of the age at which type 1 diabetes first appears. Although the age-specific incidence rises progressively from infancy to puberty and then declines, incidence rates persist at lower levels for many decades; in fact, about 30% of patients are diagnosed after the age of 20 years. In the later-onset patients, the clinical syndrome tends to evolve more slowly; in addition, islet-directed antibody titers may be lower, and HLA types may be different from those of younger patients. As a result, type 2 diabetes mellitus is initially misdiagnosed in many of these patients.
Systematic screening for asymptomatic diabetes mellitus is generally limited to high-risk populations, rendering broader prevalence estimates imprecise. Total U.S. prevalence has been estimated at 6% but likely exceeds 10 or 15% in persons older than 50 years of age; one third of these cases are thought to be undiagnosed. Type 2 diabetes is more common in Native Americans, Hispanic Americans, and African Americans than in people of European heritage; these patients also typically present at an earlier age. Prevalence rates also vary worldwide, where type 2 diabetes has a propensity for Asiatic Indians, Polynesians/Micronesians, and Latin Americans. Interestingly, African blacks, Australian Aborigines, Asians, and Pacific Islanders all have an increased risk of diabetes after emigration to the United States; this may be attributable to a genetically determined inability to metabolically adapt to “Western” behavior patterns, such as reduced physical activity and a high-fat, high-calorie diet.
Although relatively little is known about the specific genetic abnormalities associated with type 2 diabetes, the personal factors promoting disease expression are well established. Increased age, reduced physical activity, and especially obesity promote the expression of disease in genetically susceptible persons. The severity and duration of obesity contribute significantly to diabetes risk; patients with high waist-hip ratios (i.e., central or upper body obesity) are also more prone to the disease. Family history is also very important, since type 2 diabetes occurs more frequently in persons with diabetic parents or siblings. Identical twin concordance rates approach 100%; in these cases, affected twins will even develop diabetes at a similar age.
Precise statistical data regarding the prevalence of these diagnostic categories are lacking. In the United States, it is estimated that about 10–12 million people have impaired fasting glucose levels, while about 20 million have impaired glucose tolerance. The diagnoses often overlap as well: approximately 37% of patients with IFG also have IGT, and approximately 24% of patients with IGT also have IFG. Owing to the insidious nature of both conditions, precise rates of progression to overt diabetes are difficult to establish; current estimates approach 5 to 8% per year for each condition, with even higher rates if both conditions are present. In general, IGT and IFG have similar capacity to predict the future development of diabetes. IGT is also an independent risk factor for cardiovascular complications.
The gene coding for human insulin is located on the short arm of chromosome 11. Insulin is initially synthesized in pancreatic ß cells as proinsulin, a single-chain, 86 amino acid polypeptide. Subsequent cleavage of proinsulin removes a connecting strand (C-peptide) to form the smaller, double-chain insulin molecule, which contains 51 amino acid residues. Both insulin and the C-peptide remnant are packaged in membrane-bound storage granules; stimulation of insulin secretion results in the discharge of equimolar amounts of insulin and C-peptide (and a small amount of proinsulin) into the portal circulation. Although insulin is heavily metabolized during its first pass through the liver, the C-peptide fragment largely escapes hepatic metabolism; as a result, peripheral C-peptide levels provide a more precise marker of endogenous insulin secretion.
Glucose concentration is the key regulator of insulin secretion. To activate secretion, a glucose molecule must first be transported by a protein (GLUT 2) into the ß cell, phosphorylated by the enzyme glucokinase, and metabolized. The precise triggering process is poorly understood but probably involves activation of signal transduction pathways and mitochondrial signals, closure of adenosine triphosphate-sensitive potassium channels, and calcium entry into the cytoplasm of the ß cell. Normally, when blood glucose rises even slightly above fasting levels, ß cells secrete insulin, initially from preformed (stored) insulin and later from de novo insulin synthesis. The magnitude of the insulin response is determined by the amount of glucose available as well as by the mode of glucose entry; compared with intravenous administration, higher insulin levels are produced when glucose is given orally because of the simultaneous release of gut peptides (e.g., glucagon-like peptide I, gastric inhibitory polypeptide), which amplify the insulin response. Other insulin secretagogues include amino acids (e.g., leucine), vagal stimulation, sulfonylureas, repaglinide, and nateglinide (see later). Once secreted into the portal vein, 50% or more of insulin is removed by first pass through the liver. The consequence of this hepatic metabolism is that portal vein insulin levels are at least two- to four-fold higher than levels in the peripheral circulation. This point has clinical relevance with regard to insulin therapy; whereas insulin secreted by pancreatic ß cells directly enters the portal circulation, peripherally administered insulin does not raise portal insulin levels and therefore may be less efficient in inducing hepatic effects.
Insulin acts on its target tissues (liver, muscle, and fat, primarily) through a specific insulin receptor, which is a heterodimer containing two a- and two ß-chains linked by disulfide bridges. The a-subunits of the receptor reside on the extracellular surface and are the sites of insulin binding. The ß-subunits span the membrane and can be phosphorylated on serine, threonine, and tyrosine residues on the cytoplasmic face. The intrinsic protein tyrosine kinase activity of the ß-subunit is essential for the function of the insulin receptor. Rapid receptor autophosphorylation and tyrosine phosphorylation of cellular substrates are important early steps in insulin action. Thereafter, a series of phosphorylation and dephosphorylation reactions are triggered that produce insulin’s ultimate effects. A variety of post-receptor signal transduction pathways are activated by insulin, including PI3 (phosphatidylinositol 3′) kinase, an enzyme whose product appears to be critical for the eventual translocation of glucose transport proteins (GLUT 4) to the cell surface to facilitate glucose uptake.
A number of so-called “counter-regulatory” hormones oppose the metabolic actions of insulin, including glucagon, growth hormone, cortisol, and catecholamines. Among these, glucagon (and to a lesser extent, growth hormone) plays the most important role in the development of diabetes. Glucagon is normally secreted by pancreatic a cells in response to hypoglycemia, amino acids, and activation of the autonomic nervous system. Its chief effects are on the liver, where it stimulates glycogenolysis, gluconeogenesis, and ketogenesis via cyclic adenosine monophosphate-dependent mechanisms. Glucagon release is normally inhibited by hyperglycemia and hyperinsulinemia; however, in both types of diabetes, glucagon levels are absolutely or relatively elevated despite the presence of hyperglycemia. Growth hormone secretion by the anterior pituitary gland is also inappropriately increased in type 1 diabetes, a result (at least in part) of the body’s attempt to overcome a defect in insulin-like growth factor type 1 generation caused by insulin deficiency. The major metabolic actions of growth hormone are on peripheral tissues, where it acts to promote lipolysis and inhibit glucose consumption. In type 1 diabetic patients with reduced portal insulin levels, growth hormone is also capable of stimulating hepatic glucose production.
Insulin deficiency—be it relative or absolute—plays a pivotal role in the pathophysiology of diabetes mellitus. The effects of insulin lack are best appreciated by first examining the normal role of insulin in fuel homeostasis.
Fasted State.
After an overnight fast, low basal insulin levels result in diminished glucose uptake in peripheral insulin-sensitive tissues (e.g., muscle and fat). In the fasted state, most glucose uptake occurs in non-insulin-sensitive tissues, primarily the brain, which, because of its inability to use free fatty acids, is critically dependent on a constant supply of glucose for oxidative metabolism. Maintenance of stable blood glucose levels is achieved through the release of glucose by the liver (and to a small extent, by the kidney); production rates of 7 to 10 g per hour (~2 mg/kg/min) match those of the consuming tissues. The hepatic processes involved are glycogenolysis and gluconeogenesis; both play a significant role, and both depend critically on the balance between insulin and glucagon in the portal circulation. Reduced portal insulin levels decrease glycogen synthesis, which allows glucagon’s stimulatory effect on glycogenolysis to prevail. Glucagon predominance also stimulates gluconeogenesis, while concurrent low insulin

levels promote the peripheral mobilization of glucose precursors (amino acids, lactate, pyruvate, glycerol) and fuels (free fatty acids) for gluconeogenesis.
Fed State.
Ingestion of a large glucose load triggers multiple homeostatic mechanisms to minimize glucose excursions, including (1) suppression of endogenous glucose production, (2) stimulation of hepatic glucose uptake, and (3) acceleration of glucose uptake by peripheral tissues, predominantly muscle. Each of these mechanisms depends principally on insulin. In the liver, meal-stimulated insulin levels rapidly suppress glucose production. At least 30% of ingested glucose is deposited directly in the liver, via glycogen synthesis and storage; concurrently, hepatic triglyceride synthesis increases. Peripherally, insulin-stimulated glucose transport across the plasmalemma of both adipose and muscle tissue is attributable to the recruitment of glucose transport proteins (i.e., GLUT 4) from the cytosolic compartment to the plasma membrane. In muscle, glucose may then be metabolized, or it may be converted to glycogen for storage. In adipose tissue, glucose is used primarily for the formation of a-glycerophosphate, which is necessary for the esterification of free fatty acids to form triglycerides for storage in adipose tissue.
The scenario described—the ingestion of large quantities of pure glucose—is not representative of conditions during ordinary meals. If the quantity of carbohydrate consumed and resulting insulin response are small, glucose homeostasis is maintained largely by reduced hepatic glucose production rather than by increased glucose uptake, because glucose production is much more sensitive than glucose uptake to the effects of small changes in insulin secretion. The rise in insulin that accompanies the consumption of mixed meals also facilitates protein and fat storage. Because muscle is in negative nitrogen balance in the fasting state, repletion of muscle nitrogen depends on the net uptake of amino acids in response to protein feeding. In muscle, insulin acts to promote positive nitrogen balance by facilitating amino acid uptake, by inhibiting the breakdown of protein, and (to a lesser extent) by stimulating new protein synthesis. In adipose tissue, the action of insulin accelerates triglyceride incorporation by stimulating lipoprotein lipase, while simultaneously inhibiting the hormone-sensitive lipase that catalyzes the hydrolysis of stored triglycerides. In adipose tissue, the net effect of insulin is to promote the synthesis and storage of triglycerides.
In both type 1 and type 2 diabetes, fasting hyperglycemia is accompanied by an inappropriate increase in hepatic glucose production; this effect is magnified in type 1 diabetes due to absolute portal insulin deficiency. In addition, total body glucose uptake is generally increased in diabetes, largely due to mass action induced by hyperglycemia. Increased hepatic glucose production in both types of diabetes is due mostly to accelerated gluconeogenesis; the loss of insulin’s restraining effect on the a cell also leads to a relative increase in portal glucagon levels, resulting in increased uptake and conversion of glycogenic substrates to glucose within the liver. Insulin deficiency also leads to the hypersecretion of growth hormone, which further accentuates glucose overproduction. In the extreme situation of total insulin lack, excessive counter-regulatory hormone release further stimulates gluconeogenesis, while blocking compensatory increases in glucose disposal. The clinical correlate is profound hyperglycemia and glycosuria ( Fig. 242-3 ).
Diabetes is also characterized by marked postprandial hyperglycemia. In type 2 diabetes, delayed insulin secretion and hepatic insulin resistance join forces to impair both suppression of hepatic glucose production and the liver’s ability to store glucose as glycogen. Hyperglycemia ensues, even though insulin levels may eventually rise to levels above those seen in nondiabetic individuals (insulin secretion remains deficient relative to the prevailing glucose level), because insulin resistance reduces the capacity of myocytes to extract and store the excess glucose released from the liver. Under normal circumstances, muscles show increased levels of glucose-6-phosphate after sensing insulin; this rise is markedly attenuated in diabetes, which implies that the block in glycogen synthesis precedes glucose-6-phosphate formation, and thus is mediated at the level of either glucose transport (by GLUT 4) or the conversion of glucose to glucose-6-phosphate (by hexokinase). These defects are more pronounced in patients with severe hyperglycemia, in whom insulin secretion is further reduced. Type 1 patients show the most marked and prolonged elevations in blood glucose after ingestion of carbohydrate. These individuals have low portal vein insulin levels, which cannot be reversed by subcutaneous insulin therapy. Consequently, during hyperglycemia, the liver fails to arrest glucose production and fails to

Figure 242-3 The effects of severe insulin deficiency on body fuel metabolism. Lack of insulin leads to mobilization of substrates for gluconeogenesis and ketogenesis from muscle and adipose tissue, accelerated production of glucose and ketones by the liver, and impaired removal of endogenous and exogenous fuels by insulin-responsive tissues. The net results are severe hyperglycemia and hyperketonemia that overwhelm renal removal mechanisms. FFA = free fatty acids.
appropriately take up glucose for storage as glycogen. In addition, glucose uptake by peripheral tissues is impaired by the lack of insulin and by the development of insulin resistance secondary to chronic insulin deprivation and the toxic effects of chronic hyperglycemia. The net result is a gross defect in glucose disposal that can be only partially compensated by renal glycosuria.
In addition to hyperglycemia, fasting free fatty acid levels are also elevated in diabetes, because of accelerated mobilization of fat stores. In type 2 diabetes, elevated free fatty acid levels occur in the presence of normal or even increased insulin levels, suggesting that adipocytes become resistant to insulin’s inhibitory effect on lipolysis. This adipocyte resistance ultimately leads to the mobilization and inappropriate deposition of triglyceride into liver and muscle, which in turn is associated with insulin resistance in these organs.
Although free fatty acids are not directly converted to glucose, they do promote hyperglycemia by providing the liver with energy to support gluconeogenesis, as well as by interfering with muscle glucose uptake (predominantly by inhibiting glucose transport). Endogenous insulin secretion in type 2 diabetes provides sufficient portal levels of insulin to suppress the conversion of free fatty acids to ketones in the liver. In type 1 diabetes, however, mobilized free fatty acids are more readily converted to ketone bodies. The combined effects of insulin deficiency and the presence of glucagon suppress fat synthesis in the liver. This suppression of fat synthesis reduces intrahepatic malonyl coenzyme A, which together with carnitine stimulates the activity of hepatic acylcarnitine transferase I and thereby facilitates the transfer of long-chain fatty acids into mitochondria, where they are broken down via ß-oxidation and converted to ketone bodies. In addition, hypoinsulinemia, by decreasing ketone turnover, enhances the magnitude of the ketosis for any given level of ketone production. During diabetic ketoacidosis, ketone levels are further increased because of the concurrent release of counter-regulatory hormones. The rise in glucagon accelerates hepatic ketogenesis, whereas elevations of catecholamines, growth hormone, and cortisol act in concert to increase lipolysis and subsequent delivery of free fatty acids to the liver (see Fig. 242-3 ). The increase in substrate delivery may become so pronounced that it saturates the oxidative pathway, thus leading to hepatic steatosis and severe hypertriglyceridemia.
In addition to disordered glucose disposal, type 1 diabetic patients may exhibit defects in the disposal of ingested proteins and fats as well. In the absence of the normal rise in insulin, meal ingestion may produce hyperaminoacidemia, because of a failure to stimulate the

net uptake of amino acids in muscle, and hypertriglyceridemia, through the reduced activity of lipoprotein lipase. Thus, diabetes should be viewed not only as a disorder of glucose tolerance but also as a disorder of protein and fat tolerance.
Type 1 diabetes produces profound ß-cell failure with secondary insulin resistance, whereas type 2 diabetes causes less severe insulin deficiency but greater impairment of insulin action. Given their similarities overall, it is not surprising that the two major forms of diabetes share many pathophysiologic features. However, despite the apparent phenotypic similarity, the underlying pathogenetic mechanisms leading to type 1 and type 2 diabetes are strikingly different.
Type 1 diabetes results from an interplay of genetic, environmental, and autoimmune factors that selectively destroy insulin-producing ß cells (see Fig. 242-2 ).
The role of genetic factors in type 1 diabetes is underscored by data in identical twins showing concordance rates of 30 to 40%. It has been assumed that because concordance rates are not 100%, environmental factors must be important for disease expression. Although the presence of an environmental trigger is likely, it should be recognized that even identical twins do not express identical T-cell receptor and immunoglobulin genes; therefore, total concordance would not be expected for autoimmune diseases such as type 1 diabetes.
Many of the genes linked to type 1 diabetes have not been identified, but some are known. HLA genes, located on the short arm of chromosome 6, clearly play a dominant role; in nonaffected siblings, the risk of developing diabetes is 15 to 20% if they are HLA-identical, approximately 5% if they share one HLA gene, and less than 1% if no HLA genes are shared. Specific HLA haplotypes have been linked to type 1 diabetes: 90 to 95% of type 1 patients express DR3 and/or DR4 class II HLA molecules (as compared with 50 to 60% of the general population), whereas 60% express both alleles, a rate more than 10-fold that of the general population. Another class II allele, DQB1 *0602, has a negative association with the disease. Specific class II DQ haplotypes (e.g., DQ8 and DQ2) even more strongly correlate with disease susceptibility in caucasian individuals; this susceptibility is associated with polymorphisms of the allele encoding the ß-chain of the DQ class II HLA molecule. The presence of aspartic acid at position 57 protects against disease, while substitution of a neutral amino acid at this position is associated with higher disease frequency. Other polymorphisms, such as the substitution of arginine at position 52 of the DQ a-chain, may confer additional risk. Overall, it seems clear that significant genetic heterogeneity exists, and that no single class II HLA gene accounts for all HLA-associated susceptibility to disease. Association of the disease with specific class II HLA genes implies the involvement of CD4+ T cells in the autoimmune process, because these molecules are critical for both the presentation of antigenic peptides to CD4+ T cells and the selection of the CD4+ T-cell repertoire in the thymus.
Other genes likely to contribute genetic susceptibility to type 1 diabetes include IDDM 2 (chromosome 11p), a noncoding promoter region of the insulin gene that may influence insulin gene expression in the thymus (and may therefore affect thymic selection of insulin-reactive T cells), and CTLA-4 (chromosome 2q), which plays a role in T-cell action and regulation. Many other genes have also been implicated, underscoring the polygenic nature of this disease.
Although environmental factors such as diet and toxins have been proposed as triggers of diabetes, most of the scientific attention has focused on putative viruses. Epidemics of mumps, congenital rubella, and coxsackievirus have been associated with an increased frequency of type 1 diabetes. In one instance, coxsackievirus B4 was isolated from the pancreas of a child who died of diabetic ketoacidosis, and inoculation of the virus into mice caused diabetes, fulfilling Koch’s postulates. However, it is likely that acute, lytic viral infections are responsible for only an occasional case of diabetes. Instead, if viruses are involved, it is far more likely that they trigger an autoimmune response. If a virus contains an epitope resembling a ß-cell protein, viral infection could theoretically abrogate self-tolerance and trigger autoimmunity.
About 80% of patients with new-onset type 1 diabetes have islet cell antibodies. Antibodies to a variety of ß-cell constituents have been identified, including insulin, isoforms of glutamic acid decarboxylase (GAD 65 and GAD 67), and the secretory granule protein ICA 512 or IA-2, which contains a tyrosine phosphatase-like domain. The idea that type 1 diabetes is a chronic autoimmune disease with acute manifestations is supported by the fact that islet antigen-directed antibodies are present in approximately 3% of asymptomatic first-degree relatives of patients; such antibody-positive individuals are at high risk for the development of type 1 diabetes, although clinical onset may be delayed by many years. The likelihood of type 1 diabetes is greater than 50% if autoantibodies are present to more than one ß-cell antigen (i.e., insulin, GAD 65, ICA 512), whereas diabetes rarely develops in antibody-negative relatives. If antibodies appear at a young age, the risk for clinical diabetes is particularly high.
The listed antibodies appear to be markers for, rather than the cause of, ß-cell injury. ß-Cell destruction (by apoptotic and cytotoxic mechanisms) is mediated by a variety of cytokines, or by direct T lymphocyte activity. Supporting this notion, type 1 diabetes has been transferred through bone marrow cells from a diabetic patient to a nondiabetic recipient. Additionally, autopsies performed on patients dying soon after disease onset have shown islet-restricted monocytic cellular infiltrates (termed insulitis) that are composed of CD8+ and CD4+ T cells, macrophages, and B cells. Usually, as the disease progresses, the islets become completely devoid of ß cells and inflammatory infiltrates; a, d, and pancreatic polypeptide cells are left intact, thus illustrating the exquisite specificity of the autoimmune attack. At the time of clinical diagnosis, about 5 to 10% of the original ß-cell mass typically remains (see Fig. 242-2 ).
A critical role for T cells is supported by studies involving pancreatic transplantation in identical twins. Monozygotic twins with diabetes who received kidney and pancreas grafts from their nondiabetic, genetically identical sibling required little or no immunosuppression for graft acceptance. Nevertheless, the islets were soon selectively invaded with mononuclear cells, predominantly CD8+ T cells, with the subsequent recurrence of diabetes. Thus, decades after the original onset of disease, the immune system retained the ability to selectively destroy ß cells. Evidence implicating T cells also derives from clinical trials using immunosuppressive drugs. Drugs such as cyclosporine slow or prevent the progression of recent-onset diabetes, but immunosuppression must be continuous to maintain the effect. Further supporting data for a primary role for T cells derives from NOD mice, in which insulitis and islet autoantibodies develop at about 4 weeks of age, and diabetes ultimately develops after 12 to 24 weeks; in these mice, a variety of treatments designed to deplete T cells can prevent diabetes. Most importantly, adoptive transfer of T cells isolated from diabetic mice donors into immune-incompetent NOD mice rapidly produces diabetes. Both CD4+ and CD8+ T cells are generally required for transfer of disease, which suggests that both are necessary for disease expression. These diabetogenic T cells target specific ß-cell antigens, including insulin and GAD. A likely role for GAD and/or insulin is also suggested by data showing that if NOD mice are made tolerant to GAD or to insulin (or to peptides derived from these molecules) early in life, insulitis and diabetes fail to develop. Finally, the chronic, smoldering nature of type 1 diabetes suggests the presence of regulatory or protective influences. In keeping with this observation, T cells that protect the islet cell from immune attack have been isolated from the islets of NOD mice. Such findings suggest that the rate of appearance and clinical expression of disease may be modulated by the balance between diabetogenic and protective populations of T cells. “Tipping the scales” in favor of protective T-cell proliferation is the goal of protective immunization.
Hyperglycemia in type 2 diabetes likely results from complex genetic interactions, the expression of which is modified by environmental factors such as body weight and exercise. With type 2 diabetes, identical twin concordance rates approach 100%, although disease onset and course can vary greatly based on environmental factors. Hyperglycemia itself is known to impair insulin secretion and action; elevated free fatty acid levels also play a pathogenic role. By the time that hyperglycemia is detected, nearly all type 2 patients exhibit both defective insulin secretion and insulin resistance; this makes it


Figure 242-4 Elevations of circulating glucose initiate a vicious cycle in which hyperglycemia begets more severe hyperglycemia.
difficult to determine which of the two factors is primarily responsible for the vicious cycle leading to disease ( Fig. 242-4 ).
Although monogenic forms of diabetes have been identified (e.g., MODY types 1 through 6), the vast majority of cases are polygenic in nature. Type 2 diabetes shows clear familial aggregation but does not segregate in classic mendelian fashion; this implies that the disease results either from a combination of genetic defects or from the simultaneous presence of multiple susceptibility genes in the presence of predisposing environmental factors. Candidate gene mutations for polygenic forms of type 2 diabetes include mutations of the coding region of the insulin gene, peroxisome proliferator-activated receptor gamma (PPAR-?), intestinal fatty acid binding protein 2 (FABP 2), calpain 10, and the ß-3-adrenergic receptor. These and other mutations have been associated with isolated patient clusters of type 2 diabetes.
Fasting insulin levels in type 2 diabetes are generally normal or elevated, yet they are relatively low given the degree of coexisting hyperglycemia. As the disease progresses and hyperglycemia becomes more severe, basal insulin levels eventually fail to keep up and may even decline. The insulin secretory defect usually correlates with the severity of fasting hyperglycemia and is more evident following carbohydrate ingestion. In its mildest form, the ß-cell defect is subtle, involving the loss of the first-phase insulin response and the normal oscillatory pattern of insulin secretion. Although the overall insulin response may be fairly intact, this “normal” response is actually inadequate to maintain glucose tolerance when viewed in the context of simultaneous insulin resistance. During this early stage, the ß-cell defect is usually specific for glucose; other secretagogues (e.g., amino acids) maintain their potency, and insulin deficiency is thus less pronounced during the ingestion of mixed meals. Patients with more severe fasting hyperglycemia lose this capacity to respond to the other insulin secretagogues; thus, their secretory defect worsens as their disease progresses. Unfortunately, the underlying cause of the secretory defect remains uncertain and is likely multifactorial.
Studies in rodents suggest that the loss of glucose-stimulated insulin secretion is followed by a decreased expression of GLUT 2, the primary glucose transport protein of the pancreatic ß-cell. Such a loss of GLUT 2 during the clinical transition to diabetes would likely accelerate the decline of glucose-stimulated insulin secretion. Pathologic studies of islets from patients with long-standing type 2 diabetes have demonstrated amyloid-like deposits composed of islet amyloid polypeptide, or amylin, a peptide synthesized in the ß-cell and cosecreted with insulin. Chronic hypersecretion of amylin may lead to precipitation of the peptide, which over time might also contribute to impaired ß-cell function. Recent experiments in gene knockout mice suggest a potential role for impaired insulin receptor signaling in the development of impaired ß-cell function. A link between insulin resistance and secretion is also suggested by data showing that accumulation of fat within the ß-cell (as a result of hyperglycemia, insulin resistance, and increased fatty acid turnover) may further reduce insulin secretion.
With few exceptions (e.g., a subgroup of African American patients), type 2 diabetes is characterized by impaired insulin action. The insulin dose-response curve for augmenting glucose uptake in peripheral tissues is shifted to the right (representing decreased insulin sensitivity), and maximal response is reduced, particularly in the setting of severe hyperglycemia. Other insulin-dependent processes, such as inhibition of hepatic glucose production and lipolysis, also show reduced sensitivity to insulin. The mechanisms responsible for insulin resistance remain poorly understood.
Early studies of insulin resistance focused on defects of the insulin receptor. Mutation of the insulin receptor gene can produce leprechaunism, characterized by severe growth retardation, extreme insulin resistance, and early infant death. Other syndromes related to mutated insulin receptors include the Rabson-Mendenhall syndrome, also associated with tooth and nail abnormalities and pineal gland hyperplasia, and “type A insulin resistance,” most often affecting young females with acanthosis nigricans, polycystic ovaries, and hirsutism. Another example of extreme insulin resistance involves the presence of anti-insulin receptor antibodies, which is associated clinically with acanthosis nigricans and other autoimmune phenomena.
Although insulin receptors are rarely abnormal in type 2 patients, defects in more distal “post-receptor” pathways play a far greater role in insulin resistance. One important aspect of resistance is a reduced capacity for translocation of GLUT 4 to the cell surface in muscle cells. A separate defect in glycogen synthesis is also likely to be present. Whether the defects uncovered are primary or secondary to the disturbance in glucose metabolism is uncertain; possibly, a variety of genetic abnormalities in cellular transduction of the insulin signal may individually or in concert produce an identical clinical phenotype. It is uncertain whether mechanisms of insulin resistance in nonobese patients are identical to those of their obese counterparts; however, the coexistence of obesity clearly accentuates the severity of the resistant state. In particular, upper body or abdominal (as compared with lower body or peripheral) obesity is associated with insulin resistance and diabetes. Intra-abdominal visceral fat deposits, detected by computed tomography or magnetic resonance imaging, have a higher lipolysis rate and are more resistant to insulin than peripheral fat. The resulting increase in circulating free fatty acid levels promotes fat deposits in the liver and muscle, worsening insulin resistance. Intracellular free fatty acid metabolites appear to promote insulin resistance through complex mechanisms, involving serine (rather than tyrosine) phosphorylation of insulin signaling molecules. Cortisol hypersecretion and/or hereditary factors may also influence the distribution of body fat, the latter contributing an additional genetic influence on the expression of disease.
Adipocytes, once thought of as inert fat storage cells, are now known to produce a number of metabolically active hormones that may affect insulin sensitivity. Leptin, for example, acts on the hypothalamus to promote satiety and energy expenditure and may accelerate glucose metabolism. Adiponectin (Acrp30), another fat-derived hormone, circulates at levels that correlate inversely with both adiposity and degree of insulin resistance. The administration of adiponectin to obese mice causes a transient, dose-dependent, insulin-independent decrease in circulating glucose levels; adiponectin also improves insulin sensitivity by decreasing triglycerides in the liver and muscle, likely by increasing the expression of molecules involved in fatty acid combustion and energy dissipation. Finally, adipose tissue is an abundant source of the cytokine tumor necrosis factor-a, which is known to inhibit muscle glucose metabolism by inducing serine phosphorylation of insulin signalling molecules. The precise impact that these and other adipocyte-derived factors exert on insulin resistance has yet to be established; these proteins may well play an important role in the pathogenesis of diabetes.
Hyperglycemia per se impairs the ß-cell response to glucose and promotes insulin resistance. Reversal of glucotoxicity can disrupt the vicious cycle that perpetuates hyperglycemia (see Fig. 242-4 ). Circulating lipids can also adversely affect glucose metabolism; increased free fatty acid levels accelerate hepatic gluconeogenesis, inhibit muscle glucose metabolism, and may impair pancreatic ß-cell function. As is the case with glucotoxicity, the reversal of lipotoxicity can rapidly improve metabolic control and facilitate favorable therapeutic outcomes.
It remains uncertain whether insulin resistance or defective insulin secretion is the primary defect in type 2 diabetes. This issue is difficult to resolve once diabetes has developed; therefore, research attention has focused primarily on high-risk, nondiabetic subjects. Studies in high-risk populations (e.g., Pima Indians, Mexican Americans) have suggested that insulin resistance is the initial defect; similar findings have been reported in first-degree relatives of type 2 diabetic patients and in healthy prediabetic offspring of two diabetic parents. Interestingly, hyperinsulinemia has been detected in prediabetic subjects as early as one to two decades before clinical onset, suggesting that the development of diabetes can be exceedingly slow. Although these studies support the view that


Figure 242-5 A proposed sequence of events leading to the development of type 2 diabetes: insulin resistance resulting from genetic influences, central obesity, inactivity, or a combination of these factors leads over time to a progressive loss of the ß-cell’s capacity to compensate for this defect.
insulin resistance generally antedates insulin deficiency, the presence of insulin resistance alone is generally insufficient to generate disease; this implies that for diabetes to occur, impaired insulin secretion is also required ( Fig. 242-5 ). It is possible that the appearance of a secretory defect is a secondary phenomenon, possibly resulting from “ß-cell exhaustion,” excess fatty acid delivery, and/or amylin accumulation. Alternatively, diminished insulin secretion may result from an independent defect that becomes evident only upon chronic ß-cell stimulation, such as a subtle genetic defect in insulin signaling.
The sequence of events described—underlying insulin resistance followed by a secretory defect—is common but clearly does not describe all type 2 diabetic patients. For example, a subgroup of African American patients exhibits little or no insulin resistance. Additionally, diminished glucose-stimulated insulin secretion is seen in women with gestational diabetes in whom type 2 diabetes later develops. Finally, the demonstration of functional ß-cell-associated gene mutations in patients with MODY indicates that primary ß-cell defects are capable of producing a similar phenotype. Taken together, these lines of evidence strongly suggest that type 2 diabetes cannot be explained by insulin resistance alone or by any single pathogenic mechanism.
Relationship Between Diabetes Control and Its Complications
Whether the vascular and neuropathic complications of diabetes mellitus can be prevented or delayed by improved glycemic control was debated for more than a half century. To answer the question, the National Institutes of Health initiated the Diabetes Control and Complications Trial (DCCT), a 9-year multicenter study involving 1441 type 1 patients aged 13 to 39 years who were randomly assigned to either intensive insulin therapy or conventional care. Intensive therapy consisted of three or more insulin injections per day (or an insulin pump), self-monitoring of blood glucose at least four times per day, and frequent contact with a diabetes health care team. Conventional care consisted of one or more (commonly two) injections of insulin per day, less frequent glucose monitoring, standard education, and less frequent health care visits. The target goals of therapy were different as well. The intensive therapy group sought pre-meal blood sugar levels of 70–120 mg/dL, postprandial blood levels of less than 180 mg/dL, and glycohemoglobin values as close to normal as possible. In the conventional care group, the primary goal was simply to maintain clinical well-being. Patients were divided into two groups: (1) a primary prevention group, with diabetes for 1 to 5 years and no detectable complications, and (2) a secondary intervention group, with diabetes for 1 to 15 years and mild nonproliferative retinopathy. Remarkably, nearly 99% of enrolled patients completed the trial.
The DCCT achieved a clear separation of glucose levels between the groups over the entire study period. Glycohemoglobin (Hb A1c ) and mean glucose levels in the intensive therapy group were 1.5 to 2.0% and 60 to 80 mg/dL lower than in those receiving conventional care. Although considerable variability was noted among individual patients, most of the intensive care group failed to achieve normal glucose levels (glycohemoglobin averaged 1.1% above normal, or a glucose level of about 155 mg/dL). Nevertheless, intensive therapy reduced the development of retinopathy by 76% in the primary prevention group, and the progression of retinopathy by 54% in the secondary intervention group ( Fig. 242-6 )[1] ; the latter effect became apparent after only 4 years. In addition, intensive therapy reduced the risk of microalbuminuria by 39%, frank proteinuria by 54%, and clinical neuropathy by 60%. The incidence of major cardiovascular events also tended to be lower, but the number of events was insufficient for statistical proof; at the very least, intensive therapy did not pose a risk for macrovascular complications. An exponential relationship over time between the average blood glucose level (as reflected by Hb A1c ) and the progression of retinopathy in the intensive care group suggests that there may be no threshold level at which complications occur. These findings imply that any degree of improvement in glycemic control has benefit, and that normalization of glucose levels is not required to slow the progression of diabetic complications.
The benefits achieved by intensive control in the DCCT were not without risk. Weight gain was more common, and most importantly, the frequency of severe hypoglycemia (including multiple episodes in some patients) was three-fold higher in the intensive care group. In many cases, such episodes occurred without classic warning symptoms, often while the patient was asleep. Thus, in some patients, the risks of intensive therapy may outweigh the benefits; possibly included are patients with recurrent severe hypoglycemia, patients with advanced complications, young children, and patients who are unable or unwilling to participate in their management (e.g., self-monitoring of blood glucose). Such individuals are likely to benefit from less aggressive therapy designed to moderately lower glucose levels without the risk of hypoglycemia. It is noteworthy that despite

Figure 242-6 A summary of the results of the Diabetes Control and Complications Trial.

the higher rate of hypoglycemia, intensive therapy in the DCCT had no detectable long-term effects on cognitive function.
Although the DCCT did not involve type 2 diabetic patients, a small study using a similar experimental design in lean Japanese patients with type 2 diabetes showed similar results. More conclusive evidence that improved control of blood glucose is beneficial for type 2 diabetic patients derives from the United Kingdom Prospective Diabetes Study (UKPDS). The UKPDS recruited 5102 patients with newly diagnosed type 2 diabetes between 1977 and 1991. After 3 months of diet therapy, the 3867 patients with fasting glucose levels between 6.1 and 15.0 mmol/L (110 and 270 mg/dL) were randomized to a more intensive regimen consisting of sulfonylurea, metformin (for obese patients only) or insulin, or a conventional treatment regimen focused primarily on symptom reduction. Subjects were monitored for an average of 10 years. Although glycemic control gradually deteriorated in both groups, the intensified treatment group had lower mean Hb A1c than their conventionally treated counterparts (7.0% versus 7.9%). This modest improvement significantly reduced microvascular complications by 25% and reduced all diabetes-related events by 12%.[2] The intensified treatment group also had a 16% reduction in a combined end point—nonfatal or fatal myocardial infarction or sudden death—that did not quite reach statistical significance (P=.052). A continuous relationship was again noted between glycemic control and diabetic complications; also similar to the DCCT, no glycemic threshold for microvascular complications was observed. Importantly, serious adverse events were rare for all of the treatment arms in the UKPDS, and only a single death from hypoglycemia occurred in more than 27,000 patient years of intensive therapy.
What conclusions can be drawn from the DCCT and the UKPDS? The primary message is that “glucose matters.” In both type 1 and type 2 diabetic patients who are willing and able to actively participate in their management, the goal should be to achieve the best possible level of glycemic control as rapidly as possible without undue risk. The DCCT and UKPDS also demonstrate that most patients benefit from lower glucose levels, even if normalization is not achieved; for most type 2 patients, effective glucose reductions can be achieved by diet, oral agents, or less complicated insulin regimens than are required in type 1 patients. The greatest challenge for the clinician is how to effectively apply the DCCT and UKPDS results in practice, a formidable task. Both study groups were highly motivated and compliant. Furthermore, management was supervised by an experienced health care team that was able to devote more time to patients than is usually feasible. An important lesson from these studies was that successful treatment of diabetes was largely accomplished through the efforts of the patients themselves, as well as by nurse educators, dietitians, and diabetes counselors. It makes sense, then, to encourage the use of physician-directed health care teams to translate the findings of the DCCT and UKPDS into clinical practice.

Treatment of diabetes mellitus involves changes in lifestyle and may require pharmacologic intervention with insulin or with oral glucose-lowering drugs. In type 1 diabetes, the primary focus is to replace lost insulin secretion; lifestyle changes are required to facilitate insulin therapy and to optimize health. For patients with type 2 diabetes, changes in lifestyle are the cornerstone of treatment (especially in the early stages of disease), and pharmacologic intervention is a secondary treatment strategy. Although therapeutic strategies differ for the two common forms of diabetes, the treatment goals are essentially identical. In the short term, the goals of diabetes treatment are to optimize metabolic control and improve the patient’s sense of clinical well-being. Long-term therapeutic goals focus on the prevention of complications, including cardiovascular disease, nephropathy, retinopathy, and neurologic disease.
A variety of highly purified insulin preparations are commercially available that differ mainly in their pharmacokinetics ( Table 242-5 ). Premixed insulin preparations are also available and may offer added convenience for selected patients. Nearly all insulin preparations contain 100 U/mL (U-100), although a more concentrated 500 U/mL preparation of regular insulin

Lispro or Aspart
10–15 min
Regular (R)
30 min
NPH (N) or Lente
2–4 hours
Ultralente (U)
4–6 hours

2–4 hours
no peak

(U-500) for severely resistant patients can be obtained. Human insulin is now the only form of insulin sold in North America and other industrialized countries, largely because of immunologic concerns. Because human insulin generates lower titers of insulin antibodies than porcine or bovine insulin, it acts more rapidly after injection and has a shorter duration of action, allowing better synchrony between insulin peaks (after premeal injection of rapid-acting insulin) and the absorption of meals. It is noteworthy that the same insulin preparation can produce variable responses in a single patient, since the peak and duration of most insulin preparations are influenced by the site of administration, skin temperature, the depth of injection, and the magnitude of the insulin dose.
Rapid- and Short-Acting Insulin Preparations.
After subcutaneous injection, regular (R) insulin begins to act in about 30 minutes and should therefore be administered 25 to 30 minutes before a meal. Because it acts quickly and has a relatively short duration of action (5–8 hours), it is effective for blunting postprandial glucose excursions and for facilitating rapid dose adjustments based on measured blood glucose values. The properties of regular insulin are especially helpful in managing glucose elevations that occur during illness, or after the consumption of large meals. Given intravenously, regular insulin is also effective in the perioperative period and in the management of severely ill patients and acute hyperglycemic complications.
In regular insulin preparations, insulin molecules exist predominantly in hexameric form. Before being absorbed, insulin hexamers must first be diluted in subcutaneous interstitial fluid, then dissociate into single molecules; this property accounts for the slightly delayed absorption of regular insulin from subcutaneous injection sites. Recently, advances in recombinant DNA technology have led to the development of insulin analogues intended to bypass this property, allowing for more rapid absorption. In 1996, lispro insulin was introduced. This insulin analogue, in which the amino acids in positions B28 (lysine) and B29 (proline) have been reversed, has a reduced capacity for hexameric self-association and is therefore more rapidly absorbed. Its effects begin within 10 to 15 minutes of administration, and generally wane within 3 to 4 hours. Because of its rapid onset, lispro can be given just before eating (as opposed to 30 minutes prior), a feature that greatly simplifies the planning and consumption of meals; also, because its effects wane more rapidly, there is a reduced risk of “late” hypoglycemia if the next meal is delayed. Using lispro insulin, postprandial glucose and hemoglobin A1c reductions are equal to or better than those achieved with regular insulin, and there is a reduced incidence of delayed hypoglycemia. For these reasons, and because of greater convenience and flexibility, lispro is being used with increasing frequency in intensive treatment regimens.
Insulin aspart, the second rapid-acting insulin analogue approved by the Food and Drug Administration (FDA), was released in 2001. In insulin aspart, a neutral proline residue at position B28 is replaced by negatively charged aspartic acid, resulting in a reduced capacity for self-association and faster absorption. The pharmacokinetic properties of insulin aspart are similar to those of insulin lispro; insulin aspart may have a slightly longer duration of effect.
Intermediate- and Long-Acting Insulin Preparations.
The longer-acting insulin preparations have been modified to delay their absorption from injection sites, resulting in a longer duration of insulin activity. The addition of protamine and zinc yields intermediate-acting Neutral Protamine Hagedorn (NPH) insulin, whereas enlarging the size of the zinc-insulin crystal yields Lente (intermediate-acting) and Ultralente (long-acting) insulins. NPH and Lente, the

intermediate-acting insulins, have a similar time course of action; when given twice per day, they offer a compromise between some degree of meal coverage (coinciding with peak activity) and the provision of basal insulin levels. Ultralente insulin, because of its longer duration and somewhat less evident peaks, offers possible advantages for basal insulin replacement and can be given once daily in some patients. However, the pharmacokinetics of Ultralente are less predictable in clinical practice (even within a single patient), and its effects commonly require twice-daily dosing, limiting its utility.
Insulin glargine, approved by the FDA in 2001, differs from human insulin both at position A21, where asparagine is replaced by glycine, and at the C-terminus of the B-chain, where two arginine residues have been added. Insulin glargine is soluble at acidic pH and less so in physiologic conditions; injected at a pH of 4, it is neutralized in subcutaneous tissue and forms microprecipitates, delaying its absorption and prolonging its duration of activity. The primary advantages of glargine insulin are greater than 24-hour activity (allowing once-daily dosing) and the lack of peak concentrations; both characteristics are desirable for the provision of consistent basal insulin levels. Disadvantages include higher cost, a higher incidence of mild injection site discomfort, and the inability to mix glargine with other insulins. Clinical trials in type 1 diabetic patients suggest that insulin glargine produces similar or slightly larger reductions in hemoglobin A1c as compared with NPH; comparative trials have also shown a reduced incidence in nocturnal hypoglycemia when insulin glargine is used. In type 2 diabetic patients, the clinical differences between glargine and NPH are less significant; in comparative trials, hemoglobin A1c reductions are generally equivalent, and there are smaller differences in the incidence of hypoglycemia. Other long-acting insulins, intended for use in basal insulin therapy, are currently in development.
While a simple concept, the clinical use of insulin to treat diabetes mellitus can be extraordinarily complex. There are many important inter-patient (and intra-patient) variables, so a predictable algorithm cannot be uniformly applied to all patients, nor to a single patient at all times. In general, subcutaneous insulin regimens for type 1 diabetes may be classified as “conservative” or “intensive.” Modes of continuous subcutaneous insulin infusion (i.e., insulin pumps) have also gained popularity in recent years.
Conservative Insulin Therapy.
Through the early stages of type 1 diabetes, some degree of ß-cell function is usually preserved, allowing many patients to achieve glycemic control with less intensive effort. Because intermediate-acting insulins are not generally sustained over a 24-hour period, and because insulin requirements tend to increase early in the morning, these patients should start with two daily injections, consisting of a mixture of intermediate-acting and rapid-acting human insulins administered before breakfast and before dinner. Although Lente insulin has a theoretic advantage over NPH in that it does not contain a foreign protein (protamine), this difference seems to have negligible clinical significance. In fact, NPH may be preferable to Lente when insulins are mixed, because the excess zinc in Lente preparations can cause regular insulin to precipitate out of solution, delaying its absorption.
There are many acceptable approaches to the initiation of conventional insulin therapy (see the section on the treatment of type 2 diabetes). Regardless of the initiation method used, insulin dose adjustments will inevitably be required. Initially, doses of the intermediate-acting insulin should be adjusted to optimize pre-dinner and fasting (morning) glucose levels. Once these goals are accomplished, rapid-acting insulin doses should then be adjusted to optimize postprandial, pre-lunch, and bedtime glucose values. Patients should generally inject in the same anatomic region at the same time each day (i.e., in the abdomen in the morning, in the thigh at night) to ensure consistent insulin delivery; an effort should also be made to avoid exact duplication of injection sites within a 1-week period. Some patients may experience a brief “honeymoon” period, during which ß-cell function partially recovers and insulin needs are temporarily reduced. Such an improvement should not be used as a signal to reduce efforts aimed at glycemic control, since continuation of optimal insulin therapy will help to preserve residual ß-cell function.
Multiple Subcutaneous Injections.
Several years after the onset of type 1 diabetes, residual insulin secretion typically ceases. When this occurs, twice-daily insulin injections are no longer acceptable, even if they continue to successfully control diabetic symptoms. For optimal glycemic control, insulin delivery should more closely simulate the “normal” pattern of insulin secretion; namely, continuous or “basal” insulin levels are required throughout the day, while brief increases in insulin levels (“boluses”) should coincide with the ingestion of meals. The primary problem with twice-daily insulin regimens is that the glucose-lowering effect of the pre-dinner intermediate-acting insulin is greatest at the time when insulin requirements are at their lowest (i.e., around 3:00 AM). Additionally, when requirements are increasing in the pre-dawn hours, insulin levels are declining. The net results of this poorly matched insulin supply and demand are the production of nocturnal hypoglycemia and/or fasting (morning) hyperglycemia.
Successful management of diabetes begins with fasting glucose control. Failure to control morning sugars often results in the stubborn perpetuation of hyperglycemia throughout the day. Once hepatic gluconeogenesis has been activated in the morning, it is not readily suppressed by insulin injections. The key factors responsible for fasting hyperglycemia are inadequate overnight delivery of insulin and sleep-associated growth hormone release. The “dawn phenomenon” is most pronounced in patients with type 1 diabetes because of their inability to compensate by raising endogenous insulin secretion. The magnitude of the dawn phenomenon can be attenuated by designing insulin regimens to ensure that the effects of exogenous insulin do not peak in the middle of the night and dissipate by morning. Several approaches to insulin therapy can deal with this problem; some of the more common regimens are displayed in Figure 242-7 . One common approach is to use three injections: a mixture of intermediate- and rapid- or short-acting insulin before breakfast, rapid- or short-acting insulin before dinner, and intermediate-acting insulin at bedtime. The primary disadvantage of this approach is that meal sizes and schedules must be fixed rather rigidly. Alternative multidose regimens incorporate short- or rapid-acting insulin injections before each meal, with one or two daily doses of intermediate- or long-acting insulin (e.g., glargine). Pen-style insulin injectors are also available; these may help to make multidose regimens more convenient and tolerable for patients.
Continuous Subcutaneous Insulin Infusion (CSII).
In CSII, rapid-acting insulin is administered around the clock by a battery-powered, externally worn, computer-controlled infusion pump (see Fig. 242-7 ). The

Figure 242-7 Several intensive insulin regimens commonly used in the treatment of diabetes. Each is designed to provide a continuous supply of insulin around the clock and to make extra insulin available at the time of meals, thereby simulating more closely the normal physiologic pattern of insulin secretion.

pump delivers a continuous basal rate and can be programmed to vary the flow rate automatically for set time periods, such as reducing the flow rate after bedtime and increasing flow to compensate for increased insulin requirements in the pre-dawn hours. Boluses, determined by self-monitoring of blood glucose and expected meals, are given by manual pump activation. Most insulin pumps contain an insulin reservoir attached to a subcutaneous catheter (the catheter is inserted using an introducing needle, which is then removed). Catheters are generally best placed in the abdomen, to standardize absorption and maximize visibility. Overall, the CSII method provides diabetic patients with the highest degrees of lifestyle flexibility and glucose control.
The CSII approach has several limitations. One obvious disadvantage of pump therapy is the wearing of the pump itself; the device may be undesirable for patients during intense exercise, contact sports, submersion in water, or personal intimacy. Furthermore, because CSII uses rapid- or short-acting insulin, any interruption in flow (most commonly because of insulin precipitation within the catheter) can lead to rapid deterioration of metabolic control. Local infections at the catheter site occasionally occur, necessitating a site change every 2 to 3 days. Furthermore, maintenance of the pump and appropriate insulin infusion rates requires significant patient effort and sophistication.
The intensive treatment regimens described above are not for everyone. In appropriate patients, however, intensive insulin therapy should be strongly encouraged to reduce the risk of late diabetic complications. It should also be noted that pregnancy is an absolute indication for intensive therapy, and that reduction of the excess neonatal morbidity and mortality associated with diabetic pregnancies requires tight glycemic control. Ideally, intensive insulin therapy should be instituted in type 1 patients before conception, to minimize the risk of fetal anomalies. After conception, blood glucose targets are more stringently applied than at other times, with the specific aim of maintaining glucose levels in the normal range.
Diet and exercise contribute importantly to the care of patients with type 1 diabetes. Patients should be educated about balancing caloric intake (diet) with energy expenditure (exercise) and should understand the basic concepts of insulin therapy as it relates to stress and physical activity. If properly managed and sufficiently motivated, diabetic patients should be able to consume the foods they enjoy and should be able to fully participate in exercise and sports.
The introduction of intensive insulin regimens has increased meal flexibility by allowing more latitude in varying the size, content, and timing of meals. New approaches offer the opportunity for a more normal lifestyle, thus minimizing compliance problems and optimizing patient acceptance and satisfaction. Meals should be nutritionally sound and should provide sufficient calories to meet the energy needs of growing children, active young adults, and pregnant women; the 1800-kcal diet classically prescribed for type 2 patients is grossly insufficient in these and other individuals. Furthermore, diabetic diets should be specifically aimed at minimizing long-term cardiovascular risk by minimizing the ingestion of sodium, cholesterol, and saturated fats ( Table 242-6 ).
Because type 1 patients depend on exogenous insulin, proper management is facilitated by a meal plan designed to match the time course of the selected insulin regimen. Patients should learn to compensate for meal-plan departures by adjusting their insulin doses and for periods of altered activity by adjusting their consumption of food. Effort should be made to avoid long delays between meals, and small snacks may be needed at times of peak insulin action to avoid hypoglycemia. Most patients, regardless of their regimen, should incorporate a bedtime snack to reduce the risk of nocturnal hypoglycemia. Finally, the potential for insulin-induced weight gain requires special emphasis on portion control; to control hypoglycemia, patients should master the use of appropriate carbohydrate intake and avoid overcompensation.
Regular exercise is important to promote overall health and to reduce cardiovascular complications. Surprisingly, there is little evidence to suggest that exercise itself substantially improves glycemic control in type 1 diabetes, although it is known to reduce overall insulin requirements by enhancing insulin sensitivity. Through accelerated insulin absorption (due to increased local blood flow at the injection site) and increased muscle glucose consumption, exercise can rapidly reduce blood glucose levels, particularly when it coincides with the peak action of an insulin injection. In nondiabetic individuals, blood glucose levels remain stable during exercise, as decreased

I. Dietary prescription
1. Weight reduction, gain, or maintenance, to achieve and maintain ideal body weight
2. Restriction of saturated fat to <10% of total calories, to be replaced in the diet by carbohydrates and monounsaturated fats. If LDL reduction is also desired, saturated fats should be further restricted to <7% of daily caloric intake
3. Decreased cholesterol intake to <300 mg per day. If LDL reduction is also desired, cholesterol intake should be further restricted to <200 mg per day
4. Sodium restriction (<2.4 g per day) in patients with hypertension; in those with overt nephropathy, sodium intake should be further restricted to <2.0 g per day
5. Protein restriction to <20% of total calories; with nephropathy, protein intake should be further restricted to <0.8 mg/kg/day, or to ~10% of daily caloric intake
II. Exercise prescription*
1. A combination of aerobic exercise and resistance training is preferred. Avoid heavy lifting, straining, and Valsalva maneuvers, which can raise blood pressure and may aggravate proliferative diabetic retinopathy.
2. Intensity: Increase heart rate “moderately” to at least 55% of “maximal” heart rate (220 minus age in years), with adjustments based on the patient’s cardiovascular fitness. Patients with improved cardiovascular fitness can proceed to “harder” activities, achieving heart rates exceeding 70% of maximum.
3. Duration: 30 minutes, preceded and followed by stretching and flexibility exercises for a minimum of 5–10 minutes.
4. Frequency: at least 3 days per week. Results are best if exercise occurs nearly every day.
5 Avoid strenuous exercise if fasting glucose levels are =250 mg/dL. Avoid all forms of exercise if glucose levels are =300 mg/dL and/or ketosis is present.
6. Monitor blood glucose before, during, and after exercise, to learn responses to different exercise conditions and to identify when changes in insulin and/or food intake are necessary.
7. Consume added foods as needed to avoid hypoglycemia. A rapidly-absorbed carbohydrate source should be readily available during exercise and for up to 8 hours after exercise is completed.
LDL = low-density lipoprotein.

*Exercise limitations are imposed by preexisting coronary or peripheral vascular disease, proliferative retinopathy, peripheral or autonomic neuropathy, and/or poor glycemic control.

endogenous insulin secretion promotes increased hepatic glucose production to match the increased rate of glucose consumption. In diabetic patients receiving exogenous insulin, however, this “finely tuned” homeostatic mechanism is greatly disturbed. The continued presence of exogenous insulin during exercise further accelerates glucose uptake and (more importantly) blocks the compensatory increase in glucose production; as a result, circulating glucose levels can fall precipitously during exercise. Because the magnitude of this fall is not easily titrated, hypoglycemia may occur if the patient is unable to appropriately adjust diet and insulin before, during, and after physical activity ( Table 242-6 ).
Nonpharmacologic Measures
In many type 2 diabetic patients, diet and exercise are the only therapeutic interventions required to restore metabolic control. As a result, the temptation to use pharmacologic agents should be restrained at the outset, unless the patient is symptomatic or hyperglycemia is severe. On the other hand, the clinician must also resist the temptation to be satisfied by the elimination of symptoms, which is simply the first of many steps in the comprehensive treatment of diabetes. The combination of lifestyle changes and medications can reduce both cardiovascular and microvascular events by about 50%.[3]
Irrespective of initial weight, modest weight reduction (on the order of 5 kg) in obese diabetic patients leads to improved glycemic control. The dramatic impact of weight loss is mediated by changes in insulin-responsive tissues, as well as by enhanced ß-cell activity; insulin resistance diminishes, glucose production declines, and

lower glucose levels improve glucose-stimulated insulin secretion. The beneficial effects of weight loss are not restricted to glucose; dietary therapy also yields improved lipid profiles and reductions in systemic blood pressure. In general, it matters little how weight loss is achieved, provided that good health is preserved and adequate nutrition is maintained. Successful weight loss is best achieved by the combination of a supportive environment that emphasizes long-term goals, regular exercise to increase energy expenditure, and long-term behavior modification.
In sedentary diabetic patients, maintenance caloric requirements can be as low as 20 to 25 kcal per kilogram of body weight per day. In these individuals, the classically prescribed 1800-kcal diet will be ineffective in producing weight loss. It is sensible to begin with a nutritionally sound, individually tailored diet that is aimed at producing a caloric deficit of about 500 kcal per day. Because a caloric deficit of approximately 3500 kcal is required to lose 1 lb of body fat, weight loss using this method can be expected at 1 lb per week. For obese patients with a history of multiple failed weight loss attempts, very low calorie diets (<1000 kcal/day) can be useful when carried out under medical supervision. Orlistat, a gastrointestinal lipase inhibitor that reduces dietary fat absorption, can be an effective adjunct for achieving weight loss in some patients; it may also improve glycemic control and lipoprotein profiles. Regardless of the method used, experience tells us that most patients are unable to maintain low-calorie diets for an extended period of time; if successful, the majority of patients regain lost weight. In patients with type 2 diabetes, metabolic factors may also contribute to difficulty maintaining weight loss. Dieting reduces glycosuria and therefore lessens urinary caloric loss. Also, the expected decrease in basal metabolic rate during weight loss is accentuated in diabetic patients, because weight loss reverses both accelerated gluconeogenesis and the futile cycling of substrates; these conditions, commonly seen in poorly controlled diabetes, waste a good deal of energy in the hyperglycemic state.
Even when diabetic patients cannot lose weight, a careful meal plan is a valuable tool for reducing their risk of cardiovascular disease. This benefit is best achieved by restricting saturated fats and cholesterol and by raising the dietary content of carbohydrates and monounsaturated fats. It was originally thought that carbohydrate intake should be restricted in diabetes; however, it is now appreciated that a diet high in carbohydrate (>50%) may improve insulin action and glycemic control, particularly in patients with mild hyperglycemia. In patients with more severe fasting hyperglycemia or with triglyceride elevations aggravated by high-carbohydrate diets, reduced carbohydrate intake (<45% of total calories) and greater reliance on monounsaturated fats may be preferable. It has also been assumed that carbohydrate intake should be focused on complex carbohydrates (starches), and that sucrose should be avoided; however, evidence supporting these assumptions is scarce. Simple sugars appear to raise glucose levels to a similar extent as complex carbohydrates; thus, total carbohydrate intake, rather than type of carbohydrate, should be the primary consideration. Fiber-containing carbohydrates such as oats, gums, legumes, and fruit pectin may also be beneficial, since fiber blunts meal-induced glucose excursions by delaying gastric emptying and carbohydrate absorption. Fiber helps to prevent constipation and may also contribute to lowering of triglyceride and low-density lipoprotein (LDL) cholesterol levels.
Another key component of the diabetic meal plan is to alter patterns of dietary fat. The typical Western diet, high in saturated animal fat, likely contributes to the development of atherosclerosis. Diabetic patients with normal lipid profiles are encouraged to follow the recommendations of the National Cholesterol Education Program (NCEP) by limiting total fat intake to less than 30% of total calories, with less than 10% of calories as saturated fat and less than 300 mg/day of dietary cholesterol (see Table 242-6 ). If low-density lipoprotein levels are elevated, stricter recommendations apply (NCEP Step II diet), with less than 7% of calories as saturated fat and less than 200 mg/day of dietary cholesterol. As mentioned earlier, if elevated triglycerides are of concern, one should consider a moderate increase in monounsaturated fats, to replace dietary carbohydrates; of course, increasing fat intake should always be recommended with caution in patients with obesity. Despite a lack of supporting scientific evidence, moderation of dietary protein is also currently recommended for patients with diabetes; this issue assumes greater importance in patients with proteinuria and overt diabetic nephropathy.
Regular exercise is a powerful adjunct in the treatment of type 2 diabetes. Long-term studies demonstrate consistent beneficial effects of regular exercise on carbohydrate metabolism and insulin sensitivity, which can be maintained for several years. Exercise also facilitates weight loss and its maintenance, which further improves glycemic control and also has beneficial effects on cardiovascular risk: regular exercise lowers triglyceride-rich very low density lipoprotein levels, raises high-density lipoprotein levels, and improves fibrinolytic activity. In general, “moderate” levels of exercise should be prescribed most days of the week (see Table 242-6 ). Limitations may be imposed by preexisting coronary or peripheral vascular disease, proliferative retinopathy, peripheral or autonomic neuropathy, and poor glycemic control.
Pharmacologic Intervention
The hypoglycemic effect of sulfonylureas was first noted in the 1940s, during the development of sulfa antibiotics. Chlorpropamide, the first oral agent approved for use in the United States, was released in 1954. With the exception of phenformin (which was briefly available before being pulled from the market in the 1970s), sulfonylureas were the only oral agents available in the United States for more than 40 years. In contrast, since the 1995 approval of metformin by the FDA, several new classes of oral agents have become available for the treatment of type 2 diabetes ( Table 242-7 and Fig. 242-8 ). Oral agents are indicated in patients in whom diet and exercise fail to achieve treatment goals and may be favored over insulin in older patients with relatively mild degrees of hyperglycemia. Patients with more severe hyperglycemia generally require insulin during the initial phases of treatment; once glucose levels have stabilized and the “toxic” effects of severe hyperglycemia on ß-cell function and insulin action have been minimized, many of these patients can then be converted to oral agents.
Sulfonylureas are insulin “secretagogues,” which act through specific sulfonylurea receptors on the ß-cell surface. Drug-receptor binding acts to close adenosine triphosphate-dependent potassium channels, resulting in cellular depolarization, calcium influx, and the translocation of insulin secretory granules to the ß-cell surface. The resulting release of insulin into the portal vein rapidly suppresses hepatic glucose production, and later facilitates peripheral glucose utilization; insulin resistance commonly diminishes as a result of the reversal of glucotoxicity. Because sulfonylureas rely on a preserved ß-cell response, they are ineffective in the treatment of type 1 diabetes.
Although the sulfonylureas differ in relative potency, effective dosage, metabolism, and duration of action, from a clinical standpoint these differences have marginal significance (see Table 242-7 ). Each drug has similar hypoglycemic effects: at maximally effective doses, an average drop in hemoglobin A1c of 1 to 2% is expected, correlating to average fasting plasma glucose reductions of 40 to 80 mg/dL. Drugs with hepatic metabolism and a shorter duration of action have

Figure 242-8 Mechanism of action of oral glucose-lowering agents.


Insulin secretagogues
30 minutes prior to meals

Hypoglycemia, weight gain, hyperinsulinemia
First generation


K > L


L > K


L > K

Second generation


L > K


L > K
Up to 24

Glyburide micronized

L > K
Up to 24


L > K
Up to 24

Glipizide GITS

L > K

Inhibits hepatic gluconeogenesis
With meals

Gastrointestinal disturbances (abdominal pain, nausea, diarrhea), lactic acidosis

Up to 24

Metformin XR



K/L > K
Up to 24

Insulin sensitizers (PPAR-? agonists)
With meals

Fluid retention, weight gain, congestive heart failure, edema, anemia. Due to the troglitazone experience, periodic monitoring of LFTs is recommended



Up to 24

Alpha-glucosidase inhibitors
Delay carbohydrate absorption
Just prior to meals

Gastrointestinal disturbances (abdominal pain, nausea, diarrhea),? LFT elevation

Local effect


Local effect

Non-sulfonylurea secretagogues
Insulin secretagogues
15 minutes prior to meals

Hypoglycemia, weight gain, hyperinsulinemia

L > K


K > L

K = kidney; L = liver; LFT = liver function test.

*The small fraction (<2%) of acarbose that is absorbed is eliminated by the kidneys.

advantages in elderly patients with impaired renal function (who are more vulnerable to hypoglycemia) but may be less effective in practice because of noncompliance with multiple dosing schedules. Conversely, longer-acting agents can be dosed once daily, enhancing compliance but increasing the risk of prolonged hypoglycemia. After the appropriate drug is chosen, treatment is initiated at low doses, with dose increases every 1 to 2 weeks until either treatment goals are met or “maximally effective” doses are reached. Note that for all sulfonylureas, efficacy plateaus at about 50% of the listed maximum dose; above these “maximally effective” doses, there is little clinical benefit derived from dose escalation, and alternative therapies should be considered.
The majority of type 2 patients initially respond to sulfonylureas with improved glycemic control. However, 10 to 20% of patients show little or no response; these cases are known as “primary” drug failures. Additionally, many other patients will experience the loss of drug effect after years of successful therapy; these “secondary” drug failures occur at rates of 5 to 10% per year, because of progressive ß-cell failure, drug tolerance, lack of enthusiasm for diet and exercise, and/or the superimposition of comorbid illness. Glucotoxicity itself can also contribute to worsening glucose control. In clinical practice, early signs of secondary drug failure should provoke renewed attempts to reinforce diet and exercise, as well as a reassessment of drug dosage. The re-appearance of hyperglycemia despite maximally effective drug doses signals the need to add another class of oral agent, or to institute insulin therapy. Overall, only approximately 25% of patients reach glucose targets with a sulfonylurea alone; stated another way, three in four patients will require additional modes of therapy.
Advantages of sulfonylureas include low cost (especially with generics), convenience (once-daily dosing), and the proven reduction of microvascular endpoints (retinopathy, nephropathy, and probably neuropathy) in the UKPDS. Disadvantages include hypoglycemia, weight gain, and the theoretical acceleration of so-called “ß-cell exhaustion.”
Non-sulfonylurea Secretagogues.
Repaglinide, a non-sulfonylurea that interacts with a different portion of the sulfonylurea receptor to stimulate insulin secretion, was approved by the FDA in 1998. A similar agent, nateglinide, was released 2 years later. The major advantage of the nonsulfonylurea secretagogues over sulfonylureas is their rapid and relatively short duration of action, which may attenuate postprandial glucose excursions and reduce the risk of fasting hypoglycemia. Both drugs require frequent daily dosing and should be taken 0 to 15 minutes before meals. Repaglinide and nateglinide exhibit similar or diminished glucose-lowering power compared with the sulfonylureas. Both agents have a favorable side effect profile and typically produce less clinical hypoglycemia than traditional sulfonylureas. The primary disadvantages of the nonsulfonylurea secretagogues are their higher cost and multiple dosing schedules.
Metformin is the only biguanide available for use in the United States. Unlike sulfonylureas, this agent is an “insulin

sensitizer,” which acts mainly to reduce hepatic glucose production by suppressing gluconeogenesis. Metformin may also augment peripheral glucose utilization, although this effect may be secondary to reversal of glucotoxicity. Metformin exhibits a similar glucose-lowering effect to the sulfonylureas, with expected hemoglobin A1C reductions of 1 to 2%. Metformin has a relatively short half-life (it is eliminated exclusively by the kidney) and is therefore given in two or three divided doses with meals. An extended-release metformin product, released in 2001, allows for more convenient daily dosing.
Because the effects of metformin are extrapancreatic, insulin levels generally fall, a potential advantage if the theory implicating hyperinsulinemia in the development of atherosclerosis proves correct. Other advantages of metformin include mild weight loss, mild (<10%) low-density lipoprotein and triglyceride reductions, and little induced hypoglycemia. Side effects are primarily gastrointestinal, including abdominal pain, bloating, nausea, diarrhea, and anorexia; these may be partially responsible for the weight loss effect. Metformin can also rarely produce lactic acidosis (approximately 0.03 cases per 1000 patient years) and should therefore not be given to patients with renal insufficiency (serum creatinine =1.5 in males or =1.4 in females), liver disease, congestive heart failure, metabolic acidosis, or a history of alcohol abuse. The drug should also be held in dehydrated patients, and for 48 to 72 hours prior to either surgery or the administration of intravenous radiocontrast agents.
With regard to evidence-based medicine, metformin has the most proven track record among the oral agents. Like sulfonylureas, metformin reduced microvascular end points in the UKPDS; unlike sulfonylureas, it may also have produced reductions in myocardial infarction, diabetes-related death, and overall mortality. Furthermore, in the recently released results of Diabetes Prevention Program (see below), metformin showed an ability to delay the progression to diabetes in patients with impaired glucose tolerance.
Thiazolidinediones (TZDs) reduce insulin resistance, most likely through activation of PPAR-? (peroxisome proliferator-activated receptor gamma), a nuclear receptor that regulates the transcription of several insulin-responsive genes that regulate carbohydrate and lipid metabolism. The biologic effect of TZDs is principally mediated by stimulation of peripheral glucose metabolism. PPAR-? activation also reduces lipolysis and enhances peripheral adipocyte differentiation, thereby redistributing fat stores from the liver, muscle and visceral depots to subcutaneous depots, an effect that likely contributes to the “insulin-sensitizing” effects of the TZDs. In 1997, troglitazone was the first TZD approved for use in the United States; although effective, the drug was withdrawn from the market in 1999 because of concerns over idiosyncratic hepatotoxicity. Two new TZDs, rosiglitazone and pioglitazone, were FDA-approved in 1999; these agents have negligible hepatotoxicity and are currently in widespread use.
Used as monotherapy, TZDs have slightly milder (and more slowly developing) glucose-lowering effects as compared with sulfonylureas and metformin, with expected hemoglobin A1C reductions of 1.0 to 1.5%. Clinical advantages of TZDs include convenience (once-daily dosing), little hypoglycemia, and reduced levels of circulating insulin. TZDs have many other beneficial effects, including (1) lower triglyceride levels (particularly with pioglitazone), (2) higher high-density lipoprotein levels, (3) reductions in small, dense low-density lipoprotein cholesterol, (4) small reductions in blood pressure, (5) improved endothelial function, and (6) enhanced fibrinolytic activity. Studies suggest that TZDs may also slow the growth of atherosclerotic plaque in carotid arteries; ongoing clinical trials are investigating their use in cardiovascular risk reduction. Finally, there is some evidence that TZDs may also slow the decline of ß-cell function, thus delaying the clinical progression from impaired glucose tolerance to overt diabetes mellitus.
Compared to other oral hypoglycemic agents, TZDs are more costly. Side effects of the TZDs are largely related to fluid retention and fat redistribution and include weight gain, edema, mild anemia, and worsening of congestive heart failure. These drugs are therefore not recommended for use in patients with moderate-to-severe congestive heart failure or those with severe anemia. As mentioned earlier, the two newer agents appear to be relatively free of hepatic toxicity; however, because of the troglitazone experience, they should not be used in patients with active liver disease or with elevated serum transaminases (ALT=2.5 times the upper limit of normal). The manufacturers of both rosiglitazone and pioglitazone currently recommend monitoring liver function tests every 2 months during the first year of therapy, with “periodic” testing thereafter. TZDs should be discontinued if transaminases are three or more times the upper limit of normal.
Acarbose and miglitol are competitive inhibitors of a-glucosidases, brush-border enzymes in the proximal small intestine that serve to break down complex carbohydrates into monosaccharides. These agents delay the absorption of carbohydrates such as starch, sucrose, and maltose but do not affect the absorption of glucose and other monosaccharides. To be effective, acarbose and miglitol must be taken at the beginning of each carbohydrate-containing meal, usually three to four times per day. Acarbose is minimally absorbed systemically, while miglitol is absorbed and rapidly excreted (unchanged) in the urine. Perhaps as a result of improved glycemic control, both of these agents are associated with modest (<10%) reductions in circulating triglyceride levels and have no appreciable effects on low-density or high-density lipoprotein cholesterol.
In controlled trials performed in patients with type 2 diabetes, a-glucosidase inhibitors reduced postprandial glucose excursions and produced small (0.5 to 1.0%) but meaningful reductions in hemoglobin A1C . The most common side effects associated with both acarbose and miglitol are abdominal pain, bloating, flatulence, and diarrhea; these adverse events can be minimized by initiating therapy at low doses and by using a slowly escalating dose titration schedule. Still, the manufacturers of both drugs discourage their use in patients with inflammatory bowel disease, colonic ulceration, or any other significant chronic gastrointestinal disorder.
Insulin is commonly used as first-line therapy for nonobese, younger, or severely hyperglycemic patients with type 2 diabetes and is often temporarily required during times of severe stress (e.g., injury, infection, surgery) or during pregnancy. Insulin should not be used as a first-line therapy for patients who are poorly compliant, unwilling to self-monitor glucose levels, or at high risk for hypoglycemia (e.g., the very elderly). In obese patients, profound insulin resistance often necessitates the use of large doses of insulin, which can interfere with efforts to restrict caloric intake and achieve weight loss. In leaner patients, and in patients with relatively mild fasting hyperglycemia (who continue to maintain endogenous insulin secretory capacity), relatively small doses of basal insulin (e.g., 0.3 to 0.4 U/kg of body weight per day) given once or twice per day may be sufficient to achieve glucose targets. Many of these patients retain some degree of meal-stimulated endogenous insulin secretion and may therefore require less rapid-acting insulin as well.
Although it is common practice to administer a single dose of intermediate-acting insulin in the morning, the glucose-lowering effect of this regimen does not usually extend over a full 24-hour period. Because a key element of successful insulin treatment is to counteract accelerated rates of endogenous glucose production in the morning, it is generally more effective to split the dose and administer sufficient amounts of intermediate-acting insulin in the evening (preferably at bedtime) to optimize control. Alternatively, a single dose of intermediate-acting insulin given at bedtime or of insulin glargine may be effective throughout the following day in patients who have retained the capacity to secrete insulin with meals. This approach has the advantage of greater simplicity and compliance and can be combined with oral glucose-lowering agents during the day to facilitate endogenous insulin release and action.
With regard to the initiation of insulin therapy, there are many acceptable approaches. As a first step, the total daily dose of insulin should be estimated from body weight; total insulin requirements typically range between 0.5 and 1.0 U/kg/day. One classic method for starting insulin is to divide the total daily dose unevenly, with two thirds given before breakfast and the remaining one third before dinner. Each of the two doses is then further subdivided: at breakfast, two thirds of the dose is given as intermediate-acting insulin and the other one third as a rapid-acting preparation, while at dinner, the dose is divided into two equal parts. As an example, for a 90 kg man with estimated requirements of 0.67 U/kg/day, 60 U of insulin may be required. Using the above method, this patient might receive 27 units of NPH with 13 units of regular insulin before breakfast, then 10 units of NPH with 10 units of regular insulin before dinner. Please note that this is only one of many “rule-of-thumb” methods for the initiation of insulin; with the advent of intensive therapy, such methods have become largely obsolete. Furthermore, in the absence of hyperglycemic symptoms, clinicians should generally begin with more conservative doses of insulin, to minimize hypoglycemia and to smooth the patient’s transition to subcutaneous insulin therapy.

In clinical practice, most insulin-treated patients are obese, have more severe hyperglycemia, and have already failed oral therapy. Such patients have higher degrees of both insulin deficiency and insulin resistance; as a result, they may require multiple-dose insulin regimens similar to those of type 1 patients. In these patients, it is best to distribute the insulin as evenly as possible throughout the day and to provide sufficient coverage overnight to control fasting hyperglycemia. The complexity of the regimen should be individualized according to the clinical context, the patient’s ability to perform self-care, and most importantly, the patient’s level of education and motivation.
In many cases, the combination of intensive insulin therapy with oral hypoglycemic agents (TZDs or metformin) may reduce insulin dose requirements and improve glycemic control. While growing in acceptance, the potential benefit of reducing circulating insulin levels (using combination therapy) on the development of atherogenesis remains to be established. Experience with the use of intensified insulin treatment, including continuous subcutaneous insulin infusion pumps and multiple subcutaneous injection regimens, is growing in patients with type 2 diabetes. Preliminary results suggest that intensified treatment may be successfully applied to many of these patients.
Treatment Strategies for Type 2 Diabetes
In contrast to type 1 diabetes, in which insulin therapy is required, several pharmacologic options exist for the management of type 2 diabetes. The pros and cons of the various oral hypoglycemic agents have already been discussed; often, it is difficult to justify the use of one oral agent over another. In the literature, many studies have compared the glucose-lowering power of one oral agent to another; however, few studies have compared the drugs in terms of relevant clinical outcomes such as mortality, cardiovascular disease, or microvascular complications. To date, the largest study to address such outcomes in type 2 diabetes was the UKPDS.
In the UKPDS, improved outcomes produced by intensified therapy were similar for patients given insulin, sulfonylureas, or metformin therapy. The ability of the study to detect differences among the various treatments was limited because of drug crossovers and because of the frequent need for drug combinations as the study progressed. The use of metformin in the UKPDS deserves specific mention here because of conflicting results. In the study, patients initially assigned to metformin therapy showed decreased rates of microvascular complications, combined diabetes-related end points, diabetes-related deaths, all-cause deaths, and myocardial infarction as compared with conventionally treated patients; in contrast, patients treated with insulin or sulfonylureas demonstrated reductions in only two of the five categories: microvascular complications and combined diabetes-related end points. Thus, metformin therapy appeared advantageous. Late in the study, however, 537 patients failing sulfonylurea therapy were randomly assigned to either continue the sulfonylurea alone or to add metformin. Compared to the sulfonylurea subgroup, this combined-therapy subgroup had an unexpected 60% increase in all-cause mortality. The results of this “substudy” have been called into question, since it was unblinded and lacked a placebo control. In addition, 25% of the patients assigned to continue monotherapy eventually required metformin to achieve glucose targets. In summary, based on the UKPDS, it is difficult to offer an unequivocal recommendation for metformin as compared to sulfonylureas or insulin therapy.
The choice of initial pharmacologic therapy for type 2 diabetes should be influenced mainly by the severity of fasting hyperglycemia, the degree of obesity, and the presence and magnitude of hyperglycemic symptoms. Other factors such as age, education, motivation, and comorbid conditions should also be considered. To determine the effectiveness of the therapy selected, drug regimens should be adjusted over a 3-month period based on glucose self-monitoring; failure to meet glucose targets within 3 months suggests the need for combination therapy ( Fig. 242-9 ). Published clinical trials comparing drug combinations to monotherapy have generally shown additive reductions in hemoglobin A1c ; with few exceptions, the magnitude of A1c reduction is similar to that achieved when the added agent is used as monotherapy. As is the case with monotherapy, there is no convincing evidence favoring one combination regimen over another, and most combinations have been approved by the FDA. “Triple-therapy,” or combining three agents to achieve glucose targets, is also used frequently in clinical practice (although not yet with FDA

Figure 242-9 Strategy for the treatment of type 2 diabetes.
approval) and appears to be effective. Ultimately, if glucose targets cannot be met by combining oral agents, insulin remains an effective treatment option.
Self-monitoring of blood glucose has revolutionized the management of diabetes. It actively involves patients in the treatment process, allows more rapid treatment adjustments, and reinforces dietary changes. Self-monitoring provides the patient with the tools necessary to assist in managing their disease and is especially useful during periods of stress and for patients who are susceptible to hypoglycemia. Urine glucose testing is unreliable and should not be used.
Newer glucose meters are small, portable, and reliable, give a digital readout, and have computerized memory to facilitate record keeping. Blood sampling is facilitated (and made less painful) by automated, spring-operated lancet devices; recent “off-the-finger” products also allow for more comfortable testing. Self-monitoring of blood glucose is of maximal value if the patient performs tests on a regular basis, can accurately measure glucose levels, and can make use of the results. The patient must become familiar with what a normal glucose value is, what the glucose targets are, and how levels can vary with changes in diet, activity, and insulin absorption. For most insulin-dependent patients, day-to-day adjustments in short-acting insulin based on pre-meal values and a “sliding scale” can be readily accomplished. These patients also need to examine the effects of their longer-acting insulin injections, and to make adjustments if glucose levels (e.g., pre-breakfast, pre-dinner, and bedtime values) are not within the target range. At a minimum, patients should be able to adjust to repetitive patterns of hypoglycemia or hyperglycemia, as well as to periods of stress and illness (“sick days”). For patients in the latter circumstance, urine testing for ketones should also be routinely performed.
The success of insulin therapy depends on the frequency with which the patient performs self-monitoring. Patients with type 1 diabetes should be encouraged to monitor before each meal and at bedtime and whenever symptoms occur. Periodic checks 90 to 120 minutes after meals help to control postprandial hyperglycemia, and patients should occasionally monitor pre-dawn (e.g., 3 AM) glucose levels to avoid nocturnal dips. Currently, no clear guidelines have been established regarding the frequency of blood glucose monitoring for type 2 diabetes. Type 2 patients who are treated with insulin should self-test daily, usually before breakfast, before dinner, and at bedtime. The frequency of blood glucose self-monitoring will depend largely on the stability of metabolic control; testing should be more frequent during the initiation of treatment, after changes in therapy, and during all times that altered metabolic control is suspected. Type 2 patients maintained on dietary therapy should, at the very least, learn self-monitoring of blood glucose to prevent metabolic decompensation. These patients also benefit from periodically monitoring glucose levels so that they may better appreciate how changes in their diet can adversely affect glycemic control.
Traditional self-monitoring of blood glucose is often inadequate to optimize metabolic control.

Continuous glucose monitoring reveals that tight glycemic control is often being achieved at the expense of unacceptably high rates of nocturnal hypoglycemia and also that postprandial glucose excursions are often larger than expected. To minimize these highs and lows, continuous glucose monitoring systems have recently been approved for clinical use. A currently available “real-time” glucose sensor is worn “Holter-style” over a 72-hour period, to alert clinicians to previously undetected nocturnal hypoglycemia and postprandial glucose elevations. Another system worn on the wrist uses iontophoresis to measure interstitial glucose levels at frequent intervals. It is hoped that armed with this additional information, patients and clinicians can then adjust the therapeutic regimen to minimize glucose excursions and maximize patient safety. Undoubtedly, continuous glucose monitoring methods as they become more reliable and convenient will eventually replace episodic self-monitoring, especially for patients with type 1 diabetes.
Glycohemoglobin (glycosylated hemoglobin) assays have emerged as the “gold standard” for long-term glycemic control. The test does not rely on a patient’s ability to self-monitor blood glucose levels and is not influenced by acute glycemic changes or by recent meals. Glycohemoglobin is formed when glucose reacts nonenzymatically with the hemoglobin A molecule; it is composed of several fractions, the largest being hemoglobin A1C . Hemoglobin A1C (expressed as the percentage of total hemoglobin) varies in proportion to the average level of glucose over the lifespan of the red blood cell, thereby providing an index of glycemic control during the preceding 6 to 12 weeks. Several assay methods have been developed that yield different ranges for nondiabetic control subjects; clinicians should therefore become familiar with the specific assays used for testing their patients.
Although ambient glucose levels are the dominant influence on glycohemoglobin levels, other factors can confound the interpretation of the test. For example, any condition that increases red blood cell turnover (e.g., pregnancy, hemolytic anemia) spuriously lowers glycohemoglobin levels, regardless of the assay used. Some assays yield spuriously low values in patients with hemoglobinopathies (e.g., sickle cell disease or trait, hemoglobin C or D), or high values when either hemoglobin F is increased (e.g., thalassemia, myeloproliferative disorders) or when large doses of aspirin are consumed. Thus, for unexpectedly high or low values encountered in clinical practice, factors that alter the specific assay should be excluded. In most cases, discrepancies between self-monitoring of blood glucose and glycohemoglobin results reflect problems with the former rather than the latter. Although glycohemoglobin provides the most accurate estimate of overall glycemic control, it has limited value in guiding specific changes in therapy; in clinical practice, frequent blood glucose measurements are essential to properly adjust the therapeutic regimen.
A management plan should take into consideration the life patterns, age, work and school schedules, psychosocial needs, educational level, and motivation of each individual patient. The plan should include lifestyle changes, a meal plan, medications, monitoring instructions (including “sick day” management), and education regarding the prevention and treatment of hypoglycemia. Importantly, all components of the plan must be both understood and accepted by the patient. Active patient participation in problem solving, as well as ongoing support from a health care team, is critical for the successful management of diabetes. At each visit, the management plan should be reviewed, and an assessment should be made of the patient’s progress in achieving glucose targets; if goals are not being met, causes need to be identified, and the plan should then be modified accordingly. At each visit, the history and physical examination should focus on early signs and symptoms of retinal, cardiovascular, neurologic, and podiatric complications, and on reinforcement of the diet and exercise prescription. A complete ophthalmologic examination, assessment of cardiovascular risk factors, and measurement of urinary albumin excretion (through either a timed collection or two “spot” urine albumin-to-creatinine ratios) should all be performed annually. Specialized podiatric care is also recommended for all patients with evidence of pedal neuropathy.
Formulation of individual glycemic goals must take into account the results of the DCCT (type 1 diabetes) and the UKPDS (type 2 diabetes) in the context of the patient’s capacity to implement the treatment plan, risk for hypoglycemia, and other factors that would alter the risk-benefit ratio. Table 242-8 presents target glycemic guidelines

Preprandial plasma glucose (mg/dL)
<80 or >140
2-hour postprandial glucose (mg/dL)
Bedtime plasma glucose (mg/dL)
<100 or >160
Hemoglobin AIc (%)
Low-density lipoprotein cholesterol (mg/dL)
High-density lipoprotein cholesterol (mg/dL)
>45 (men), >55 (women)
Fasting triglycerides (mg/dL)
Blood pressure (mg Hg)

*Clinical targets vary for individual patients, depending on assessment of overall health and risk-to-benefit ratio.
†Interventions may include dietary therapy, exercise prescription, and/or pharmacologic intervention.
‡Hemoglobin AIc control to <6.5% is advocated by some authorities, especially in type 2 diabetics.

for nonpregnant diabetic patients as well as targets for other clinical factors (e.g., blood pressure, lipids) that increase the potential for diabetic complications.
Intensive insulin therapy rarely (if ever) restores glucose homeostasis to levels achieved in nondiabetic individuals. As a result, the search for more effective methods of treatment remains a crucial long-term goal of diabetes research. Pancreas transplantation is promising in this regard; with growing experience in recent years, there have been substantial improvements in the outcome of pancreas transplant surgery. In major centers, 80 to 90% of patients emerge from the perioperative period with a functioning graft; once insulin independence is established, the majority of patients remain stable for several years. Successful pancreas transplantation improves the quality of life of patients with diabetes, primarily by eliminating the need for dietary restrictions, insulin injections, and frequent glucose self-monitoring. Although pancreas transplantation is only partially able to reverse long-term diabetic complications, it effectively eliminates acute complications such as hypoglycemia and diabetic ketoacidosis.
Unfortunately, because of the need for long-term immunosuppression, pancreas transplantation is at present an option for only a select group of patients, mainly for type 1 diabetics who will already require immunosuppression for a renal allograft. In such individuals, successful pancreas transplantation is also effective in preventing nephropathy in the grafted kidney. In the absence of indications for a kidney transplant, pancreas transplantation should generally be considered only in diabetic patients with a history of frequent, severe metabolic complications (e.g., hypoglycemia, ketoacidosis), in whom insulin therapy consistently fails to achieve metabolic control.
Pancreatic islet cell transplantation holds many potential advantages over the whole-gland transplant, since it is simpler to perform and less costly. Until recently, islet transplantation has had disappointing results with regard to long-term insulin independence; however, a recent series published out of Edmonton, Alberta, Canada, suggests that outcomes may be improving and that islet transplantation may become a therapeutic option. In this 12-patient series (median follow-up, 10 months), 4 patients had normal glucose tolerance, 5 had IGT, and 3 had a stable diabetes characterized by endogenous insulin production and a low risk of hypoglycemia. Interestingly, compared with diabetic patients, islet-cell transplantation has been easier in patients with chronic pancreatitis, many of whom have successfully undergone total pancreatectomy followed by intraportal injection of pancreatic islets. The implication here is that with diabetes, the use of immunosuppressive drugs, chronic low-grade rejection of the foreign islet grafts, and/or the activation of an autoimmune response may account for transplant failure. If these inferences are correct, the

future of islet transplantation therapy for diabetes depends mainly on manipulating the islet and/or the immune response and the availability of donor islets, rather than on technical surgical advances.
Prevention of Diabetes
As the pathogenesis of both types of diabetes becomes better understood, the potential for prevention of these diseases is more realistic. Two large, multicenter disease prevention trials have already been completed in the United States, and several more are planned.
In the Diabetes Prevention Program, more than 3000 overweight subjects with IGT were randomized into four treatment arms: (1) intensive lifestyle changes aimed at reducing body weight by 7% through a low-fat diet and 150 minutes of weekly exercise; (2) treatment with metformin, 850 mg twice per day; (3) treatment with placebo pills, twice per day; and (4) treatment with troglitazone, 400 mg once per day (this arm was discontinued because of concerns over liver toxicity). The latter three groups also received standard information regarding diet and exercise. On the advice of the Diabetes Prevention Program’s external data monitoring board, the trial was stopped a year early because of definitive results: 29% of patients in the placebo group developed diabetes during the average follow-up period of 3 years, compared with 22% of patients taking metformin and only 14% of patients undergoing intensive diet and exercise.[4] Put another way, patients taking metformin reduced their risk of diabetes by 31% versus standard care, whereas patients undergoing intensive lifestyle interventions reduced their risk by an impressive 58%. This suggests that patients with IGT (more than 20 million patients in the United States alone, according to recent estimates) can sharply lower their immediate risk of diabetes with intensive lifestyle changes (or in some cases with metformin), and puts the onus on clinicians to screen, identify, and appropriately treat patients with IGT. Postmenopausal estrogen and progestin can reduce the incidence of type 2 diabetes by 35%,[5] but the adverse effects of such therapy may outweigh its benefits ( Chapter 256 ).
Results of the Diabetes Prevention Trial Type 1 have also been recently published. In this study, “high-risk” relatives of type 1 diabetic subjects (based on antibody screening and HLA typing) were randomly assigned to no treatment or to low-dose insulin injections, a therapy used successfully in rodent models of spontaneous autoimmune diabetes to prevent disease expression. After 5 years of observation, nearly 60% of these “high-risk” patients developed diabetes, as predicted by clinical models; unfortunately, there was no difference in incidence between the insulin and no treatment arms.[6] Another substudy of the Diabetes Prevention Trial Type 1, testing the prevention of diabetes using oral insulin in patients at more moderate risk for disease, is still underway. Other putative preventive strategies are also under investigation.

Goldman: Cecil Textbook of Medicine, 22nd ed., Copyright © 2004 W. B. Saunders Company

The pathogenesis of the microvascular and neuropathic complications of diabetes is complex and poorly understood. Two well-researched mechanisms proposed for glucose-induced cell injury are advanced glycosylation end-products (AGEs) and an accelerated polyol pathway with consequent protein kinase C activation. These and other potential contributors are briefly discussed.
Proteins are readily glycosylated in vivo in direct proportion to prevailing levels of glucose. This nonenzymatic glycosylation is nonspecific, involving a wide range of proteins, including hemoglobin, collagen, laminin, low-density lipoproteins, and peripheral nerve proteins (tubulin). The consequent AGEs accumulate in a variety of tissues (including the kidneys and blood vessels) and are thought to contribute to cell injury through a variety of mechanisms, including stimulation of cytokines, complement activation, and upregulation of growth factor synthesis. AGEs also stimulate oxidative reactions, and their cross-linking capabilities render them resistant to natural degradation. In experimental diabetic animals, inhibition of AGE formation reduces tissue deposition of these end products and inhibits both the expansion of glomerular volume and urinary protein excretion.
In the polyol pathway, increased activity of intracellular aldose reductase leads to an accumulation of sorbitol and fructose, resulting in osmotic cell injury, decreased glutathione antioxidant activity (via decreased NAD+), and the enhanced formation of diacylglycerol. Diacylglycerol formation can in turn activate specific isoforms of protein kinase C, which stimulate transforming growth factor-ß release and play an important role in cell proliferation and vascular permeability. Beneficial effects of both aldose reductase inhibitors and specific protein kinase C inhibitors have been consistently demonstrated in animal models of diabetes. To date, their value in human subjects is uncertain.
Other potential mechanisms through which glucose could impair cell function include (but are not limited to) (1) formation of reactive oxygen species (hydrogen peroxide, superoxide), (2) activation of cytokines (angiotensin II, endothelin), (3) growth factor stimulation (transforming growth factor-ß, vascular endothelial growth factor), and (4) depletion of basement membrane glycosaminoglycans. Interventions directed at each of these mechanisms are currently under investigation.
Hemodynamic changes in the microcirculation may also contribute to microangiopathy. In the diabetic kidney, GFR is increased out of proportion to renal plasma flow, owing to an elevation in the transglomerular pressure gradient. It is assumed that raised glomerular pressures promote the passage of proteins and AGEs; with time, their accumulation in the mesangium could trigger the proliferation of mesangial cells and matrix production, eventually leading to glomerulosclerosis. Compensatory hyperfiltration would develop in less affected glomeruli, but even these would ultimately succumb because of progressive glomerular damage. Clinical studies support this view. Unilateral renal artery stenosis diminishes diabetic pathologic lesions in the affected kidney, and angiotensin converting enzyme (ACE) inhibitors (which reduce transglomerular pressure) are known to slow the progression of diabetic nephropathy. The diabetes-associated increase in microcirculatory hydrostatic pressure may also contribute to generalized capillary leakage of macromolecules in diabetic patients.
These theories would predict the benefits of optimal glycemic control reported by the DCCT in patients with mild or no complications. Whether similar benefits can be expected once severe damage has occurred is less clear. Extensive glycosylation of proteins with slow turnover rates would not be readily affected by correction of hyperglycemia. Moreover, the hemodynamic theory for nephropathy predicts that once glomerular injury causes compensatory hyperfiltration, progressive injury may continue in the remaining glomeruli, regardless of the prevailing metabolic state.
Diabetic retinopathy refers to progressive pathologic alterations in the retinal microvasculature, leading to areas of retinal nonperfusion, increased vascular permeability, and the pathologic proliferation of retinal vessels. In the United States, diabetes is the leading cause of blindness in persons aged 20 to 74 years. Retinopathy in patients with poorly controlled type 1 diabetes occurs in about 25% of patients 5 years after diagnosis, in 60% at 10 years, and in more than 95% at 15 years. Blindness occurs 25 times more frequently in diabetic patients than in control subjects and is seen most often after the disease has been present for at least 15 years, in the setting of advanced retinopathy. Approximately 10 to 15% of type 1 diabetic patients will become legally blind (visual acuity of 20/200 or worse in the better eye). In type 2 diabetes, though the incidence of blindness is lower, higher disease prevalence results in an even larger number of patients affected with severe visual loss.

The earliest pathologic changes associated with retinopathy are termed mild nonproliferative diabetic retinopathy (mild NPDR). In type 1 patients, these changes generally begin 3 to 5 years after diagnosis. The first signs of mild NPDR are microaneurysms, which arise most often in areas of capillary occlusion. Subsequently, increasing vascular permeability leads to retinal blot hemorrhages (round, with blurred edges) and “hard” exudates (sharply defined and yellow). Infarctions of the nerve fiber layer, known as “soft” exudates or “cotton-wool spots,” appear as white or gray, rounded swellings. At this early stage of retinopathy, visual acuity is generally unaffected, and the risk of progression to high-risk proliferative diabetic retinopathy (PDR) (see later) is about 15% at 5 years. Moderate NPDR is characterized by intraretinal microvascular abnormalities, including venous caliber changes, beading, and increased capillary dilatation and permeability. Later changes, termed severe or very severe NPDR, include progressive retinal capillary loss and ischemia, with further development of extensive hemorrhages, exudates, and microaneurysms. At 5 years, moderate and severe NPDR are associated with a 30% and 60% risk of progression to high-risk PDR, respectively.
Proliferative diabetic retinopathy involves neovascularization, the growth of fine tufts of new blood vessels and fibrous tissue from the inner retinal surface or the optic head. Early proliferative changes are confined to the retina, but later invasion of the vitreous body constitutes high-risk PDR; during this end stage, fibrosis and contracture of the neovasculature results in retinal detachment and hemorrhage, the most important determinants of blindness. Occasionally, new vessels can invade the iris and anterior chamber, leading to sight-threatening closed-angle glaucoma.
Clinically significant macular edema (CSME) results from vascular leakage at the macula and can occur either with or without the stages of retinopathy described earlier. CSME is suggested by hard macular exudates on fundoscopic examination and can be confirmed with slit lamp biomicroscopy. In general, maculopathy is more common in type 2 patients, in whom it is an important contributor to the loss of visual acuity. As will be discussed, the treatment of CSME runs parallel to the treatment of other forms of diabetic retinopathy.

At present, medical management of diabetic retinopathy is aimed at controlling risk factors for progression. The value of tight glycemic control was proven by the DCCT, whose primary prevention arm demonstrated an impressive 76% risk reduction for the onset of retinopathy with intensive therapy. In the secondary prevention arm, patients with early NPDR undergoing intensive therapy demonstrated a 47% risk reduction in the development of severe NPDR or PDR, a 51% risk reduction in the need for laser treatment, and a 26% risk reduction in the development of CSME. Other targets for medical management, all associated with accelerated retinal damage, include (1) hypertension, (2) hyperlipidemia, (3) treatment of nephropathy, and (4) careful follow-up during pregnancy, where accelerated retinal pathology has been linked to preexisting diabetes (but not gestational disease).
Surgical management of retinopathy is aimed at slowing disease progression, as baseline visual acuity is difficult to recover. In the

No to minimal NPDR
Not recommended
12 months†
Mild to moderate NPDR
Not recommended
6–12 months†
Severe to very severe NPDR
2–4 months
Early PDR
2–4 months
High-risk PDR
2–4 months
CSME = clinically significant macular edema; NPDR = nonproliferative diabetic retinopathy; PC = photocoagulation; PDR = proliferative diabetic retinopathy.
Adapted from Aiello LP, et al: Diabetic retinopathy (Technical Review). Diabetes Care 1998;21:143–156.

*If retinopathy and CSME coexist, focal PC for CSME should always precede panretinal PC.
†In these patients, follow-up is recommended in just 2–4 months if CSME is also present.

1980s, large-scale prospective clinical trials such as the Diabetic Retinopathy Study and the Early Treatment Diabetic Retinopathy Study established photocoagulation as the treatment of choice when retinopathy threatens vision. Most patients with PDR, and selected patients with severe NPDR, are now treated primarily with scatter (panretinal) photocoagulation; cryotherapy or vitrectomy may be required if laser treatment is unfeasible for technical reasons or because of extensive disease. CSME is near-universally treated with focal photocoagulation, with the possible exception of patients exhibiting no or minimal NPDR. In such patients, close follow-up at 2- to 4-month intervals is an acceptable option. A treatment chart, adapted from a thorough technical review by Aiello et al, is shown in Table 242-10 . Note that the decision to treat depends not only on stage of retinopathy and extent of CSME but also on general medical status, compliance with follow-up, and status of the contralateral eye.
These considerations make it imperative for physicians to prospectively identify diabetic patients at risk for retinopathy and visual loss. Nonspecialists, including house officers, internists, and diabetologists, are known to have difficulty diagnosing the stages of retinopathy; studies show that such physicians arrive at the correct diagnosis in fewer than half of cases. Accordingly, patients should be referred to an experienced ophthalmologist for a complete examination, to include a dilated fundoscopic examination, tonometry, and slit lamp biomicroscopy. The most recent ADA position statement recommends initial eye examination within 3 to 5 years of diagnosis of type 1 diabetes, and at the time of diagnosis in type 2 patients. Two special circumstances deserve a footnote here: (1) since children rarely develop retinopathy before puberty, early-onset type 1 patients generally do not require screening before 10 years of age, and (2) the acceleration of retinopathy during pregnancy demands that all patients with preexisting diabetes be examined during the first trimester. Follow-up of all patients should occur at least on a yearly basis, with the possible exception of retinopathy-free type 2 diabetics. Even in the latter cases, the ADA does recommend yearly examinations to avoid lost follow-up and to identify patients with more aggressive ocular disease.

Felices Vacaciones de Verano

La Fe no es nada sino va acompañada de acciones.

Fragmento de Cuento para antes de Dormir:

Y palabras de esa muchacha que habían estallado en el alma del hombre como bengalas de fiesta, y una noche de abrumadora belleza, los dos sentados en las escaleras de una hostería, cuando todavía era tiempo de promesas y en alguna zona del cielo podían descubrir el milagro de una estrella fugaz.

1989 by Marcelo David Caruso



Introduce tus datos o haz clic en un icono para iniciar sesión:

Logo de

Estás comentando usando tu cuenta de Cerrar sesión / Cambiar )

Imagen de Twitter

Estás comentando usando tu cuenta de Twitter. Cerrar sesión / Cambiar )

Foto de Facebook

Estás comentando usando tu cuenta de Facebook. Cerrar sesión / Cambiar )

Google+ photo

Estás comentando usando tu cuenta de Google+. Cerrar sesión / Cambiar )

Conectando a %s


A %d blogueros les gusta esto: