Diabetes Mellitus

From WiserWiki

[edit] Diabetes Mellitus

Jay S. Skyler

Irl B. Hirsch

Diabetes mellitus is a group of metabolic diseases characterized by hyperglycemia resulting from defects in insulin secretion, insulin action, or both. As a function of time, the chronic hyperglycemia of diabetes is associated with long-term damage, dysfunction, and failure of various organs, especially the eyes, kidneys, nerves, heart, and blood vessels.

The vast majority of cases of diabetes fall into two broad categories. In one category (type 1 diabetes), the cause is an absolute deficiency of insulin secretion, a result of autoimmune destruction of the pancreatic islet β-cells. In the other, much more prevalent category (type 2 diabetes), the cause is a combination of resistance to insulin action and an inadequate compensatory insulin secretory response.

[edit] EPIDEMIOLOGY AND PUBLIC HEALTH ASPECTS

In the United States today, diabetes mellitus is a public health nightmare.^[1] Consider the following:

• Of the estimated 15.6 million people nationwide who have diabetes (approximately 1 in 16 people), a projected 5.4 million people are unaware of it.
  
• Nearly 800,000 Americans develop diabetes every year, or approximately 2,200 every day, but individuals may have diabetes and remain undiagnosed for an average of 5 to 10 years.
  
• Each year 182,000 deaths are linked to diabetes, making it the third largest killer in the country, with 57,000 of those deaths directly attributable to diabetes.
  
• Diabetes has the highest direct costs for health care of any disease category (estimated by the National Institutes of Health [NIH] to be $91.1 billion in 1995), and is responsible for one in every seven health care dollars spent in the country.
  
• Total medical costs for diabetic patients are staggering, per capita almost fourfold that of people without diabetes.
  
• Diabetic retinopathy is the leading cause of blindness in working age adults—24,000 cases of legal blindness every year—but an estimated 90% of lost vision is preventable.
  
• Diabetic nephropathy is the leading cause of end-stage renal disease (ESRD)—42% of all cases—but an estimated 90% of future ESRD is preventable.
  
• Diabetes is the leading cause of nontraumatic amputations—67,000 limbs lost per year, a rate of amputation fifteenfold to fortyfold greater than that in the nondiabetic population—but an estimated 85% of limb loss is preventable.
  
• Diabetes results in a twofold to sixfold increased risk of heart disease and a twofold to fourfold increased risk of stroke.
  
• Risk can be dramatically reduced by careful glucose control, aiming for near-normal levels of hemoglobin A_1c (HbA_1c), but national surveys demonstrate that only 12% of patients achieve that, with the average HbA_1c among diabetic patients 9.1% (normal 3.9% to 6%); even worse, the majority of patients do not have HbA_1c measurements performed at all.
  

[edit] PATHOPHYSIOLOGY

The brain and nervous system are obligate users of glucose for fuel metabolism (Fig. 96-1). They do so independent of insulin. In contrast, insulin-stimulated glucose uptake occurs in most peripheral tissues (e.g., muscle, adipose), with muscle consuming the largest portion of postprandial glucose (see Fig. 96-1). Glucose entry into the circulation is from the gastrointestinal system as a consequence of food consumption and metabolism of carbohydrates to monosaccharides. In nondiabetic individuals, hepatic glucose production (HGP) matches brain and nervous system glucose utilization in the basal state, almost on a precise quantitative basis. Basal insulin secretion modulates HGP, thus regulating glycemia (see Fig. 96-1).

Figure 96-1 Scheme of regulation of blood glucose. Glucose input is from food intake via the gastrointestinal system, or during the basal state from hepatic glucose production, which is modulated by basal insulin secretion. The brain and nervous tissue use glucose independent of insulin, while insulin stimulates glucose uptake and utilization by peripheral tissues (here represented by muscle and adipose tissue).

Hyperglycemia emerges when there is relative deficiency of insulin action at target tissues—muscle and peripheral tissues in the postprandial state, liver in the basal state. Thus hyperglycemia occurs either when circulating levels of insulin are low or when cellular sensitivity to insulin is impaired. Without insulin to stimulate glucose transport into cells, blood glucose concentrations increase and symptoms of diabetes, including polyuria, polydipsia, polyphagia, and fatigue, develop. With milder degrees of insulin deficiency, the abnormalities may be manifest only during the fed state when plasma insulin levels are normally high. With more severe degrees of insulin deficiency, the effects that insulin normally has in the fasted state (i.e., modulation of hepatic glucose production and inhibition of catabolism) are also impaired, and thus the metabolic derangement is more severe (e.g., proteolysis, lipolysis, weight loss). The most extreme degree of impairment results in ketogenesis and ketoacidosis.

Diabetic ketoacidosis (DKA) develops when excess free fatty acids taken up by the liver are preferentially shunted toward the formation of acetoacetate and β-hydroxybutyrate. As the concentration of ketone bodies increases, serum buffering capacity is exceeded and metabolic acidosis occurs. Hyperglycemia induces an osmotic diuresis that eventually leads to dehydration and volume depletion.

[edit] Type 1 Diabetes

Type 1 diabetes mellitus arises as a consequence of selective destruction of the insulin-producing β-cells in the pancreatic islets of Langerhans.^[2] Thus there is absolute insulinopenia, and the dependence on exogenous insulin therapy for survival. Because of the absolute insulinopenia, patients with type 1 diabetes are prone to ketosis (and possible ketoacidosis) even under basal conditions. Type 1 diabetes generally has its onset in the first 2 decades of life (thus the previous name juvenile-onset diabetes), but may occur at any age. In the United States, among Caucasians, thedisease typically affects 1 in 300 persons in the population. Relatives of type 1 patients have a higher risk (approximately 3% to 6%), the highest risk being in monozygotic twins (30% to 50%).

Multiple loci have demonstrated associations or linkages to type 1 diabetes. Both human leukocyte antigen (HLA) and non-HLA genes contribute to diabetes susceptibility. The best characterized susceptibility loci are IDDM1 (HLA) and IDDM2 (INS-VNTR), while several other loci linked to diabetes await further characterization. IDDM1 appears to confer ˜50% of the genetic susceptibility to type 1 diabetes. Interestingly, genetic protection from type 1 diabetes is associated with specific alleles at the IDDM1 (DQB1*0602) and IDDM2 (INS-VNTR class III) loci.

The current concept is that islet β-cells in genetically susceptible individuals are destroyed by an autoimmune response mediated by T lymphocytes (T cells) that react specifically to one or more β-cell proteins (autoantigens). The disease process is insidious, evolving over a period of years. During this time, a number of immune markers appear that indicate the presence of ongoing β-cell damage (e.g., immunofluorescent islet cell autoantibodies [ICA], insulin autoantibodies [IAA], and autoantibodies to glutamic acid decarboxylase [GAD65] and to a transmembrane tyrosine phosphatase [ICA512]). This is accompanied by a progressive decline of β-cell function (loss of first-phase insulin response (FPIR) during an intravenous glucose tolerance test (IVGTT). Ultimately, the clinical syndrome of type 1 diabetes becomes evident when a majority of β-cells have been destroyed and hyperglycemia supervenes.

The generally accepted sequence, genetic susceptibility → environmental trigger → immunologically mediated pancreatic islet β-cell destruction, however, is not straightforward. Rather, the pathogenesis of type 1 diabetes appears to involve a disruption of balance between forces propelling the progression of disease and forces retarding or preventing that progression. Diabetes appearance is influenced by the net effects of genetic and environmental factors on immunoregulatory responses. Thus there are susceptibility genes and protective genes; environmental triggers and environmental factors associated with protection; and autoreactive immune/inflammatory processes leading to insulitis and immunoregulatory mechanisms restraining the destructive processes. The predominant regulatory cells appear to be various subsets of helper T cells, acting via the cytokines they produce. Thus β-cell destruction is enhanced by the T-helper-1 (Th1) subset of CD4⁺ T cells and the type 1 cytokines (interleukin-2 [IL-2] interferon-γ [IFN-γ], tumor necrosis factor-β [TNF-β]) they produce. In contrast, there is inhibition of β-cell destruction by the T-helper-2 (Th2) and T-helper-3 (Th3) subsets of CD4⁺ T cells and the type 2 cytokines (IL-4, IL-5, IL-10) and type 3 cytokines (transforming growth factor-β [TGF-β]) they respectively produce. The cytokines produced by Th1-cell activation activate lymphocytes and macrophages to kill islet β-cells by a variety of mechanisms. Nonspecific immune/inflammatory killing mechanisms involved in islet β-cell destruction are mediated by molecules released from activated T cells (both CD4⁺ and CD8⁺ T cells) and macrophages. These include destructive cytokines (IL-1, TNF-α, TNF-β, IFN-γ), oxygen free radicals (O₂^·−, H₂O₂, OH^·), nitric oxide (NO^·), and peroxynitrite (ONOO⁻). These reactions may be counterbalanced by endogenous or exogenous protective mechanisms (e.g., antioxidants, nicotinamide, or superoxide dismutase activity). Antigen-specific CD8⁺ cytotoxic-T cells (that interact with a β-cell autoantigen-MHC class I complex) may kill β-cells by receptor (Fas/FasL)-mediated mechanisms and/or by secretion of cytotoxic molecules (granzymes and perforin). Once the initial immune destruction commences, secondary and tertiary immune responses also are activated, with virtually the whole immunologic army attacking β-cells.

Enhancement of protective forces has been demonstrated in animal models of diabetes. For example, diabetes development can be prevented by administering β-cell autoantigens—insulin or GAD—by the parenteral, oral, intranasal, or aerosol inhalation routes. The protective effects of these treatments have been attributed to activation of T cells that produced one or more suppressor cytokines (IL-4, IL-10, and TGF-β). Based on this, plus pilot studies in high-risk relatives of type 1 diabetic patients, a full-scale randomized, controlled clinical trial, the Diabetes Prevention Trial of Type 1 Diabetes (DPT-1), is being conducted to determine whether intervention with insulin (parenteral or oral) can delay the appearance of overt clinical diabetes. Also underway is another randomized, prospective, controlled clinical trial, the European Nicotinamide Diabetes Intervention Trial (ENDIT), to evaluate the effects of nicotinamide in high-risk relatives of individuals with type 1 diabetes. The concept is that nicotinamide may work by restoring β-cell content of nicotinamide adenine dinucleotide (NAD) toward normal (via inhibition of poly-ADP-ribose polymerase), by scavenging free radicals, or by inhibiting cytokine-induced islet nitric oxide production.

[edit] Type 2 Diabetes

Type 2 diabetes mellitus is a heterogeneous disorder characterized by impaired β-cell function (defective insulin secretion) and by diminished tissue (liver, muscle, adipose) sensitivity to insulin (insulin resistance).^[3] Insulin secretion is not absent, so that these individuals are not dependent on exogenous insulin for survival, and are not prone to ketosis. It is the more common form of the disease, accounting for over 90% of all cases. There is a higher risk among relatives of type 2 patients, with the highest risk being in monozygotic twins (80% to 90%). Thus there appears to be an important genetic predisposition. However, central obesity, a sedentary lifestyle, and the sarcopenia of aging are all contributing factors. Type 2 diabetes generally has its onset after age 40 (thus the previous name adult-onset diabetes), but may occur at any age. Indeed, there is a growing epidemic of type 2 diabetes in adolescents, perhaps related to the epidemic of obesity in that age group, accelerated by a progressively sedentary lifestyle.

Multiple genetic loci have demonstrated associations or linkages to type 2 diabetes. These associations and linkages are not observed in all populations studied. The most important locus found thus far is on chromosome 2(NIDDM1). In the study group of Mexican-American sibling pairs with type 2 diabetes from Starr County, Texas, NIDDM1 accounted for ˜30% of genetic susceptibility to type 2 diabetes.

There has been considerable debate as to which of the two defects, impaired insulin secretion or insulin resistance, is the initial lesion in the pathogenesis of type 2 diabetes. For individuals with significant fasting hyperglycemia (plasma glucose > 180 to 200 mg/dl [>10 to 11.1 mmol/L]), it is clear that both defects are present. Furthermore, defects in insulin secretion can lead to insulin resistance and vice versa. Thus, in a given individual it is impossible to determine if the initial defect was the impairment of insulin secretion or the impairment of insulin action.

Most individuals with genetic defects in insulin secretion do not develop hyperglycemia (glucose intolerance or diabetes) unless there is superimposed insulin resistance (e.g., pregnancy, weight gain, physical inactivity) or the simultaneous presence of diabetogenic genes predisposing to insulin resistance. Likewise, insulin resistance alone does not result in hyperglycemia unless there is an impairment in the ability of the β-cell to compensate for the insulin resistance. Although patients with type 2 diabetes may have insulin levels that seem normal or elevated, the higher blood glucose levels in these diabetic patients would be expected to result in even higher insulin values had their β-cell function been normal. Thus insulin secretion is defective in these patients and insufficient to compensate for the insulin resistance.

The impairment in β-cell function is manifest in several ways. There is an attenuated sensitivity of insulin response to glucose, or “blindness” to hyperglycemia. This blindness is selective for glucose, and the insulin response to other stimuli is not impaired. In addition, there is blunted or absent first-phase insulin response to glucose, so that insulin secretion is delayed and fails to restore prandial glycemic excursions in a timely manner. Insulin response is normally biphasic. The first phase is the critical determinant of the magnitude of hyperglycemia following carbohydrate intake. Its decrease results in an overall delayed insulin secretory response. Although variable, in mild and moderate type 2 diabetes, second-phase insulin response is sufficient to restore prandial plasma glucose excursions to basal levels. With more significant hyperglycemia, both the first phase and second phase of insulin release are impaired, and there is decreased overall insulin secretory capacity.

Decreased insulin action (insulin resistance) has been demonstrated in type 2 diabetes, in impaired glucose tolerance (IGT), and in first-degree relatives of individuals with type 2 diabetes despite having normal glucose tolerance. It is also seen in obesity, essential hypertension, and acromegaly, or with use of glucocorticoids or estrogens.

Insulin resistance occurs at several sites. Normally, insulin binds to its receptor on a target cell, initiating a signaling cascade that eventually results in the effects of insulin, including glucose entry into the cell. Abnormalities before insulin's interaction with the cell are termed prereceptor defects. These include binding of insulin by antibodies, increased degradation of insulin, and molecular abnormalities in the structure of the insulin molecule that make it difficult to bind to its receptor. These causes of insulin resistance are rare. At the receptor level, insulin resistance may be caused by decreased binding of insulin to the insulin receptor due to decreased number or affinity of receptors for insulin, or by decreased ability of the receptor to generate a transmembrane signal for example, by decreased receptor tyrosine kinase activity for both autophosphorylation and phosphorylation of substrates (e.g., insulin receptor substrate-1 [IRS-1]). Beyond the level of insulin interaction with its receptor, postreceptor defects may occur. There may be an impairment in any number of the biologic effects of insulin (e.g., translocation of the glucose transporter [GLUT4] to the cell membrane).

Hyperglycemia itself may aggravate the impairments both in insulin secretion and in insulin action. The components of defective insulin secretion and action that are reversible by lessening of hyperglycemia are known as glucose toxicity.

There is a progressive loss of β-cell function during the course of type 2 diabetes. This may be related to amyloid deposits that are seen in the islets of most people with type 2 diabetes. The material consists of fibrils formed by a 37–amino acid peptide called islet amyloid polypeptide (IAPP) or amylin. Within the β-cell, a precursor is cleaved to form IAPP, which is contained within secretory granules and cosecreted with insulin. IAPP accumulation within the β-cells may contribute to their destruction via some form of direct cytotoxicity.

[edit] Other Types

There are a number of other specific types of diabetes.^[4] One is a group of disorders associated with monogenetic defects in β-cell function. These are frequently characterized by onset of mild hyperglycemia at an early age (generally before age 25 years) and were formerly referred to as maturity-onset diabetes of youth (MODY). They are characterized by impaired insulin secretion with minimal or no defects in insulin action. They are inherited in an autosomal dominant pattern. Abnormalities at three genetic loci on different chromosomes have been identified to date: MODY1, mutations on chromosome 20q in a hepatic transcription factor (HNF-4α) gene region; MODY2, mutations in the glucokinase gene on chromosome 7p; and MODY3, mutations on chromosome 12 in a hepatic transcription factor (HNF-1α) gene region. Point mutations in mitochondrial DNA are associated with diabetes mellitus and deafness. Genetic abnormalities that result in the inability to convert proinsulin to insulin have been described. There are rare genetic defects in insulin action, previously termed type A insulin resistance. Leprechaunism and the Rabson-Mendenhall syndrome are two syndromes that have mutations in the insulin receptor gene. In addition, there are a number of genetic syndromes sometimes associated with diabetes.

Secondary diabetes can occur as a consequence of diseases of the exocrine pancreas or endocrinopathies. The pancreatic diseases may include pancreatitis, trauma, infection, pancreatectomy, pancreatic carcinoma, cystic fibrosis and hemochromatosis. Fibrocalculous pancreatopathy may be accompanied by abdominal pain radiating to the back and pancreatic calcifications on x-ray. Pancreatic fibrosis and calcium stones in the exocrine ducts have been found at autopsy. Endocrinopathies that may cause diabetes include acromegaly, Cushing's syndrome, glucagonoma, and pheochromocytoma, particularly in individuals with preexisting defects in insulin secretion. Somatostatinoma-and aldosteronoma- induced hypokalemia can cause diabetes by inhibiting insulin secretion.

Many drugs can impair insulin secretion (e.g., thiazide diuretics, β-blockers), and thereby precipitate diabetes in individuals with insulin resistance. There are also many drugs and hormones that can impair insulin action (e.g., nicotinic acid and glucocorticoids).

Atypical diabetes mellitus (ADM) is a form of diabetes that may be seen in African-Americans. In ADM, there is acute onset of hyperglycemia, with ketosis or ketoacidosis as a common feature, followed by a clinical course more characteristic of type 2 diabetes. It appears to be inherited as an autosomal dominant disorder, often being seen in two to three consecutive generations. In one of 10 families with ADM, a unique glucokinase mutation was identified.

Gestational diabetes mellitus (GDM) is defined as any degree of glucose intolerance with onset or first recognition during pregnancy. The definition applies regardless of whether insulin or only diet modification is used for treatment or whether the condition persists after pregnancy. In the majority of cases of GDM, glucose regulation will return to normal after delivery.

[edit] DIAGNOSIS

The diagnosis of diabetes mellitus is established by the demonstration of hyperglycemia (Table 96-1). The diagnostic criteria for diabetes mellitus were changed in 1997.^[4] The previous criteria were based mainly on the use of the oral glucose tolerance test (OGTT), which was sufficiently inconvenient not to be used in clinical practice. As a consequence, the default criterion for diagnosis became the fasting glucose alone, which was pegged at a plasma glucose level of 140 mg/dl (7.8 mmol/L). There were at least two problems with this cut point. One, it simply was too high, based on retinopathy risk. Two, it did not correspond to the OGTT level. This left undiagnosed patients with disease sufficient to lead to complications. Consequently, the American Diabetes Association (ADA) commissioned an expert committee to examine the available data and make recommendations that might allow individuals with diabetes to be more easily diagnosed in clinical practice. The expert committee analysis found that by lowering the fasting plasma glucose (FPG) cut point to ≥126 mg/dl (7 mmol/L), most people with undiagnosed diabetes would become recognized, without very much risk of false-positive diagnosis. Moreover, those at risk of retinopathy would be detected earlier. Thus 126 mg/dl (7 mmol/L) is a surrogate for an OGTT 2-hour value of 200 mg/dl (11.1 mmol/L). This change does not increase the number of people with diabetes. Rather, it increases the number of people with known diabetes.

Table 96-1 Categories of Glycemia

The revised criteria include three ways to diagnose diabetes, each of which must be confirmed on a subsequent day by any one of the three methods. The criteria are (1) unequivocal symptoms (polyuria, polydipsia, unexplained weight loss) and casual (any time of day regardless of time since last meal) plasma glucose ≥200 mg/dl (11.1 mmol/L); (2) fasting (no caloric intake for at least 8 hours) plasma glucose ≥126 mg/dl (7 mmol/L); and (3) 2-hour plasma glucose ≥200 mg/dl (11.1 mmol/L) during an OGTT using a 75-gram oral glucose load.

The old criteria used a FPG of <115 mg/dl (6.4 mmol/L) for normal. In contrast, the new criteria use a FPG of <110 mg/dl (6.1 mmol/L) for normal. Individuals having FPG levels 110 to 125 mg/dl (6.1 to 6.9 mmol/L) are now defined as having impaired fasting glucose (IFG) (Table 97-1), and are at increased risk of diabetes, similar to those with IGT, who have OGTT 2-hour values of 140 to 199 mg/dl (7.8 to 11 mmol/L).

HbA_1c measurement is not currently recommended for diagnosis of diabetes, although some studies have shown that the frequency distributions for HbA_1c have characteristics similar to those of the FPG and the 2-hour PG. However, both HbA_1c and FPG (in type 2 diabetes) have become the measurements of choice in monitoring the treatment of diabetes, and decisions on when and how to implement therapy are often made on the basis of HbA_1c. The revised criteria are for diagnosis and are not treatment criteria or goals of therapy.

Screening is important because meticulous glycemic control slows the course of development of diabetic complications. Thus prolongation of normoglycemia should reduce risk of complications. Studies suggest that in the earlier stages (IFG and IGT) interventions, such as diet and exercise, may forestall the evolution of type 2 diabetes. Screening for type 2 diabetes is simple, only a FPG is needed. The OGTT is no longer the primary screening tool. Screening and early diagnosis of type 2 diabetes should be highly cost-effective. All adults over age 45 should be screened every 3 years. All individuals at higher risk (based on obesity, ethnicity, family history, previous gestational diabetes) should be screened annually, starting at an earlier age.

[edit] PATIENT EVALUATION

A comprehensive medical history can uncover symptoms that may help establish the diagnosis in a patient with previously unrecognized diabetes. It provides information essential for providing high-quality care. The hallmark symptoms of diabetes mellitus are polyuria, polydipsia, and polyphagia. However, many individuals with type 2 diabetes present only with fatigue, which may be dismissed as related to aging. Patients with type 2 diabetes may have only a gradual onset of frequent urination that prevents recognition of symptoms. Many attribute their polyuria to their polydipsia and do not seek medical attention. Nocturia is a particularly helpful indicator of polyuria and the onset of significant hyperglycemia; most patients remember how many times they get up during the night. The association of thirst or dry mouth with frequent nocturia is more consistent with elevated nighttime blood glucose than with benign prostatic hyperplasia or other urogenital conditions. Visual changes such as transient blurriness are very suggestive of fluctuating hyperglycemia.

In patients with known diabetes, the history should review previous treatment, evaluate glycemic control, seek evidence of complications, assist in formulating a management plan, and provide a basis for continuing care. It should focus on the following 15 elements:

• Details of previous treatment programs, including nutrition and diabetes self-management education
  
• Current treatment program, including medications, meal plan, and results of glucose monitoring and patients' use of the data
  
• Prior HbA_1c records
  
• Nutritional status, weight history, and eating patterns
  
• Physical activity and exercise pattern
  
• Lifestyle, cultural, and psychosocial factors that might influence diabetes management
  
• Growth and development in children and adolescents
  
• History of acute complications (e.g., ketoacidosis and hypoglycemia)
  
• Prior or current infections
  
• Symptoms and treatment of chronic complications (eye; kidney; nerve; genitourinary, including sexual; bladder; gastrointestinal; cardiovascular; cerebrovascular; peripheral vascular; foot)
  
• Other medications
  
• Risk factors for atherosclerosis (smoking, hypertension, obesity, dyslipidemia, family history)
  
• History and treatment of other conditions
  
• Family history of diabetes
  
• Gestational history (hyperglycemia, delivery of an infant weighing >9 lb, toxemia, stillbirth, polyhydramnios, or other complications of pregnancy)
  

The physical examination in people with diabetes should note the following:

• Height and weight measurement (and comparison with norms in children and adolescents)
  
• Blood pressure determination (with orthostatic measurements when indicated)
  
• Ophthalmoscopic examination (preferably with dilation)
  
• Cardiac examination
  
• Abdominal examination (for hepatomegaly)
  
• Evaluation of pulses
  
• Foot examination
  
• Skin examination (including insulin-injection sites)
  
• Neurologic examination
  

Laboratory evaluation focuses on determination of the degree of glycemic control and the identification of complications and risk factors. These include FPG; HbA_1c; fasting lipid profile (total cholesterol, high-density lipoprotein [HDL] cholesterol, triglycerides, and low-density lipoprotein [LDL] cholesterol); serum creatinine; urinalysis (glucose, ketones, protein, and sediment) test for microalbuminuria (e.g., timed specimen or the albumin-to-creatinine ratio) in pubertal and postpubertal type 1 patients who have had diabetes for at least 5 years and in all patients with type 2 diabetes); urine culture if sediment is abnormal or symptoms are present; thyroid function tests when indicated; and electrocardiogram.

Comprehensive dilated eye and visual examinations should be performed annually by an ophthalmologist or optometrist who is knowledgeable and experienced in the management of diabetic retinopathy for all postpubertal patients who have had diabetes for 3 to 5 years, all patients diagnosed after age 30, and any patient with visual symptoms and/or abnormalities.

All individuals with diabetes should receive a thorough foot examination at least once a year to identify high-risk foot conditions. This examination should include an assessment of protective sensation, foot structure and biomechanics, vascular status, and skin integrity.

[edit] MANAGEMENT

[edit] Type 1 Diabetes

[edit] Goals of Therapy.

The importance of glycemic control in diminishing risk of complications in type 1 diabetes was unambiguously demonstrated by the Diabetes Control and Complications Trial (DCCT).^[5] The DCCT showed that intensive treatment with a goal of meticulous control decreased the frequency and severity of retinopathy, nephropathy, and neuropathy, by 50% to 70% (Fig. 96-2). The intervention group in DCCT achieved a HbA_1c of 7.2%. However, there was a continuous relationship between glycemic exposure and risk of complications, without a “glycemic threshold.” Similar findings were also obtained in the Stockholm Diabetes Intervention Study (SDIS).^[6] The beneficial effects and impact of effective glycemic control in type 1 diabetes also have been seen in a number of other smaller intervention studies, which collectively were subjected to a meta-analysis^[7] that produced findings consistent with the DCCT. Thus there are a number of randomized controlled clinical trials that support the ADA Standards of Medical Care for Patients With Diabetes Mellitus,^[8] which recommend the treatment targets shown in Table 96-2.

Figure 96-2 Relative risk reductions for microvascular complications, seen with intensive glycemic control, as demonstrated in the DCCT.

Table 96-2 American Diabetes Association Glycemic Targets

[edit] Medical Nutritional Therapy.

Contemporary dietary practice allows flexibility in when and what individuals eat. A meal plan based on the individual patient's lifestyle, food preferences, and eating habits should be determined and used as a basis for integrating insulin therapy into lifestyle. Insulin regimens are integrated with lifestyle and adjusted for deviations from usual eating and exercise habits. Patients monitor blood glucose levels and adjust insulin doses for the amount of food usually eaten. To accomplish this, patients and their families learn a system that incorporates calorie and nutrient content of foods (e.g., exchanges, carbohydrate counting). They also should learn general principles of the influence of various foods and of activity on glycemia, and the balancing of these to achieve glycemic control.

Diabetic patients should follow sound general nutritional practices. This particularly means avoiding excess intake of saturated fats and cholesterol, which influence serum lipids, which also may be elevated if there is suboptimal diabetic control. It also means limiting salt consumption, which may aggravate the risk of blood pressure elevation and may alter vascular reactivity.

Dietary protein should be ˜10% to 20% of total caloric content, with the other 80% to 90% of calories distributed between dietary fat and carbohydrate. In general, 30% or less of the calories should be from total fat, with less than 10% of calories from saturated fats and up to 10% calories from polyunsaturated fats. This leaves 60% to 70% of total calories from monounsaturated fats and carbohydrates. Dietary cholesterol is limited to 300 mg or less daily. Sucrose and sucrose-containing foods must be substituted for other carbohydrates and not simply added to the meal plan. In making such substitutions, nutrient content of concentrated sweets and sucrose-containing foods, as well as the presence of other nutrients frequently ingested with sucrose such as fat, must be considered. Saccharin, aspartame, and acesulfame-K may be used as nonnutritive sweeteners.

[edit] Exercise.

Regular physical activity contributes to the determination of dietary calorie content and insulin dose and regimen. Sporadic physical activity, which departs from daily routine, requires compensatory action to avert hypoglycemia (e.g., 10 to 15 grams of carbohydrates every 30 to 45 minutes during the activity). Blood glucose monitored before, during, and after the activity determines effectiveness of the extra carbohydrate. Insulin dose reductions may be used in addition to or instead of extra carbohydrate. Quick-acting, rapidly absorbed carbohydrate should be available during activity in case of hypoglycemia.

Moderately intensive exercise may deplete glycogen stores, resulting in sustained food requirement to replace the glycogen. Thus may occur well after exercise (e.g., 12 hours later). Therefore patients should be cautious when planning evening physical activity.

[edit] Insulin.

Although any discussion on the treatment of type 1 diabetes focuses on insulin therapy, successful treatment cannot be accomplished without sufficient understanding about how physical activity and diet affect blood glucose levels. Furthermore, it is critical to appreciate that a team of individuals expert in diabetes therapy is required to achieve optimal goals. This team does not need to be located in the same clinic or office as the physician supervising the care, but all team members need to be aware of everyone else's responsibilities. At the least, a diabetes nurse educator and a nutritionist need to be available for routine treatment. Furthermore, these other team members should be certified diabetes educators (CDEs), which guarantees they have the fundamental skills required to teach and manage many aspects of diabetes therapy.

Modern management of type 1 diabetes focuses on replication of normal insulin secretion.^[9][10] Also called flexible diabetes therapy, this strategy calls for insulin delivery that comprises a basal and prandial component, frequent home self-monitoring of blood glucose, and less restrictive dietary plans (compared with previous recommendations), yet with specific guidelines on how to alter therapy based on carbohydrate intake. Most important, patients require the self-management skills to correct alterations in metabolic control at the time such occur. This could include a change in insulin dose for premeal hyperglycemia, treatment of hypoglycemia, or addition of carbohydrate at bedtime to compensate for exercise earlier in the day. In addition, patients should understand basic principles of diabetes management during illness—sick day guidelines. All too often, patients with type 1 diabetes either fail to administer enough insulin during a viral gastroenteritis or they fail to measure urinary ketones during illness. As a consequence, life-threatening ketoacidosis may develop.

Available insulins are noted in Table 96-3. Although insulins have traditionally been categorized as being long-acting, intermediate-acting, and short-acting, it would be more relevant to consider them as being used as a basal or prandial insulin component. The former is that part of the regimen responsible for suppression of hepatic glucose (and ketone) production, and the latter is the insulin available for mealtime caloric ingestion. There are several options patients may choose, and each one has advantages and disadvantages.

Table 96-3 Time Course of Action of Human Insulin Preparations

A diabetes algorithm is the term used to describe actions patients take to prevent or correct any alteration in diabetes management. These include insulin supplements (additional insulin used to correct hyperglycemia) and adjustments (a change in the usual or prevailing dose of insulin). Supplements usually are with rapid-onset insulin (regular insulin, insulin lispro, or insulin aspart), whereas adjustments may be made to any insulin component. An adjustment is made when a pattern of blood glucose levels outside the target range is noted. For example, for an individual who takes bedtime NPH insulin and fasting blood glucose levels are consistently above target, an adjustment—in this case an increase—in the bedtime dose of NPH would be suggested. Another important part of a diabetes algorithm is altering the lag time for the onset of a dose of prandial insulin. Premeal hyperglycemia, for example, in a patient using a rapid-onset insulin analog (lispro or aspart) may be effectively treated by increasing the lag time between the dose and eating from 5 minutes to 15 or 20 minutes.

The traditional split-mix regimen (Fig. 96-3) may seem to be less complex than others, but it limits flexibility, especially with timing of lunch. Even if lunch is not delayed, many patients find it necessary to consume a midmorning snack to prevent hypoglycemia. Because both regular and NPH insulin have actions around lunchtime, many patients find it difficult to avoid midday hypoglycemia. Furthermore, many (if not most) patients are above their blood glucose target by dinner, due to dissipation of NPH insulin effect. An increase of the morning NPH dose creates more problems with midday hypoglycemia and assists little with late afternoon hyperglycemia. In addition, NPH insulin dissipation from the presupper injection results in similar problems in the morning with fasting hyperglycemia. Nocturnal hypoglycemia also is more problematic with presupper NPH insulin. One solution to these problems is adding a prandial injection of regular insulin at lunchtime and moving the NPH injection to bedtime (Fig. 96-4). The morning NPH insulin now does not have to act as a prandial insulin component for lunch and thus can better function as a basal component. Nocturnal control, often considered the Achilles' heel of type 1 diabetes therapy, should provide less of a risk of hypoglycemia and provide higher levels of insulin when they are often required (i.e., around the time of awakening, to counteract the dawn phenomenon).

Figure 96-3 Schematic representation of idealized insulin effect provided by insulin regimen consisting of two daily injections of regular (REG) insulin and intermediate-acting insulin (NPH or lente). Arrows indicate time of insulin injection, 30 minutes before meals. B, Breakfast;L, lunch;S, supper;HS, bedtime snack.

Figure 96-4 Schematic representation of idealized insulin effect provided by multiple-dose regimen providing preprandial injections of regular (REG) insulin before meals, and basal regimen consisting of two daily injections of intermediate-acting insulin (NPH or lente). Arrows indicate time of insulin injection, 30 minutes before meals. B, Breakfast;L, lunch;S, supper;HS, bedtime snack.

For an improved flexible regimen, prandial insulin may be administered with each meal. Patients learn to estimate the appropriate dose based on anticipated carbohydrate intake and the prevailing blood glucose at the time of the meal. Often, individual patient experience becomes an important part of daily management. What becomes clear is that frequent home blood glucose monitoring is required, and patients need to be able to review the large amount of information they may generate. Written glucose log books allow patients to think about each glucose level at the time it is tested. Among other things, insulin doses may be written down, and comments for changes in daily schedule may be noted. Many blood glucose meters now have memories that can be downloaded to a computer. A vast array of statistical information, graphs, and charts can be generated. These may suggest trends in glucose that are not obvious with the log book. Therefore many patients and physicians prefer using both of these techniques for reviewing blood glucose data.

For the basal insulin component, one may use bedtime NPH or lente insulin, with or without a morning injection of the same insulin (Fig. 96-4). Alternatively, ultralente insulin or the new basal insulin analog, insulin glargine, may be used. These have less of a peak than NPH or lente. If ultralente or glargine is used as the basal component, it still may be desirable to use two daily doses (before breakfast and dinner[Fig. 96-5]). Another option is to give ultralente or insulin glargine before breakfast, together with bedtime NPH as the overnight basal component, the latter to target the severe insulin resistance some patients experience in the mornings, and thus overcome the dawn phenomenon.

Figure 96-5 Schematic representation of idealized insulin effect provided by multiple-dose regimen providing preprandial injections of regular (REG) insulin before meals, and basal long-acting insulin (ultralente or insulin glargine). Arrows indicate time of insulin injection, 30 minutes before meals. B, Breakfast;L, lunch;S, supper;HS, bedtime snack.

There are also several options for the prandial insulin component. Regular insulin has the advantage of being better suited for meals high in fat and protein. However, postprandial hyperglycemia occurs with higher carbohydrate meals. Furthermore, to be most efficacious, regular insulin should be administered 20 to 30 minutes prior to eating, unless there is premeal hyperglycemia in which case one should wait longer. The rapid-onset insulin analogs, insulin lispro and insulin aspart, are better suited for higher carbohydrate-containing meals. Slower absorbed meals that are high in fat and protein content may result in greater problems with postprandial hyperglycemia after the analog (lispro or aspart) has dissipated. These analogs can be used as prandial insulin together with basal insulin provided either by intermediate-acting insulin, NPH or lente (Fig. 96-6), or by long-acting insulin, ultralente or insulin glargine (Fig. 96-7). Compared with regular insulin, hypoglycemia is decreased with insulin lispro and insulin aspart, since there is less interaction with basal insulin, especially NPH. Hypoglycemia related to exercise also tends to be less problematic when the activity occurs greater than 2 hours after the last injection of lispro or aspart.

Figure 96-6 Schematic representation of idealized insulin effect provided by multiple-dose regimen providing preprandial injections of rapid-acting insulin (insulin lispro or insulin aspart) before meals, and basal regimen consisting of two daily injections of intermediate-acting insulin (NPH or lente). Arrows indicate time of insulin injection, 30 minutes before meals. B, Breakfast;L, lunch;S, supper;HS, bedtime snack.

Figure 96-7 Schematic representation of idealized insulin effect provided by multiple-dose regimen providing preprandial injections of rapid-acting insulin (insulin lispro or insulin aspart) before meals, and basal long-acting insulin (ultralente or insulin glargine). Arrows indicate time of insulin injection, 30 minutes before meals. B, Breakfast;L, lunch;S, supper;HS, bedtime snack.

Many patients have learned that for certain situations they do better by mixing regular insulin and a rapid-onset analog. For example, if insulin lispro or aspart is usually used but a particular meal will have more fat than usual, it might be wise to mix half the dose as regular insulin. Alternatively, for individuals who are doing well on regular insulin, insulin lispro or insulin aspart may be used as supplemental insulin for premeal hyperglycemia.

The most precise way to replicate normal insulin secretion is to use an insulin pump in a program of continuous subcutaneous insulin infusion (CSII) (Fig. 96-8). By delivering microliter amounts of insulin on a continual basis, basal insulin secretion is replicated. Programming a variable basal rate counteracts the dawn phenomenon and other variations in insulin sensitivity that otherwise result in disruption of glycemic control. The pump is activated before meals to provide prandial insulin increments as meal boluses, allowing total flexibility in meal timing and replicating physiologic prandial insulin availability. If a meal is skipped, the prandial bolus is omitted. If a meal is larger or smaller than usual, a larger or smaller bolus is used.

Figure 96-8 Schematic representation of idealized insulin effect provided by continuous subcutaneous insulin infusion using insulin lispro or insulin aspart. B, Breakfast;L, lunch;S, supper;HS, bedtime snack.

Typical insulin dose in type 1 diabetes is 0.5 to 1 unit per kg body weight per day, less during the “honeymoon” period of relative remission early in the disease, more during the adolescent growth spurt or intercurrent illness. Basal insulin is about 40% to 60% of the total daily insulin dose, with the rest divided among the meals.

[edit] Self-monitoring of Blood Glucose (SMBG).

SMBG is essential for guiding the therapeutic plan. At a minimum, SMBG should be done four times daily—before meals and at bedtime. Nocturnal measurements and occasional postprandial values also are helpful. It is important that the patient understands his or her individualized target blood glucose levels (Table 96-4).

Table 96-4 Representative Target Blood Glucose Levels Suitable for a Young Otherwise Healthy Patient With Type I Diabetes Mellitus

[edit] Type 2 Diabetes

[edit] Goals of Therapy.

The importance of glycemic control in diminishing risk of complications in type 2 diabetes was demonstrated by the large United Kingdom Prospective Diabetes Study (UKPDS)^[11][12] and a small study from Kumamoto University in Japan.^[13] In both of these studies, meticulous glycemic control decreased the frequency and severity of microvascular complications (Fig. 96-9). The intervention group in UKPDS achieved a HbA_1c of 7%. Moreover, a linear relationship between glycemic exposure and risk of complications was found across the spectrum of HbA_1c, without a glycemic threshold. The consistent and substantial beneficial effects of improved glycemic control in both type 1 and type 2 diabetes suggest that the impact of glycemic control on complications may be generalized to all categories of patients. This supports the ADA standards and treatment targets shown in Table 96-2.

Figure 96-9 Relative risk reductions for major endpoints, seen with intensive treatment policy, as demonstrated in the UKPDS.

[edit] Medical Nutrition Therapy.

Medical nutrition therapy is an essential component of successful diabetes management. The emphasis in the plan in type 2 diabetes should be placed on achieving glucose, lipid, and blood pressure goals.

In obese patients with type 2 diabetes a major focus is weight reduction and restriction of total calorie intake. Even mild to moderate weight loss (10 to 20 pounds [5 to 10 kg]) improves short-term glycemic control. Weight loss is best approached by a moderate decrease in calorie intake, coupled with an increase in caloric expenditure.

Additional principles that facilitate glycemic control are a balanced nutrient intake; emphasis on appropriate alterations for achieving lipid and blood pressure goals; adequate spacing between meals (i.e., 4 to 5 hours apart); consumption of dietary fiber; and avoidance of excessive intake of rapidly absorbed, simple sugars (i.e., sucrose, glucose, maltose), confining their use to substitution for other carbohydrates and not as simply added to the meal plan. Coexisting conditions—dyslipidemia, renal disease, hypertension—may require alteration in nutrient content.

[edit] Exercise.

Exercise improves insulin sensitivity and facilitates insulin action; increases energy expenditure; improves cardiovascular conditioning; facilitates control of hypertension; and improves dyslipidemia. A formal exercise training program is not necessary. Patients should increase physical activity to a tolerable level. Evaluation should include an exercise-stress electrocardiogram in all individuals >35 years old to detect silent ischemic heart disease. Patients should self-monitor their glycemic response to exercise.

[edit] Pharmacologic Therapy.

One of the greatest changes in the past several years in clinical medicine is the increase in treatment options for management of type 2 diabetes (Table 96-5). With all these options, it is tempting to disregard the dramatic effects of exercise and diet on glycemic control. Nonpharmacologic treatments need to be emphasized indefinitely because none of the drug therapies will have their maximum impact otherwise. Despite all of the medications, there appear to be “secondary failures” after initial response. These are recognized by deterioration of glycemic control. Such deterioration in control may occur because of disease progression, lack of dietary adherence, intercurrent illness, or loss of pharmacologic effect (drug failure). Usually, however, these are not true drug failures, but masking of pharmacologic effect by disease progression. The fact that the underlying pharmacologic effect remains intact can be demonstrated by adding another agent with a different mode of action. In clinical trials with that design, marked improvement of glycemic control is seen when a second agent is added, whereas little change in glycemia is seen by switching to the new agent. Nevertheless, in view of possible loss of effect of any pharmacologic agent, glycemia should be regularly monitored to regulate dosage and verify that beneficial effects are sustained.

Table 96-5 Characteristics of Oral Antidiabetic Agents Available in the United States

Including insulin, there are currently six classes of drugs available for the treatment of type 2 diabetes. There is no consensus as to which drug should be used first, although the UKPDS suggested that metformin may be preferable as first-line therapy in obese patients with diabetes.^[12]

[edit] Insulin Secretagogues.

Two categories of drugs, sulfonylureas and meglitinides, act to stimulate insulin secretion. Secretagogues improve β-cell function, thus correcting one of the two fundamental defects that characterize type 2 diabetes. They act by binding to a receptor unit on the β-cell, which results in closure of an ATP-dependent potassium ion channel. This causes depolarization of the plasma membrane, resulting in intracellular calcium accumulation, that in turn causes migration of insulin secretory granules to the membrane surface, there poised for insulin release on signaling in response to hyperglycemia. Excess hepatic glucose production is suppressed by secretagogues, presumably as a result of secretion of insulin into the portal circulation. Some studies suggest chronic improvement in insulin sensitivity, presumably secondary to correcting glucose toxicity.

The newest sulfonylurea formulations, glipizide-GITS and glimepiride, are long-acting sulfonylureas taken once daily, with sufficient duration to both control fasting glycemia and improve meal-related insulin secretion. Traditional formulations of second-generation drugs—glipizide and glyburide—must be taken two to four times daily, and less adequately control fasting glycemia, while increasing the risk for hypoglycemia due to large sustained postprandial peaks of drug. First-generation sulfonylureas have a greater likelihood of side effects and of interactions with other drugs, and are only rarely used.

On the other hand, the meglitinides such as repaglinide and nateglinide are ultrashort in duration of action, stimulating insulin to coincide with meals, and thus are taken together with any meal consumed. This offers flexibility in dosing, which some patients may find advantageous.

There is a high initial response rate to secretagogues, with 80% to 90% of subjects showing a response (defined as a decrement in plasma glucose of at least 30 mg/dl). Moreover, the response time is rapid, with improvements in glycemia seen within 24 to 48 hours. As monotherapy, these agents result in a HbA_1c decrease of ˜1.5% to 2%. Although very effective, they may lose their effectiveness over time. One clue that a secretagogue is still effective during a period of otherwise high blood glucose levels is when hypoglycemia occurs, perhaps after exercise or a missed meal. The presence of hypoglycemia proves that the secretagogues are still effective.

Disadvantages of secretagogues include the major side effects of hypoglycemia (which in some instances may be prolonged or severe); weight gain; drug interactions (especially with the first-generation compounds); and hyponatremia with chlorpropamide.

Advantages of secretagogues include improvement in a primary pathophysiologic impairment, that of insulin secretion; a physiologic route of insulin delivery (from the pancreas into the hepatic portal circulation); a high initial response rate; and no lag period before response. Additional advantages of repaglinide are that it may be used in renal insufficiency and that there is flexible dosing in relationship to meals.

[edit] Biguanides.

The glucose-lowering effects of the biguanides were first shown in the nineteenth century, although this class was not introduced clinically until the 1950s. One biguanide, phenformin, was withdrawn by the Food and Drug Administration (FDA) in 1977 because of propensity to lactic acidosis, sometimes fatal. Another biguanide, metformin, which had been available for 4 decades, was released in the United States in 1995. Biguanides have a complex and poorly understood mechanism of action, but appear to directly effect glucose metabolism, thus improving insulin sensitivity, particularly at the liver, where they decrease hepatic glucose output, while also increasing peripheral (muscle) glucose uptake and utilization.

Metformin may be used as initial monotherapy in type 2 diabetes, particularly in obese patients because, unlike secretagogues, it is not associated with weight gain on improvement in glycemia, and in fact a modest weight loss may be seen. Metformin enjoys particular success in combination with secretagogues, since biguanides and secretagogues have complementary mechanisms of action. Indeed, most patients require this combination to achieve glycemic targets. On the other hand, in the setting of inadequate glycemic regulation, switching from a secretagogue to metformin, or vice versa, does not result in improvement of glycemic control.

As monotherapy, metformin results in a HbA_1c decrease of ˜1.5% to 2%. There is a high initial response rate, with 80% to 90% of subjects showing a response (defined as a decrement in plasma glucose of at least 30 mg/dl). Because of the need for slow dose titration in order to minimize the gastrointestinal side effects (nausea, abdominal discomfort, and less frequently, diarrhea) that are seen on initiation of therapy, the full effectiveness of metformin may not be seen for 4 to 6 weeks.

It is important to be sure that patients receive adequate doses of metformin. A dose-response study demonstrated that maximum glucose lowering was achieved with a dose of 1000 mg twice daily, which is substantially higher than that prescribed by most physicians.

Unlike phenformin, metformin is rarely associated with lactic acidosis, provided its use is avoided in patients with elevated serum creatinine, hepatic disease, congestive heart failure, or cardiovascular compromise. Following radiographic procedures involving contrast dyes, metformin should be withheld until it is ensured that acute renal insufficiency has not been induced by the contrast media.

Disadvantages of metformin include gastrointestinal side effects on initiation of metformin (forcing gradual dose increments) and the risk of lactic acidosis in the circumstances noted above.

Advantages of metformin include improvement in a primary pathophysiologic impairment, that of insulin resistance; a high initial response rate; a long record (40 years) of relative safety; and the fact that it is associated with absence of weight gain and may result in modest weight loss.

[edit] α-Glucosidase Inhibitors.

The α-glucosidase inhibitors competitively bind to the carbohydrate-binding region of α-glucosidase gastrointestinal enzymes (sucrase, maltase, isomaltase, amylase, glucoamylase), thus slowing digestion of complex carbohydrates, oligosaccharides, and disaccharides. These actions result in retardation of gut glucose absorption. In the United States, acarbose became available in 1996, miglitol in 1999, and voglibose is in clinical trials. These agents inhibit intestinal brush border enzymes and reduce postprandial hyperglycemic excursions. This results in a modest improvement in HbA_1c, which on average is about 0.5% to 1%. Although occasionally used alone, their primary role is in combination with other agents when glycemic targets are not met.

Because carbohydrates remain in the gut, these drugs frequently produce gastrointestinal side effects, particularly high frequency of flatulence, which is often severe. Other gastrointestinal side effects include nausea, abdominal discomfort, borborygmi, and diarrhea. To some extent, these can be reduced by very slow dose titration to therapeutic levels, but continuing flatulence is bothersome to many patients.

Caution must be exerted in treatment of hypoglycemia by patients using α-glucosidase inhibitors. Because these enzymes are inhibited, hypoglycemia cannot be treated with sucrose, maltose, or starch, since glucose is not readily available for gut absorption. On the other hand, lactose (e.g., milk) may be used to treat hypoglycemia, since lactase is a β-glucosidase and not inhibited. Monosaccharides, including glucose itself and fructose, may be used to correct hypoglycemia.

Disadvantages of α-glucosidase inhibitors include the fact that a high carbohydrate diet is required for efficacy, since it is a competitive enzyme inhibitor; the flatulence and other gastrointestinal side effects; its limited effect on fasting plasma glucose; and the need to alert patients about the ineffectiveness of many usual treatments of hypoglycemia.

Advantages of α-glucosidase inhibitors include a good safety profile; lack of weight gain with improved glycemic control; and a unique mechanism of action that allows these drugs to be combined with any other class of glucose-lowering agent.

[edit] Thiazolidinediones.

The most recent class of oral agents for the treatment of diabetes is the thiazolidinediones (glitazones). These agents act by binding to nuclear receptors called peroxisome proliferator-activated receptors (PPARs). These receptors are important regulators of lipid homeostasis, adipocyte differentiation, and insulin action. The thiazolidinediones act by binding to the PPAR gamma subtype (PPARγ), resulting in the expression of a number of gene-encoding proteins that enhance cellular insulin action on glucose and lipid metabolism. As a consequence, there is improvement in insulin sensitivity, particularly resulting in increased peripheral (muscle and adipose) glucose uptake and utilization, with only a modest effect at the liver. The effect of thiazolidinediones on improving target cell insulin action has been demonstrated in type 2 diabetes, in IGT, and in obese individuals with normal glucose tolerance.

In the United States, troglitazone became available in 1997 and rosiglitazone and pioglitazone in 1999, and several other thiazolidinediones are in various stages of development.

Thiazolidinediones are most effective when given in conjunction with insulin in patients with type 2 diabetes, in which circumstance insulin doses can be lowered and improvement in glycemic control achieved, including attaining control in some individuals who previously were refractory to glucose lowering in spite of large doses of insulin. Effectiveness is also seen in combination with secretagogues, metformin, or both (so-called triple therapy). When used in monotherapy, the response rate is variable, and there is no easy way to anticipate who will be “responders.” Rosiglitazone and pioglitazone appear to have a higher response rate than troglitazone, although this point is moot, since troglitazone is no longer on the market.

The major problem with troglitazone was that of idiosyncratic liver disease, which in some cases was associated with acute hepatic necrosis, resulting either in death or need for liver transplantation. It was recommended that liver enzyme levels be measured at the start of therapy and monthly thereafter. In spite of monitoring, cases of severe liver dysfunction were missed. Although this was a rare problem, it is unfortunately unpredictable. Therefore troglitazone was withdrawn from the market. Fortunately, the clinical trial experience with rosiglitazone and pioglitazone suggests that these agents appear devoid of this risk of liver disease. If that proves to be the case, this class of agents will enjoy widespread usage.

Other disadvantages of thiazolidinediones include the fact that there is a delayed onset of action (up to 3 weeks) and a prolonged time to see the full effect (10 to 12 weeks); weight gain; a relatively high nonresponse rate in monotherapy; increased levels of LDL cholesterol; and unknown long-term side effects since this is a relatively new class of drug.

Advantages of thiazolidinediones include improvement in a primary pathophysiologic impairment, that of insulin resistance; a unique mechanism of action that allows it to be combined with any other class of glucose-lowering agent; once-daily dosing; lowering of serum triglycerides and increase in HDL cholesterol; and that some agents may be used in renal insufficiency.

[edit] Insulin Therapy in Type 2 Diabetes.

Because of the progressive nature of β-cell dysfunction in type 2 diabetes, insulin therapy often becomes necessary. It should also be appreciated that enough insulin will almost always overcome insulin resistance and improve glucose metabolism. In addition, insulin therapy can be used to overcome glucose toxicity and correct the reversible components of the defects in insulin secretion and insulin action.

Disadvantages of insulin therapy in type 2 diabetes include induction of hypoglycemia, weight gain, the need for injections and nonacceptance of that by patients or physicians, and the fact that subcutaneous injections represent a nonphysiologic route of administration (i.e., the peripheral circulation rather than the hepatic portal circulation).

Advantages of insulin therapy in type 2 diabetes include flexibility in dosing and lifestyle by virtue of the multiple preparations with different action profiles, the ability to control all patients (although this may require very high doses), and that it can be used to overcome glucose toxicity.

It should be noted that temporary insulin therapy may be used to attain glycemic control, to overcome glucose toxicity, or to re-regulate decompensated patients. Insulin often is used in combination with various oral medications. All classes of oral medications can be used with insulin. In combination therapy, bedtime NPH insulin has been shown to be an effective strategy of initiating insulin therapy in patients with type 2 diabetes already on an oral agent. It targets the fasting glucose, and thus lowers the level of glycemia above which daytime glucose excursions occur. Alternatively, such basal insulin may be given as ultralente insulin or insulin glargine, both long-acting insulins, or by continuous insulin infusion. However, over time, insulin deficiency becomes more pronounced and a more physiologic regimen will be required, similar to what is used in individuals with type 1 diabetes (see above).

Insulin therapy is often used in the elderly as a last resort, after failure of dietary management and maximum doses of oral hypoglycemic agents. The aim of therapy in the elderly is to relieve symptoms and prevent both hypoglycemia and acute complications of uncontrolled diabetes (e.g., hyperosmolar states). Schedules for the injection of insulin should be kept as simple as possible, since self-administration may be difficult and dosage errors are not uncommon. Premixed insulins may be particularly desirable due to their simplicity of use.

Insulin therapy in type 2 diabetes need not be permanent. With correction of glucose toxicity, there is improvement in both endogenous insulin secretion and insulin sensitivity. As a consequence, it may be possible to discontinue exogenous insulin therapy.

[edit] A Therapeutic Strategy.

Although there is controversy about which drug to use initially, it is clear that many agents work well in combination. It also is clear that most patients need more than one drug. It would seem logical to favor combinations of drugs that work by complementary mechanisms of action, such as a secretagogue with either metformin or a thiazolidinedione. Although both are insulin sensitizers, metformin and a thiazolidinedione may be used together, since the former has its greatest effect on the liver, and the latter has its greatest effect on peripheral glucose utilization. Adding oral agents, especially metformin or a thiazolidinedione to insulin may be very effective, or initiating insulin therapy to a patient not reaching targets on sulfonylureas, metformin, or a thiazolidinedione may be very effective.

It is important to emphasize that frequent SMBG levels at home and regular HbA_1c measurements are required to assist in deciding how to optimally use the diabetes drugs. SMBG in particular has become a critical tool because it allows patients to see how different foods and exercise change blood glucose, but it also assists patients to make changes in insulin at the time the test is performed. Furthermore, one should not assume that a deterioration of glycemic control is from the failure of the agent, although that may be the case. If diabetes control is achieved initially but is not maintained (secondary failure), other possibilities should be entertained before taking further action. For example, life stresses, illness, travel, or a change in activity may all cause a deterioration of glycemic control and should not necessitate the addition of a new agent. Temporary insulin therapy may be considered in such circumstances, even if it is just an injection or two of rapid-acting insulin (e.g., insulin lispro or insulin aspart) to correct prevailing hyperglycemia.

[edit] ACUTE COMPLICATIONS

[edit] Diabetic Ketoacidosis

DKA occurs in 2% to 5% of patients with type 1 diabetes each year. Depending on the age group and population, death still occurs in 1% to 10% of patients. The metabolic derangements in DKA result from absolute or relative insulin deficiency and elevated counterregulatory hormones, resulting in severe hyperglycemia, ketonemia and acidemia, and volume depletion.

Laboratory values for DKA include an arterial pH < 7.3, plasma bicarbonate ≤15 meq/L, blood glucose usually >250 mg/dl, and ketonuria with ketonemia. Although the clinical diagnosis is usually apparent in an individual with known type 1 diabetes, it also needs to be considered in a comatose elderly patient with type 2 diabetes or a child with rapid breathing perceived as having a respiratory tract infection. The clinical signs and symptoms of DKA are noted in Box 96-1. Abdominal pain in association with a leukocytosis and high amylase level is not uncommon, but if it persists with therapy, consideration should be given to appendicitis or bowel perforation.

Any intercurrent illness or stress may precipitate DKA, including myocardial infarction, stroke, trauma, and particularly infection. Some patients may mistakenly withhold their insulin during an acute illness due to decreased appetite. This is particularly common during a bout of gastroenteritis, during which patients may be prone to ketosis anyway. Alternatively, some patients (especially adolescent girls) intentionally withhold insulin as a strategy for weight loss.

The most important component for the treatment of DKA is close monitoring. For this reason, many believe that it is best to treat these individuals in an intensive care unit. There should be frequent assessment of fluid intake and output, vital signs, and pertinent laboratory values (glucose, bicarbonate, potassium, sodium, urea nitrogen, and creatinine). Although some authors recommend the measurement of a routine arterial blood gas, especially on admission, this is probably not necessary. Indeed, if a simple metabolic acidosis is discovered, there is no need for subsequent blood gas measurements. A flow sheet with all of the pertinent clinical and laboratory information is invaluable. The essential elements of treatment are listed in Box 96-2.

Volume depletion is present in all patients. It is reasonable to assume a weight loss of 5% to 10% of total body weight, or 3.5 to 7 L of fluid in a 70-kg patient. Isotonic solutions are the fluids of choice; hypotonic solutions run the risk of rapid reduction of plasma osmolality, with large fluid shifts precipitating cerebral edema and hypovolemia. Appropriate solutions are either 0.9% saline or, alternatively, a solution of one ampoule (˜50 mEq) of sodium bicarbonate (plus appropriate potassium supplements) added to each liter of 0.45% saline.

Insulin replacement should be accomplished by an intravenous infusion of regular insulin. Insulin should be delayed if the patient is hypokalemic. A bolus of 0.1 units/kg may be administered, and then infused at 0.1 units/kg/hr. Only in unusual situations (e.g., myocardial infarction or sepsis) will this amount of insulin not result in an improvement in the acidemia and the anion gap. For this situation, the rate of insulin infusion should be doubled and consultation with a diabetologist should be considered.

Glucose is always normalized more quickly than the acidosis is corrected, but the insulin infusion needs to be continued until both are corrected. Five percent dextrose is added to the infusion when the blood glucose reaches 250 mg/dl. When plasma bicarbonate level reaches 18 meq/L and the anion gap has decreased to 15 meq/L, the insulin infusion may be reduced (by about 50%).

All patients with DKA are total body potassium depleted, even if they present with normokalemia or hyperkalemia. Prior to initiating potassium replacement, it is necessary to establish urine output and obtain an electrocardiogram to quickly estimate hyperkalemia or hypokalemia. If low T waves with U waves are noted (indicative of severe hypokalemia), potassium should be initiated even if the serum potassium level is not available from the laboratory. For most situations, potassium replacement should be initiated no later than 1 to 2 hours after starting the insulin infusion.

The use of bicarbonate in the treatment of DKA is controversial. Although bicarbonate may be used as part of the intravenous solution, routine bolus bicarbonate infusion is not recommended. Some authorities recommend bicarbonate only for severe acidemia (i.e., pH < 7), particularly when associated with hypotension, shock, and arrhythmias. When needed, bicarbonate may be given as an infusion of 1 to 2 meq/kg over 2 hours, and then the plasma bicarbonate level should be remeasured.

Phosphate replacement in the treatment of DKA also is controversial. Routine phosphate replacement has not been shown to be of significant benefit. However, with low initial phosphate levels, it is reasonable to replace phosphate as potassium phosphate.

Other considerations in the treatment of DKA are the following. Low-dose subcutaneous heparin is recommended in the elderly. Cerebral edema has been described in children, most likely due to correction of the hyperglycemia too quickly and using excessive amounts of hypotonic saline. Therefore, in children blood glucose should be maintained at 250 mg/dl for the first 12 to 24 hours. Overaggressive fluid replacement may induce congestive heart failure. Frequent auscultation of the lungs is recommended.

[edit] Hyperosmolar Hyperglycemic Nonketotic Syndrome (HHNS)

If DKA is at one end of the spectrum of severe metabolic decompensation, HHNS is at the other end. With HHNS there is a lack of significant ketosis and higher average glucose levels than noted with DKA. By definition, plasma glucose is above 600 mg/dl and plasma osmolality is greater than 320 mOsm/L. Many cases of HHNS are associated with acidemia, often related to lactic acid accumulation or uremia. Average initial laboratory values include a plasma glucose of 1000 mg/dl, serum osmolality of 360 mOsm/L, and a blood urea nitrogen (BUN) and creatinine of 65 mg/dl and 3 mg/dl, respectively. Precipitating factors are similar to those seen in DKA.

The cornerstone of HHNS therapy is fluid replacement. These patients are both volume contracted (loss of isotonic saline), resulting in tachycardia and hypotension, and dehydrated (pure water loss), reflected by hypertonicity. Since the tonicity of isotonic saline (308 mOsm/L) is always hypotonic to the patient's tonicity, most authors suggest initially infusing isotonic saline (usually at 1 L/hr) until blood pressure and heart rate are normalized. With more severe dehydration and hypernatremia, some combination of isotonic saline (to correct volume deficits) and hypotonic saline (to correct hypertonicity and water deficits) will be required. Most patients respond well to receiving the first liter of fluid in the first hour, and then the next liter over the next 2 hours. Thereafter, the rate and tonicity of the fluid must depend on the clinical status of the patient. Elderly patients need to be managed more cautiously.

Rehydration itself will dramatically lower blood glucose levels. However, all patients require insulin therapy, although they are less resistant than those in DKA. Most authors recommend 3 to 4 units/hour initially, although many do well with 1 to 2 units/hour, especially once plasma glucose levels reach 250 mg/dl. At this stage, 5% dextrose should be added to the intravenous solutions, as in the treatment of DKA.

Hypokalemic emergencies are less of a concern in HHNS compared with DKA, most likely because these patients usually have less vomiting before admission. Therefore the rates of potassium replacement should be more cautious, based on renal function and serum potassium level.

[edit] MICROVASCULAR COMPLICATIONS

[edit] Retinopathy

Diabetic retinopathy is the leading cause of blindness in the United States.^[14] Approximately 5% of patients with diabetes progress to severe visual loss of 5/200 or less. Several risk factors exist for the development of diabetic retinopathy. The first is duration of disease. The risk of having any diabetic retinopathy after 15 years duration is 98% for individuals with type 1 diabetes and 78% for those with type 2 diabetes. The second risk factor, now definitively proven, is glycemic control, which should be conceptualized as glycemic exposure. This is considered the amount of time exposed to a given level of hyperglycemia. For example, in the DCCT it was shown that 9-years' exposure to a HbA_1c level of 8% yields approximately the same risk of retinopathy as 2.5 years at a HbA_1c level of 11%. The final important risk is hypertension. In the UKPDS, an average blood pressure of 144/82 compared with 154/87 resulted in a 47% decreased risk in the deterioration of visual acuity from diabetic retinopathy.^[15]

Diabetic retinopathy may be classified as nonproliferative (background) or proliferative retinopathy. Nonproliferative diabetic retinopathy (NPDR) is characterized by structural abnormalities of the retinal vessels, varying degrees of retinal hypoperfusion, retinal edema, lipid exudates, and intraretinal hemorrhages. Neovascularization is the hallmark of proliferative diabetic retinopathy (PDR), and may occur on the optic disc, elsewhere in the retina, or on the iris (rubeosis iridis). Neovascular tissue contains both a vascular and fibrous component. The former may cause preretinal or vitreous hemorrhage, and the latter may interact with the vitreous to produce traction on the retina and subsequent retinal detachment. It needs to be appreciated that both PDR and NPDR may result in visual loss. For example, macular edema is a serious form of NPDR that may only be diagnosed by viewing the macula stereoscopically.

For the primary care physician caring for the patient with diabetes, management of diabetic retinopathy can be divided into prevention, screening, and treatment. For prevention, it is clear that both blood glucose control and blood pressure control can decrease the risk for the development and progression of diabetic retinopathy.

Because treatments for diabetic maculopathy and proliferative retinopathy are most effective when initiated early, it is incumbent on the physician to obtain ophthalmic consultation on patients with diabetes who complain of decreasing vision and/or demonstrate any retinal vascular abnormalities. Screening recommendations include a yearly dilated retinal examination for individuals with type 2 diabetes. For those with type 1 diabetes, yearly screening examinations should begin after 5 years of disease but not before puberty. Women with diabetes who become pregnant should have a dilated eye examination in the first trimester of pregnancy and close follow-up throughout the pregnancy. Screening examinations should be performed by a trained eye care professional and should include a stereoscopic examination.

The Diabetic Retinopathy Study proved that argon laser panretinal photocoagulation significantly decreases the likelihood that eyes with proliferative retinopathy will progress to severe visual loss.^[16] Other controlled studies proved the effectiveness of focal argon laser photocoagulation in decreasing or stabilizing diabetic macular edema and the value of vitrectomy in managing the severe complications of diabetic retinopathy.

It needs to be appreciated that numerous studies have shown that rapid improvement of glycemic control may actually result in a worsening of preexisting retinopathy. Patients at highest risk of this are those with longstanding poor control and a moderate stage of NPDR. Although there was no serious visual loss noted in subjects participating in the DCCT, other reports have noted a decrease in visual acuity. Patients at high risk of early worsening should have more frequent ophthalmologic evaluations. Although not yet formally studied, many recommend a slower improvement of glycemic control in these patients. For those individuals who already have advanced retinopathy, it would be prudent to delay improvement of glycemic control until after photocoagulation treatment is completed.

Although not considered microvascular disease, there are other ocular complications that occur more frequently in people with diabetes. The first is cataract formation, which is present in 22% of adults with diabetes compared with 3% of those without. For those diagnosed with diabetes after reaching the age of 30 years, cataract formation is responsible for more decrease in vision than diabetic retinopathy. Similarly, glaucoma is more common, being present in 7% of individuals with diabetes compared with 1% in the general population. Glycemic control may be important for these complications as well. For example, in the UKPDS the intensive group had a 24% reduction in need for cataract extraction.^[11]

[edit] Nephropathy

Diabetic nephropathy occurs with an overall prevalence of approximately 20% to 30%. Although it is often noted that nephropathy is more common in type 1 than in type 2 diabetes, this point is now under some debate. Younger patients with type 1 diabetes frequently develop ESRD without significant cardiac disease. On the other hand, older individuals with type 2 diabetes may have significant nephropathy (manifested by heavy proteinuria) with normal renal function when they succumb to coronary artery disease. Proteinuria is an independent risk factor of cardiac death, and these older patients have other risks such as dyslipidemia and hypertension. Furthermore, diabetic nephropathy is more common in certain ethnic populations (African-Americans, Native Americans, Mexican-Americans) for which the prevalence of type 2 diabetes is increasing. Individuals with type 2 diabetes are the largest group entering treatment programs for ESRD in the United States. The mean 5-year life expectancy for patients with diabetes-related ESRD is less than 20% in the absence of transplantation.

The major risk factors for the development or progression of diabetic nephropathy are poor glycemic control and the presence of hypertension. Both of these risk factors are well supported by both observational and prospective studies. Other reported risk factors include duration of diabetes, family history of diabetic nephropathy, male gender, cigarette smoking, and ethnicity (as noted above). There is increasing support that one or several genes may predispose to nephropathy. Hypercholesterolemia may be an independent risk factor for the development and progression of diabetic nephropathy. Indeed, it has been reported that treating patients with nephropathy with a 3-hydroxy-3-methylglutaryl coenzyme A (HMG CoA) reductase inhibitor may slow the progression of established renal disease.

Although the natural history of diabetic nephropathy has been described for type 1 diabetes, it appears to be similar in type 2 diabetes. The hallmark of diabetic nephropathy is albuminuria, which reflects both histologic and functional abnormalities of the kidney. Microalbuminuria is the earliest laboratory finding in nephropathy, and occurs 5 to 8 years before the onset of overt proteinuria. Normoalbuminuria is present with albumin excretion rates below 30 mg/day, whereas microalbuminuria is defined by an albumin excretion rate of 30 to 300 mg/day. An albumin excretion rate greater than 300 mg/day defines clinical nephropathy, also known as overt proteinuria, dipstick positive proteinuria, or macroalbuminuria. At this stage, the usual commercial dipstick for proteinuria is positive.

Without intervention, there is relentless progression from microalbuminuria to clinical nephropathy. Intervention in the microalbuminuria stage (at which time renal function is normal) can reverse the albuminuria and halt or slow the progression of the nephropathy. If intervention does not occur until clinical nephropathy is present (when renal function is normal or slightly decreased), the progression of the nephropathy can be slowed considerably, but it is rare to have it stabilize and at this point it will not reverse. Without any intervention at this stage, the average rate of decline in glomerular filtration rate (GFR) is about 1 ml/min/month, but there is wide variation in this. Without intervention, most patients with type 1 diabetes develop ESRD within 18 years after the diagnosis of their diabetes. With improved screening and intervention, this dismal figure hopefully will improve.

There are four interventions to consider once albuminuria is discovered. The first is meticulous glycemic control. Numerous studies have shown that near-normal glucose control can reduce the albumin excretion rate and prevent the progression to overt proteinuria. There are less data about the impact of glycemic control with more advanced nephropathy, but available evidence suggests the more advanced the renal disease the less impact glycemic control will have on slowing its progression.

The second intervention is scrupulous blood pressure control. Studies from the early 1980s clearly showed that controlling systemic blood pressure slows the rate of decline of renal function and improves survival.^[17] In these studies, the primary drugs used were cardioselective β-blockers and loop diuretics.

Later, it was shown that angiotensin-converting enzyme (ACE) inhibitors have an incremental beneficial effect, by virtue of selective efferent arteriolar dilation.^[18] Indeed, ACE inhibitors offer beneficial effects on microalbuminuria even in the absence of systemic hypertension.^[19] ACE inhibitors also have proven to be of benefit in patients with established nephropathy and mild renal insufficiency.^[20]

The fourth intervention for diabetic nephropathy is dietary protein restriction.^[21] It is thought that this strategy is useful in reducing renal plasma flow, thus improving glomerular hemodynamics in diabetic renal disease. The ADA currently recommends 0.8 grams of protein/kg/day (or about 10% of daily total calories) in patients with clinical nephropathy.

With the identification of these intervention strategies, screening for diabetic nephropathy needs to be a routine aspect of diabetes care. The ADA recommends yearly screening for individuals with type 2 diabetes, and yearly screening for those with type 1 diabetes after 5-years' duration of disease (but not before puberty).^[8] Several screening techniques are available; a random albumin-to-creatinine ratio from a spot collection and a timed (e.g., overnight) or 24-hour urine collection for albuminuria and creatinine are all acceptable. Positive results need to be confirmed with a second measurement due to the high variability in albumin excretion in people with diabetes.Box 96-3 lists substances and circumstances that may result in false-positive screening. Urine dipsticks for microalbuminuria are reasonable for an initial screen but are “semiquantitative,” and long-term data on their use are not available.

With our increasingly effective interventions for diabetic nephropathy, one of the most debilitating of diabetic complications, it is imperative that screening becomes routine. Due to the great expense of dialysis and renal transplantation, it is not surprising that screening has been shown to be extremely cost-effective.

[edit] Neuropathy

There are several classifications of the various diabetic neuropathies. One such clinical classification is noted in Box 96-4.