Annabelle Rodriguez-Oquendo
Robert I. Gregerman, MD wrote this chapter in previous editions.
Definition and Classification
Classically, diabetes mellitus has been considered to be present when an individual exhibits an abnormal elevation of blood sugar in the fasting state or glucose intolerance on glucose challenge. However, diabetes is not a single disease entity, as can be seen from the classifications discussed below.
Until recently, the 1979 classification and terminology proposed by the National Diabetes Data Group of the National Institutes of Health (based on insulin dependence) (1) were widely followed. In 1997, the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus of the American Diabetes Association (ADA) developed a new classification (2) (Table 79.1), followed in this chapter, that is based on etiology. However, there has been some modification of this classification, as new subtypes have been identified.
Type 1 Diabetes Mellitus and Its Slowly Progressive Form
Type 1 diabetes, which accounts for approximately 10% of cases in the United States, generally has its onset in childhood, as early as infancy but more often near puberty, and was once called juvenile-onset diabetes mellitus. However, that designation was misleading because this type of diabetes also can develop in adulthood (3). Its essential characteristic is that insulin dependence is eventually absolute; at this stage, without replacement insulin therapy, ketosis–acidosis ensues within hours. Type 1
P.1294
diabetes mellitus is now considered a chronic autoimmune disease (4), characterized by destruction of β cells in pancreatic islets and eventual failure to synthesize sufficient insulin. A variety of autoantibodies (e.g., to islet cells and their enzymes, and to insulin) have been identified early in the course of the illness. There is clearly a genetic susceptibility to this process (see below), but environmental influences also play a role. Those influences are not yet well defined. There have been claims, for example, that viral infection (5) or antibodies to bovine albumin (6) may be important triggers, but the precise pathogenesis of type 1 disease has not been established.
|
TABLE 79.1 Classification and Clinical Characteristics of Diabetes Mellitus |
||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||
Classification of individual patients into type 1 and type 2 categories on strictly clinical criteria is unreliable. Either type can mimic the other at onset, and sometimes several years pass before the precise diagnosis is made (7, 8, 9). For example, it is now recognized that type 1 diabetes can develop over years, with a prolonged period of hyperglycemia before the onset of ketoacidosis. As a result, the World Health Organization proposed a new subgroup of diabetes that it designated “slowly progressive autoimmune type 1 diabetes,” perhaps better known as latent autoimmune diabetes in adults (LADA) (10). This group is estimated to account for nearly 10% of all cases of diabetes. If this estimate is correct, type 1 may account for as much as 20% of the diabetic population, 10% with onset in children and 10% as LADA.
Up to 80% of persons with type 1 diabetes have islet cell antibodies (10); as stated, a variety of other autoantibodies may be present as well. People with type 1 diabetes who do not have such antibodies have identical presentations to those who do. However, young diabetic patients who are antibody negative may have maturity-onset diabetes of the young (MODY), a variant of type 2 diabetes, with a distinct genetic profile (see Type 2 Diabetes Mellitus) (11).
Inheritance
The genetics of type 1 diabetes are incompletely understood, and prediction of the occurrence of diabetes in
P.1295
offspring of diabetic parents is presently impossible. Even statistical estimates are crude. The prevalence of overt diabetes in offspring of conjugal type 1 diabetic parents is remarkably low, ranging from 3% to 12% in most reports (12). Only 2.5% of siblings of type 1 diabetic patients develop diabetes; when tested initially, these people may show only impaired glucose tolerance (IGT).
If expression of disease in type 1 diabetes was based entirely on genetics, one would expect 100% concordance for diabetes in monozygotic twins. However, this is not the case; both members of pairs develop diabetes only about one-third of the time (13). Thus, environmental factors also must be important (14).
Parents who have a child with insulin-dependent diabetes often wish to know the risk to future offspring. Prenatal histocompatibility leukocyte antigen (HLA) typing of fetal cells obtained at amniocentesis could be compared with that of the sibling with diabetes. A fetus with the same HLA identity would have an increased risk, but the accuracy of the prediction would still be only approximately 50%. The imprecise nature of this assessment is in contrast to the nearly 100% certainty of predicting Tay-Sachs disease or Down syndrome. Thus, even with an accurate family history and pedigree, together with chemical assessment of diabetes (glucose tolerance testing), only crude predictions can be made for a couple who wish to know their own chances of developing diabetes and the risk for their offspring.
At this point, only a few generalities seem safe. Prospective parents should not be told to avoid procreation merely because one parent has diabetes. Even when both parents have diabetes, the risk of having a child likely to develop diabetes is relatively low.
Type 2 Diabetes Mellitus
Type 2 diabetes is often only one component of a complex of abnormalities variously termed the metabolic syndrome, syndrome X, themetabolic syndrome X, or the insulin-resistance syndrome (15). Hyperinsulinemia with or without obvious hyperglycemia is present and denotes the presence of insulin resistance. Other components of the syndrome are obesity (central type), hypertension, fasting and postprandial hyperlipidemia, abnormal concentrations of blood coagulation factors, and premature cardiovascular atherosclerosis. Its pathogenesis remains unclear.
Obesity may be the first manifestation of the metabolic syndrome. Only after some years does an obvious state of diabetes emerge in which all or most of the various features of the metabolic syndrome also become apparent. As the condition evolves, insulin resistance progresses to glucose intolerance and finally to diagnosable type 2 diabetes.
Type 2 diabetes is the most common form of diabetes mellitus and accounts for approximately 80% of patients presenting with an overt abnormality of glucose metabolism. Ordinarily, patients with type 2 disease are neither absolutely dependent on treatment with insulin nor ketosis prone. Nonetheless, some patients being treated with oral hypoglycemic drugs may require insulin to control hyperglycemia or ketoacidosis during stress. Most patients are older than age 40 years at the time of diagnosis, but this type of disease is also seen in young people, which is why older terms such as maturity-onset diabetes have been abandoned.
A current concept of the evolution of the common form of type 2 disease is that obesity-related insulin resistance is superimposed on an individual with a limited ability to secrete sufficient insulin (i.e., β-cell dysfunction, presumably genetic) to compensate for the insulin resistance. In this scenario, obesity-related insulin resistance is the immediate cause of the glucose intolerance (hyperglycemia). Without obesity, an overt diabetic state would presumably not develop (16). However, 10% to 15% of type 2 patients are not obese. Some of these individuals are also insulin resistant, presumably secondary to other causes, perhaps genetic as well, but not all of the cases are, in fact, insulin resistant. At this point, the scenario becomes more speculative because longitudinal data on individual patients are not available. Long-standing hyperinsulinemia may eventually diminish, the result of progressive impairment in the patient's ability to secrete insulin, possibly as a result of the accumulation in pancreatic islets of amylin, an amyloidlike protein. The degree of β-cell “exhaustion” is related in part to the duration of the illness. Some patients ultimately develop insulinopenia to the point that they become totally dependent on exogenous insulin and are at risk of developing ketoacidosis if stressed.
Included in the group of type 2 diabetic patients are those who develop the disease before they reach adulthood Maturity Onset Diabetes of the Young (MODY). These young adults (or older children) have what was once considered to be a “variant” of type 2 diabetes shown to be a heterogeneous genetic disorder; usually they are not obese (11). MODY is relatively rare in most populations, accounting for approximately 2% of all cases of type 2 disease (17), but it may be increasing in prevalence. MODY should not be confused with a more common problem: An increasing number of children and young adults are becoming obese and developing “ordinary” type 2 diabetes. This problem is considered to be of epidemic proportions in the United States (18) and is also seen worldwide, mostly in children older than the age of 10 years. When diabetes is diagnosed, most prove to have a strong family history of type 2 diabetes. Nonobese young diabetic individuals should be tested for autoantibodies and C peptide to exclude type 1 diabetes and if these tests are negative or low, respectively, MODY should be considered.
Atypical diabetes is a variant of type 2 diabetes that was first described in adults who seemed to have type 2 but who
P.1296
developed ketoacidosis during stress and required control with insulin. Adult African American patients exhibit this response with some regularity, and it was recently seen in obese African American children. When the ketoacidosis is seen in a child, the issue of type 1 disease is raised, but obesity is the clinical clue that “atypical diabetes” is the correct diagnosis. MODY does not need to be considered in African Americans; to date, it has been described only in whites, Japanese, and, rarely, Chinese persons.
Behavioral, and possibly environmental, factors appear to be involved in the onset of type 2 diabetes. Especially prominent is the role of excessive caloric intake and subsequent obesity in most cases. Although it is clear that obesity somehow aggravates the underlying genetic predisposition for the development of the metabolic abnormalities of the diabetic state and that weight loss often ameliorates them (above), it is also clear that the factors driving the development of obesity are themselves multifactorial. In type 2 diabetes, association with certain HLA subtypes and with antibodies to islet cells has not been found. Blood insulin levels vary depending on the stage of the disease and may be supranormal in the early years and subnormal later in the disease. Insulin resistance is the rule, but measurement of insulin concentration has no clinically diagnostic or therapeutic usefulness. Measurement of insulin C peptide is sometimes useful, along with antibody determinations (see the discussion of LADA under Type 1 Diabetes Mellitus and Its Slowly Progressive Form).
Inheritance
Type 2 diabetes mellitus is genetically and clinically heterogeneous (17,19) with a much more obvious familial pattern of expression than type 1 disease. Impaired first-phase insulin secretion is the earliest detectable abnormality in type 2 disease and is commonly abnormal in the first-degree relatives of patients with type 2 disease, even when their conventionally measured oral glucose tolerance is normal. In contrast to type 1 diabetes mellitus, there is essentially complete concordance for type 2 diabetes in monozygotic twins (20).
Studies of ethnic groups show distinctive patterns of inheritance of type 2 diabetes and superimposed geographic (environmental) effects on these patterns. The most easily apparent correlate is obesity. Certain Native American tribes (e.g., Pima, Navajo) show a remarkably high prevalence of diabetes, with approximately 50% of the adults having the disease; obesity appears to be the major factor in expression of diabetes in these groups. The Hispanic population (Mexican American) of the southwestern United States also shows similar, but less frequent, expressions of obesity and diabetes. In nonnative populations in the United States, no such distinctive ethnic or racial patterns in the pathology of type 2 diabetes have been recognized to date.
|
TABLE 79.2 Diseases Associated with Diabetes Mellitus |
||||||||||||||||
|
Other Types of Diabetes
Sometimes diabetes is associated with another disease (Tables 79.1 and 79.2); usually the association is infrequent but more common than in the general population. This heterogeneous group includes some disorders in which there is a clear relationship between the associated disease and the diabetes (e.g., chronic pancreatitis) and many others in which an association has been noted but is not well understood (e.g., primary hyperaldosteronism).
Problems in Classification of Individual Patients
On occasion, classification may be difficult. For example, an adult with ketoacidosis may be erroneously classified as a type 1 diabetic when the diabetes is type 2, with insulin dependence having been precipitated by the temporary stress of infection or trauma (see the discussion of “atypical diabetes” under Type 2 Diabetes Mellitus). Similarly, the process of distinguishing between a patient with type 1 diabetes and a thin patient with type 2 disease for whom insulin has been prescribed may require diagnostic procedures to exclude the possibility that the diabetes is one of the “other types” (Table 79.2).
Clinical Presentation
Most diagnoses of diabetes mellitus are now made at an asymptomatic stage of the disease as a result of routine blood tests that reveal elevation of plasma glucose (PG) concentration. When the diagnosis is actively sought, oral glucose tolerance tests (OGTTs) reveal additional cases because up to one-fourth of patients with a diagnostic OGTT have a normal fasting plasma glucose (FPG) concentration. Unless the fasting glucose is elevated, patients do not have enough glucosuria to become symptomatic. Of patients with overt hyperglycemia who are symptomatic at time of diagnosis, most complain of polyuria, which is
P.1297
caused by the osmotic diuresis induced by the glucosuria; polydipsia; and, if the disease is very severe, polyphagia, often associated with weight loss if the increased food intake falls short of full compensation for the caloric loss that results from heavy glucosuria. All these symptoms are manifestations of hyperglycemia and secondary glucosuria. Other symptomatic manifestations include blurred vision (osmolality-related changes in the shape of the lens of the eyes), vaginitis (usually caused by monilial infection), and skin infections. Furuncles and carbuncles, once common, are now rarely seen, but intertriginous candidiasis is common in the obese, and thrush is common in patients with poor oral hygiene or poorly fitting dentures.
Usually, these symptoms are present for weeks or months before medical attention is sought. The onset of symptoms is often insidious and may be attributed by the patient, or even by the clinician, to emotional factors or a common problem such as a urinary tract infection. Indeed, the diagnosis may be missed for a time because the clinician, failing to consider the evolving character of the disease, believes that the patient does not have diabetes on the basis of previous evaluation.
Many patients with type 2 diabetes present with minimal or no symptoms of hyperglycemia and glucosuria but have already developed complications such as neuropathy or, more commonly, vascular disease. Also, it is common to encounter patients who believe that their long-standing diabetes is “mild” only to find themselves with severe complications of the disease. Occasionally, a patient may be completely unaware of having diabetes and yet present with retinopathy or nephropathy.
Diagnosis
Hyperglycemia is the hallmark of diabetes mellitus. Glucosuria alone is not a pathognomonic finding because rare patients may have a renal tubular glucose leak (renal glucosuria) at normal concentrations of blood sugar. Often, patients show diagnostic hyperglycemia (fasting; postglucose load in a glucose tolerance test; or, postprandially) before glucosuria develops.
Criteria for Diagnosis of Diabetes Mellitus
The following revised criteria were established in 2003 by the Expert Committee of the ADA (2): (a) unequivocal elevation of PG concentrations associated with classic symptoms of diabetes mellitus, or (b) elevation of FPG on more than one occasion, or (c) elevation of PG after an oral glucose challenge (standardized OGTT) on more than one occasion. A single elevated FPG or a single OGTT does not establish the diagnosis (Table 79.3).
The ADA's decision (2) to promote the FPG, with a new and lower cut point rather than the OGTT as the primary basis for a diagnosis of diabetes, was based in part on the practical consideration that few physicians were performing glucose tolerance tests in any event and that the OGTT would best be reserved for research purposes. They further argued that, because of its simplicity, careful attention to the FPG would result overall in a larger number of diagnoses of diabetes than was being made with the OGTT. The diagnostic cut point of 126 mg/dL (7.0 mmol/L) is not arbitrary; it is the level in several studies at which the risk begins for the development of diabetic retinopathy, whereas 110 mg/dL is the point above which acute-phase insulin secretion is lost in response to intravenously administered glucose, the hallmark of early diabetes. The ADA did not completely abandon the OGTT or modify its long-time cut points in the OGTT; a diagnosis of diabetes is established by either an elevated FPG or the OGTT. Nonetheless, a number of studies of different populations have shown that many fewer diagnoses of diabetes will be made when the ADA's fasting glucose limit is used, even though the new fasting diagnostic level is lower than the old one of 140 mg/dL. These studies clearly show that ADA's criteria for diagnosis are less sensitive than the OGTT (10). There is also poor correlation between the new category of impaired fasting glucose (IFG, see below) and IGT, the latter term applying only to the values obtained using the OGTT (Table 79.3). There is also poor correlation between the new category of impaired fasting glucose (IFG) and impaired glucose tolerance (IGT), the latter term applying only to the values obtained using the OGTT (Table 79.3) (see Impaired Fasting Glucose and Impaired Glucose Tolerance).
|
TABLE 79.3 Interpretation of Values for Plasma Glucose |
||||||||||||||||||||||||
|
||||||||||||||||||||||||
P.1298
In modern laboratories in the United States, glucose is determined in plasma or serum. Plasma and serum values are identical, but both are 5% to 15% higher than those obtained in whole blood from which they are derived. Portable devices for measuring blood glucose are not sufficiently accurate to be used in diagnosis, and abnormal “fingerstick” results should be confirmed with standard tests on venous blood.
Fasting Plasma Glucose
The ADA cut point for making a diagnosis of diabetes is a FPG ≥126 mg/dL (Table 79.3). Two values of 126 mg/dL (7.0 mmol/L) or greater obtained on different days are needed for a definitive diagnosis. Values above 100 mg/dL and less than 126 mg/dL designate impaired fasting glucose. It should be remembered that FPG may be elevated transiently by stress or illness.
Oral Glucose Tolerance Test
The OGTT, for many years an accepted diagnostic standard for the diagnosis of type 2 diabetes, is no longer recommended by the ADA for routine clinical use (see Criteria for Diagnosis of Diabetes Mellitus). It is less convenient, more costly, and more variable than the FPG; it is, however, more sensitive for establishing a definitive diagnosis. The test still has a place in diagnosis during pregnancy.
Table 79.3 lists the diagnostic criteria for the OGTT. The test may be falsely abnormal in people who have had a recent stressful illness, have had a reduced food intake (less than 150 g carbohydrate per day), or who have been taking one of a variety of drugs (e.g., glucocorticoids and most diuretics). Even smoking or caffeine or performance of the test in the afternoon can cause an abnormal test result. Also, the values in the OGTT tend to increase with age; that is also true of the FPG, but the latter increase with age is very small.
Measurement of a random 2-hour postprandial blood sugar should never be done for screening purposes; it has low sensitivity, specificity, and reliability.
Previous and Potential Abnormalities of Glucose Tolerance
According to the currently accepted scheme, people with a normal OGTT who previously showed either IGT or overt hyperglycemia should be classified as having a “previous abnormality of glucose tolerance.” These people should not be considered diabetic and should not be labeled with the terms prediabetic or latent diabetic. Terms such as subclinical, preclinical, chemical, and borderline diabetes should also be avoided.
Impaired Fasting Glucose and Impaired Glucose Tolerance
Patients whose FPG levels or whose glucose levels obtained during an OGTT fall between normal and diabetes (Table 79.3) are now classified by the ADA into a group having impaired fasting glucose. The term impaired glucose tolerance separates those with glucose intolerance during an OGTT from those who meet the diagnostic criteria for diabetes mellitus.
Significance of Impaired Fasting Glucose or Impaired Glucose Tolerance for Development of Diabetes and Cardiovascular Disease
Both IFG and IGT are risk factors for the development of diabetes. In this combined group, one can expect 1% to 5% per year to develop diagnosable diabetes mellitus. On the other hand, many patients eventually show normalization of glucose tolerance, and still others remain in the IFG or IGT range. The higher the blood sugar within the range of IGT, the greater the tendency to progress to diabetes (21).
Perhaps some of the most convincing evidence on IGT progression has come from long-term studies of Pima Indians (22). The risk of progression to overt diabetes in this group is clearly related to the level of glucose within the range of 160 to 200 mg at 2 hours (three times the risk of that of people with lower values). In this group, however, the rate of decompensation to overt diabetes is still only 3% per year.
Studies of treatment of patients with IFG or IGT with oral antidiabetic (hypoglycemic) agents to prevent or delay the eventual development of diabetes are in progress. Metformin has been shown to have a modest effect (see Prevention of Diabetes Mellitus).
However, IFG and IGT are generally considered to be risk factors for cardiovascular disease (21). Moreover, even the level of fasting blood glucose within the normal range clearly modestly predicts cardiovascular death in nondiabetic men (24). There is no risk of microvascularcomplications (retinopathy or nephropathy) in people with IFG or IGT unless they develop diabetes.
Prevention of Diabetes Mellitus
Prospective studies show that altering “lifestyle,” by modifying diet and increasing physical activity, can prevent or delay the development of diabetes in high-risk patients showing IGT (25). Reduction of progression from IGT to diabetes ranged from 31% to 58% over 3 to 6 years. Whether similar results could be obtained under nonstudy conditions is problematic. Modest weight reduction (mean: about 8.8 lb [4 kg] or 5% to 7% of body weight) and exercise (widely variable but about 2 to 4 hours per
P.1299
week) seem to be the most important contributors to success. Other studies have identified consumption of fiber from cereals and a low glycemic index as contributors to prevention. Total fat intake and type do not seem to relate to the development of diabetes.
A large multicenter randomized prospective study in the United States (Diabetes Prevention Program) included a group who were treated with metformin but who did not receive the other interventions (26). These individuals showed a more modest decrease in the rate at which diabetes developed over the 3 years of study. Metformin is not approved for such preventive use.
Treatment of Diabetes Mellitus
Patient Education
For all patients with diabetes, the following factors are important: the impact of diet and patterns of eating on diabetes, the implications of having diabetes on ordinary activities and of ordinary activities on diabetes, recognition of the signs of worsening diabetes, the importance of proper foot care and of regular eye examinations, and the clarification of misconceptions about diabetes. For patients receiving insulin, the following additional factors are important: correct administration and timing of insulin injections, the unique constraints that insulin therapy places on dietary management and changes of activity, recognition of the symptoms of hypoglycemia, and adjustment of insulin dosage during intercurrent illness.
Patient responses to being informed of a diagnosis of diabetes varies widely. Many patients already suspected the diagnosis as the result of previous observations of similar symptoms in family members. These patients are often aware of the complications of the disease (loss of vision, amputations) and the use of the needle (insulin self-administration). Transient or even prolonged anxiety or depression is common and should be anticipated by the caregiver. Similar problems at this time are commonly seen in close relatives or friends of the patient. Effective approaches to assisting patients with coping strategies are described in Chapters 4 and 20.
Many patients are reluctant to accept the need for self-injection of insulin, and many clinicians are unwilling to press the issue. The result is poor control, inappropriate use of oral hypoglycemic drugs, or both. Reluctance of both patient and clinician may stem from unfamiliarity with the techniques of insulin injection. In fact, insulin injection is simple and almost without discomfort. A firm attitude on the part of the clinician and input from nurses and, if necessary, other patients can overcome patient reluctance in almost all cases. The use of disposable syringes has eliminated the inconvenience of sterilization, and modern thin, very sharp, plastic-hubbed, or syringe-attached needles render the injections practically painless. Aspects of technique are described under Insulin Therapy.
A substantial proportion of newly diagnosed patients have difficulty making the behavioral changes required for optimal management of their illness. Hence, a multifaceted approach is required, which often involves a diabetes educator, a nutritionist, the primary caregiver, and an endocrinologist. The ADA is a useful resource as well for educational material (see http://www.hopkinsbayview.org/PAMreferences). General principles and strategies for educating, motivating, and empowering patients and for helping them make desired behavioral changes are discussed in Chapter 4.
Diet Therapy
Different diet strategies guide therapy for diabetes, depending on whether one is dealing with an obese patient with type 2 diabetes or a patient of appropriate weight who has type 1 disease. For the obese patient with type 2 diabetes, the immediate and long-term goal is weight reduction (see Chapter 83). In type 1 patients, timing of meals must be matched to the administration of insulin to prevent excessive postprandial hyperglycemia and to avoid hypoglycemia. In type 2 patients, timing of meals is still important, whether the patient is using insulin or oral hypoglycemic agents. Diet composition is shown in Table 79.4.
Most obese patients with type 2 diabetes are not severely symptomatic and do not require immediate therapy with insulin or oral hypoglycemic agents for control of symptoms; rather, diet therapy is instituted for correction of hyperglycemia and weight. The blood sugar may fall rapidly on initiation of a diet (i.e., within a few days). This effect is caused by caloric restriction and occurs before significant weight loss is seen. Oral agents, if used along with diet, have the advantage that they do not usually produce severe hypoglycemia, but simultaneous institution of a weight-reduction program and treatment with insulin can lead to hypoglycemia and must be done cautiously. No attempt at tight control should be made until active efforts to lose weight have ended. Some, perhaps 10%, of type 2 patients are not overweight. Such patients should not be advised to lose weight.
Population studies indicate that most type 2 diabetes is either made manifest by obesity in genetically predisposed persons or is actually caused by obesity. Overt diabetes in obese patients is potentially preventable or can be ameliorated by weight reduction; sometimes loss of even 5 or 10 pounds (4.5 kg) has a salutary effect. However, most patients are unable to achieve or maintain a weight that will reverse overt diabetes. In one study, a group of patients with IGT were shown to lose weight over a 1-year period using a reduced-fat diet (average weight loss 7.3 lb [3.3 kg]);
P.1300
this reduced the number of patients who progressed to diabetes. Over the following 4 years, however, weight was generally regained to baseline and glucose tolerance deteriorated (27).
|
TABLE 79.4 Distribution of Major Nutrients in Diabetic Diets (United States) |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||
A guiding principle for formulating diabetic diets should be the recognition that individual food preferences must be respected whenever possible. The dietitian should obtain the patient's preferred dietary history and then should attempt to construct the diet around these preferences. Such an approach is demanding for the dietitian, but the issuance of a standardized “American” diet to a patient from an ethnic minority who has diabetes is unlikely to be helpful. Chapter 83 on obesity and Chapter 4 on patient education deal with these principles in greater detail.
Prevention of Atherosclerosis
A goal of diet therapy, beyond weight reduction, is the prevention of atherosclerotic disease. This problem is both more prevalent and accelerated in all types of diabetes and accounts for approximately 25% of deaths among patients with type 1 diabetes with onset before age 20 years. Without treatment, adults with type 2 diabetes are two to four times more likely than those in the general population to die from coronary artery disease. A large portion of this excess mortality is undoubtedly caused by the abnormalities of lipids that are so common in diabetes mellitus.
The evidence that atherosclerosis in the patient with diabetes may be preventable is based to a large degree on comparisons of the prevalence of atherosclerotic disease in different populations with widely varying diets (28,29). The diabetic subjects in the United States who followed conventional, widely used, high-fat, low-carbohydrate diabetic diets—at least until about 1970—had the highest rate of coronary disease seen anywhere in the world (three times the rate of the general population). For this reason and because of the evidence from population studies, the ADA recommended that its old standard diabetic diets should be abandoned. Ironically, no strong evidence exists to support the notion that only the high-fat diets used for diabetes were responsible for the high rate of coronary atherosclerosis, although they sometimes contained up to 70% of calories as fat. Type 2 diabetes is now recognized to be part of a syndrome that includes hyperlipidemia and a propensity to the development of atherosclerosis, probably regardless of diet (see the discussion of the metabolic syndrome under Definition and Classification). Currently, the ADA recommends that diet should be individualized to accommodate individual preferences along ethnic lines, but also advocates a diet high in carbohydrates (50% of calories) and relatively low in fat (30%), a diet similar to that recommended by the American Heart Association for people without diabetes (Table 79.4).
The ADA nutritional recommendations are in general, but not completely, followed in this chapter. For details, the reader is referred to the ADA's position paper (30). Although such diets have been successfully used in diabetic patients studied on metabolic wards, their reported beneficial effects on blood lipids may result from other factors: control of caloric intake with concomitant weight reduction, very low cholesterol content, high fiber content, and absence of sucrose. Several studies in which patients with type 2 diabetes received such diets resulted in unchanged low-density lipoprotein (LDL) cholesterol, lowered high-density lipoprotein (HDL) cholesterol, and increased triglyceride levels (31,32), as well as in accentuated postprandial lipemia with accompanying potential for increased atherogenicity. In contrast, another study used a high monounsaturated fat diet (50% of calories) low in carbohydrates (35%) that resulted in improved PG, triglycerides, and HDL cholesterol when compared to the ADA diet (33). A later report comparing a high-carbohydrate (60% of calories) diet to a high-fat diet (40% of calories) rich in monounsaturates saw no differences in lipid profiles between these regimens (34). It is unclear whether
P.1301
seemingly small differences of carbohydrate and monounsaturated fat are critical or whether other factors are involved in this apparent discrepancy. The optimal diet for the control of blood lipids and glucose in the patient with diabetes must still be considered unsettled, but it is reasonable currently to recommend up to 50% of calories from fat, provided that it is high in monounsaturates, and 35% of calories from carbohydrates.
Role of Alcohol (Ethanol)
Objective discussion of the role of moderate amounts of alcohol (ethanol) in the diet is confounded by cultural, social, and religious considerations, and by concern for potential abuse and addiction. Moderate alcohol use in men is associated with decreased development of type 2 diabetes (35) and in women with diabetes, with a reduced risk of atherosclerotic heart disease (36). Therefore, given the high risk of atherosclerosis, it is reasonable, until evidence to the contrary is presented, for clinicians to tolerate moderate amounts of alcohol in the diet, especially if the patient is already a user of alcohol. In people without diabetes, two to three drinks per day for men and one to two per day for women seem to be optimal for reducing cardiovascular risks, although the dose–response is difficult to define (37). On the other hand, alcohol, especially when ingested in the fasting state, can readily produce hypoglycemia in both nondiabetic and diabetic individuals. Diabetics receiving oral agents and/or insulin are probably especially vulnerable to this effect of alcohol. Although alcohol generally increases HDL, which appears to be at least part of its mechanism in preventing cardiovascular disease (see Chapter 82), it induces hypertriglyceridemia. If lipid control proves difficult with ordinary therapy in a particular patient, attention should be paid to a possible contributing role of excessive alcohol consumption, which may be unreported. Alcohol can also induce or worsen hypertension, another major problem in diabetes.
Role of Hyperglycemic Control in Control of Hyperlipidemia
Before hypolipidemic drug therapy is considered, efforts to control hyperlipidemia (especially hypertriglyceridemia) in diabetic patients should include at least an attempt at near normalization of the fasting blood glucose (along with initiation of an appropriate diet [Table 79.4] for at least 3 months). At present, no evidence is available that favors insulin or oral hypoglycemic drugs to achieve this goal in type 2 diabetes, although a theoretical advantage for glipizide has been suggested (38). Treatment of coexisting diseases that can cause hyperlipidemia (e.g., hypothyroidism) is also necessary.
It is inappropriate and usually ineffective to introduce hypolipidemic drug therapy (see Chapter 82) if hyperglycemia is not controlled. However, gross elevations of triglycerides (above 1,000 to 1,500 mg/dL) can predispose to acute pancreatitis. Early institution of drug therapy (gemfibrozil or fenofibrate) is indicated under these circumstances (i.e., even before glucose control is achieved).
Glycemic Index
The magnitude of the increase in blood glucose in the 3 hours after a meal is determined not only by the carbohydrate content of the meal but by the carbohydrate type(s) consumed. By comparing the percent increase of the blood glucose to a reference food, usually white bread or potato with the equivalent carbohydrate content, an index can be calculated to predict the glucose elevating effect of the food. Diets incorporating foods with low glycemic indices do reduce postprandial hyperglycemia. When combined with a high fiber content, which probably acts by slowing absorption, meals with low glycemic indices can be beneficial in terms of the total period of postprandial hyperglycemia. Adopting an optimal diet in terms of the glycemic indices and content of fiber may be quite effective and even comparable with the effect of an oral diabetic drug in a type 2 diabetic, but this approach requires a highly motivated patient and the services of a skillful dietitian (39).
Fiber
It is now recommended that most of the carbohydrate in the diet of both type 1 and type 2 diabetics be in the form of high-fiber foods (fruits and vegetables, especially legumes). The ADA currently recommends 25 g of such foods should be eaten each day, but there is evidence that 40 g per day may be preferable (40). A high-fiber diet results in lower mean blood glucose levels and may allow the administration of lower doses of hypoglycemic agents. There is often a modest reduction in LDL concentrations as well (see Chapter 82). In many patients these diets produce a variety of unpleasant side effects, including increased frequency of stools, diarrhea, abdominal pain, and flatulence. The formulation of fiber-rich diets is difficult, and most patients do not accept the major alterations of diet that are necessary to produce the desired effects on blood glucose levels.
Estimation of Caloric Needs
Caloric requirements for maintenance of weight vary considerably from person to person and are influenced by activity level. Required calories are approximately 40 kcal/kg or 20 kcal/lb per day for an adult with normal activity. Thus, a person weighing 70 kg may require 2,800 kcal, although some lean men performing ordinary activities may require
P.1302
as much as 3,000 to 3,500 kcal per day. Individuals who perform manual labor may need 4,000 or more kcal per day, whereas sedentary people may need only 2,000 kcal, or less, per day.
In prescribing diets, caloric requirements are often underestimated. Clinicians commonly prescribe a 1,800-kcal diet for maintenance even if it is grossly inadequate for a particular patient's caloric needs. Prescription of such a diet leads to frustration and noncompliance. Overzealous decreases of calories for weight reduction may be equally defeating. When maintenance of weight is the goal, a careful dietary history by a skilled dietitian may be a good starting point for establishment of a patient's needs; the prescribed diet should then become simply a modification of that patient's ordinary pattern.
Diet during Conventional Insulin Therapy
Any patient who receives insulin faces a special problem. Unlike patients who are not receiving insulin and whose total intake can vary from day to day, patients who are receiving standard therapy with insulin require fixed patterns of food intake; greater flexibility is possible with intensive therapy (see Insulin Therapy). Total caloric intake must be distributed among the meals of the day, which usually include bedtime snacks. Occasional patients strive to reduce insulin dosage by senseless restriction of intake, incorrectly reasoning that disease severity will somehow be less if they can treat their diabetes with less insulin. Needless to say, they must be dissuaded from such practices.
The exact composition of the diet for the patient with type 1 diabetes is less important for blood sugar control than is the constancy of distribution of the amount of food at each meal from day to day. Insulin effect (duration, intensity), even for a particular type of insulin, varies from patient to patient. Accordingly, avoidance of extremes of blood sugar concentration (hypoglycemia and hyperglycemia) requires some adjustment of food apportionment for each patient. However, one should attempt to simulate as closely as possible the patient's usual and preferred pattern of food intake. The main modification is usually to add between-meal snacks. Once an acceptable food pattern has been established and insulin dosage adjusted to that pattern, the patient must adhere to the program if extremes of blood sugar are to be avoided. Patients learn by trial and error how much latitude they can tolerate. Problems, not easily solved, are encountered in individuals who engage in strenuous sports or work that varies from day to day. Such people may have to eat more on some days than others or make frequent adjustments of their insulin dosage. Intensive therapy (see below) actually allows for better control of blood sugar and greater variation in diet and in the level of physical activity.
Exchange Lists and Special Foods
After a dietitian estimates the constituents that will be acceptable to a patient, joint discussion should be held with the spouse or other involved family members. Cooperation and participation of a spouse in the process may be essential for successful adaptation, which, for practical reasons, may require that both partners participate in the diet modifications.
The intelligent use of diet exchange lists (food equivalents) is helpful for many patients. Such lists are available from the ADA, the American Dietetic Association, and most hospital dietetic units.
Special diabetic or dietetic foods are expensive and usually are unnecessary. Some such foods do contain less simple sugar than is ordinarily the case, but sucrose has no worse a glycemic index than bread or potatoes. The patient must read the labels carefully to avoid deception.
Exercise as Therapy
Historically, exercise was recommended for control of hyperglycemia as part of a basic program of diet and insulin for type 1 patients, but conclusive data indicating such a benefit are not available. Although diabetics, like others, derive health benefits from regular exercise, the complexities of avoiding hypoglycemia during exercise in type 1 patients makes blanket recommendations tenuous (see Exercise During Insulin Therapy). The case for exercise is better in type 2 diabetes. Regular exercise may actually help prevent the emergence of type 2 diabetes (41). The conditioning effect of regular exercise also decreases insulin resistance and can improve hyperglycemia. In patients with type 2 diabetes who receive sulfonylurea drugs, the likelihood of provoking hypoglycemia by an exercise program is not great, but obese patients on low-calorie weight-reduction diets who exercise at high intensity may be severely limited by lack of muscle glycogen unless they consume additional carbohydrates immediately before exercising. Walking or cycling may be the least-threatening form of exercise for these patients; attention to a period of adaptation is vital. All patients must avoid exercise that aggravates latent or existing problems (e.g., foot trauma that can lead to ulceration). It should be assumed that patients with long-standing diabetes who may have occult cardiac disease must be especially cautious when initiating an exercise program. A stress electrocardiogram is a prudent but minimal measure in such patients (see Cardiovascular Problems). The clearest rationale for exercise as therapy is as an adjunct to weight-reduction programs. With weight loss, sensitivity to endogenous insulin may be restored in obese insulin-resistant patients and normoglycemia may ensue, sometimes obviating the need for oral agents or insulin. The blood sugar-lowering effect of exercise often antedates significant weight loss. Approaches to exercise
P.1303
therapy in healthy people and in patients with heart disease are described in Chapters 16 and 63, respectively.
Selection of Patients for Insulin or Oral Therapy
Type 2 Diabetes
Many patients with type 1 diabetes are started on insulin during an episode of ketoacidosis. The insulin dependence has been established by the occurrence of the acute episode. Unless this acute event was precipitated by stress in a type 2 diabetic patient, insulin dependence is usually absolute and permanent. Occasionally in adults (more often in children), the insulin requirement may decrease or even disappear over several months, but relapse is the rule in such cases. Many nonobese adults were in the past considered to have type 2 disease when first diagnosed because they had not developed ketoacidosis on presentation. It is now understood that such patients may have LADA (seeType 1 Diabetes Mellitus and Its Slowly Progressive Form). Conversely, patients presenting with ketoacidosis and thought to have type 1 may prove to have type 2 diabetes. The presence of type 1 diabetes can be suspected from the lack of obesity, lack of response to sulfonylurea, an unequivocally low level of insulin C peptide (a marker for endogenous insulin secretion), and appropriate antibody determinations (anti-islet cell and antiglutamic acid decarboxylase [anti-GAD] antibodies).
Type 2 Diabetes
There is no direct relationship between the level of glycemia and the type of pharmacologic therapy to be used. Although it is true that the magnitude of the hyperglycemia suggests which patients are likely to respond to oral agents, the glycemic goal of therapy should determine the selection of the agent(s) to be used. Obviously, if a single oral agent can achieve the glycemic objective, it is the simplest if not always the least expensive route. Other factors, including age of the patient, life expectancy, and the presence of comorbid disease(s) (such as renal disease or dementia), are important considerations in selecting a therapeutic regimen. As noted, the initial approach to the obese non–insulin-dependent patient should be caloric restriction and weight reduction. Such therapy, if successfully followed, can be expected to reduce if not normalize the blood sugar within a few weeks. However, significant caloric restriction may induce a marked fall in the blood sugar within a few days, well before significant weight loss has occurred. If FPG is less than 200 to 250 mg/dL, hyperglycemia and glucosuria will not ordinarily produce enough symptoms to be troublesome during this period and no additional drug therapy (oral hypoglycemics or insulin) is needed. Even an FPG of 300 mg/dL may be tolerated. These patients are not prone to ketosis; no urgency exists for instituting drug therapy. On the other hand, symptomatic hyperglycemia or glucosuria, persisting for weeks despite efforts at (or actual) weight loss, should not be ignored. In this case, drug therapy is indicated for symptomatic relief of polyuria and thirst and can be discontinued if weight reduction is successful. Most patients with symptomatic type 2 diabetes are treated initially with oral hypoglycemic drugs, although some require insulin or a combination of oral agents or oral agents plus insulin.
The imperative for use of insulin in patients with asymptomatic type 2 diabetes, whose blood sugar cannot be controlled with oral agents is no longer in question. The evidence is clear that modest elevations of blood sugar do indeed relate to at least the microvascular complications of diabetes. The United Kingdom Prospective Diabetes Study (UKPDS) data (42) plus other studies and experimental observations have settled this issue. Patients with type 2 diabetes who do not respond to initial treatment with oral hypoglycemic agents are termed primary failures. Others, adequately controlled by oral hypoglycemic drugs for a time, become unresponsive to these agents (secondary failures). Insulin therapy may become essential in such cases. Other patients with type 2 disease develop grossly uncontrolled hyperglycemia during stress (trauma, infection, surgery, glucocorticoid therapy). Whether or not ketosis ensues, the gross hyperglycemia may produce severe osmotic diuresis and its sequelae. Such patients require control of hyperglycemia with insulin therapy, which may be discontinued as soon as the situation warrants. Occasionally adults, usually not obese and not necessarily exhibiting much glucosuria, may exhibit unexplained weight loss and lack of well-being. Such patients may show dramatic improvement with insulin.
Determination of glycosylated hemoglobin (HbA1c) in cases of modest elevation of blood sugar is an important guide to therapy. Using the best available methodology for measurement, the normal mean HbA1c is approximately 5% and the upper limit of normal is 6.5% (3 standard deviations). A near-normal value (e.g., below 7.0%) might deter a recommendation for drug or insulin therapy, whereas an elevated value would suggest that long-term benefit might outweigh the possible risks or inconvenience of treatment. Recommendations for or against therapy under these circumstances are currently determined not only by the clinical circumstances, including lipid abnormalities, but by the long-term deleterious effect of hyperglycemia.
Special Considerations in the Treatment of the Geriatric Patient
Many elderly patients are best treated with oral agents. Simple symptomatic therapy may be the foremost consideration for these patients. Insulin therapy may present
P.1304
special problems for elderly diabetics (because of poor vision, poor manipulative skill, or cognitive decline) that make self-administration of insulin difficult. On the other hand, many elderly patients can manage insulin therapy, especially of the type that is not excessively aggressive, and age alone should not deter the clinician from instituting insulin therapy. As noted earlier, insulin syringes can be prefilled and stored in the refrigerator for 1 to 2 weeks; this plan is useful for the older person who cannot accurately draw up the correct amount of insulin. The elderly are especially likely to have multiple diseases and to use multiple drugs. The risk of drug interactions in this group is therefore greater than in younger people; insulin therapy avoids this problem. A study of various insulin regimens in the elderly, as well as of insulin–sulfonylurea combinations, concluded that twice-daily insulin administration was the simplest, most effective, and most cost-effective regimen (43).
Normoglycemia as a Goal of Therapy
Type 1 Diabetes Mellitus
The results of a National Institutes of Health-sponsored multicenter study, the Diabetes Control and Complications Trial (DCCT) were released in 1993 (44). The study's definitive results have profoundly altered the goals of clinical practice in patients with type 1 diabetes and have provoked new efforts to control blood glucose in type 2 diabetes. The DCCT enrolled only patients with minimal evidence of complications at entry, and the beneficial results were striking. The results of this trial were summarized in a Position Statement by the ADA (45):
The Diabetes Control and Complications Trial (DCCT) [was] designed to test the proposition that the complications of diabetes mellitus are related to elevation of the plasma glucose concentration. The study design was simple. Two groups of patients [with Type 1 diabetes mellitus, all younger than 30 years of age] were followed long term, one treated conventionally (goal: clinical well-being; called the standard treatment group) and another treated intensively (goal: normalization of blood glucose; called the intensive treatment group). The intensive treatment group was clearly distinguished from the standard treatment group in terms of glycated hemoglobin levels and capillary blood glucose values throughout the duration of the study. Normalization of glucose values was not achieved in the intensively treated cohort as a group because mean glucose values were ~40% above normal limits. Nonetheless, over the study period, which averaged 7 years, there was an ~60% reduction in risk between the intensive treatment group and the standard treatment group in diabetic retinopathy, nephropathy, and neuropathy. The benefit of intensive therapy resulted in a delay in the onset and a major slowing of the progression of these three complications. Finally, the benefits of intensive therapy were seen in all categories of subjects regardless of age, sex, or duration of diabetes.
A computer model that used the DCCT data projected considerable gains for patients with type 1 diabetes who maintain over their lifetimes a near-normal blood sugar: an extra 5 years of longevity, 8 years of sight, and 6 years’ delay of renal failure, amputations, and neuropathy. The costs of the required intensive therapy were two to three times as much as those for conventional therapy. The DCCT is the longest and largest, although not the only, prospective study showing that lowering blood glucose concentration slows or prevents the development of diabetic complications. As such, it has major therapeutic implications for healthcare providers and their patients.
A primary treatment goal in type 1 diabetes should be blood glucose control at least equal to that achieved in the intensively treated cohort of the DCCT. This goal may not apply to all patients with type 1 disease and its pursuit must be based on sound clinical judgment. Of importance, intensively treated patients had a threefold greater risk of hypoglycemia than did patients in the control group. Because serious hypoglycemia is dangerous and is not entirely avoidable, the goal of near normalization of blood sugar may, after an initial effort, have to be abandoned for some patients.
There is no favored form of treatment to achieve tight control of blood glucose levels in type 1 diabetes. However, the goal is certainly not achievable in type 1 diabetic patients by use of a single-dose, or even a two-dose, insulin regimen. The decision to use multiple injections of insulin versus an insulin pump depends on patient preference and the ability of the healthcare team to provide the necessary resources and support, but even with these regimens normoglycemia can be achieved in only about half of the patients. The improvement seen in some of the others may, nonetheless, be worthwhile.
Young patients with type 1 diabetes in the early years of their disease stand to gain the most from normalization of blood sugar (tight control) because prevention of complications is the goal. Patients with type 1 disease who already have advanced complications of diabetes will not benefit at all because such complications are irreversible and probably cannot even be stabilized.
At what point in the course of type 1 diabetes should clinicians consider initiation of intensive therapy? After an initial period of conventional therapy for several months, the issue of intensive therapy should be considered and discussed with patients who are suitable candidates. In those to whom tight control is suggested, the clinician must explain the current view that maintained normoglycemia
P.1305
prevents the long-term complications of diabetes mellitus. The magnitude of the effort that is necessary to maintain normoglycemia must also be explained, including the need for self-monitoring of blood glucose (SMBG). One of the frequent-dose intensive insulin therapy schemes (basal-bolus or multidose) or its alternative, infusion pump delivery of insulin, must also be presented. If the patient understands and accepts the problems and effort required, the clinician may consider a program of tight control. However, serious consideration should be given to referral of the patient to an endocrinologist familiar with such a program because the process is difficult, very demanding of the clinician's time, and usually requires a team approach using a specially trained physician's assistant or nurse. The demands on the patient and the clinician are greatest at onset of intensive therapy. When is conventional glycemic control rather than intensive therapy appropriate in type 1 diabetes? Often intensive therapy proves to be less than intensive, regardless of the initial intent. Conventional therapy attempts to achieve near normalization of fasting PG, as opposed to near normalization of blood sugar throughout the day. Even this degree of control is simply not possible using conventional therapy. Many clinicians, failing to realize the limitations of one- and two-dose schedules, nonetheless still go through an agonizing trial of conventional therapy with such patients, only to have the effort end in failure and produce frustration for all involved. In such futile efforts, several types and mixtures of insulin are often tried along with both one- and two-dose schedules. At this point, the options include acceptance of a simplified treatment scheme that merely avoids excessive symptomatic glucosuria with resultant symptoms and prevents development of ketoacidosis (minimal therapy) or reconsideration of intensive therapy, perhaps under the direction of a specialist team.
Type 2 Diabetes Mellitus
The major conclusion of the DCCT—that control of hyperglycemia prevents microvascular complications in type 1 diabetics—has since been proven valid for type 2 diabetics as well. The UKPDS, the largest of the studies of type 2 diabetes, reported on 5,102 patients studied for an average of 10 years (42,46,47). The UKPDS produced some clear results. With intensified therapy with insulin, a sulfonylurea, or metformin, a decrease of HbA1c was accompanied by a decrease of microvascular complications (retinopathy, nephropathy, and possibly neuropathy) of 25%. For every percentage point decrease of HbA1c (e.g., from 9% to 8%), there was a 35% decrease in the risk of these complications.
Thus, young or even middle-aged patients with type 2 diabetes stand to gain as much from tight control as do young patients with type 1 disease. On the other hand, in patients whose life expectancy is limited by age, complications of diabetes, or concurrent disease, the problems associated with intensive insulin therapy should temper the clinician's approach to glucose control.
A major problem in type 2 diabetes is macrovascular (atherosclerotic) disease. Whereas hyperglycemia per se may directly contribute to the development of atherosclerosis, the clinical evidence is weak. Although the DCCT trial produced a favorable trend for type 1 diabetes, the data were not statistically significant. In fact, DCCT was so small and was conducted in such young patients (younger than age 30 years) that it could not have been expected to yield useful information on slowing atherosclerosis through control of glycemia. The UKPDS did not demonstrate a significant effect on the development of cardiovascular complications except in a subgroup of obese patients treated with metformin (47).
Oral Therapy
Obese type 2 diabetic patients who have not responded to a weight reduction diet within 3 to 4 months or who, having started on a diet, need interim symptomatic relief from hyperglycemia that is producing osmotic diuresis (polyuria, polydipsia) may benefit from an oral hypoglycemic drug. Typically, these patients are older than age 40 years and are more likely to respond if their diabetes has been present for only a few years. Other candidates are those who are unwilling to accept insulin therapy or in whom the risks of insulin-induced hypoglycemia seem unacceptable. The latter might include patients with occupations involving hazardous conditions (vehicle or dangerous equipment operators). Still others include nonobese patients in whom insulin therapy is unacceptable but for whom persistent hyperglycemia is a risk factor for microvascular disease.
Oral agents should not be prescribed for patients with a history of ketoacidosis, unless the latter has developed in relation to stress. Available oral agents may be problematic in patients with severe cardiac, hepatic, or renal disease, although the correct choice of an agent may make such therapy possible.
When selecting oral hypoglycemic agents, consideration should be given to the onset of action of the different medications and the need to ameliorate symptoms. In general, if patients are symptomatic, consideration should be given to using insulin secretogogues (sulfonylureas), which have a shorter onset of action, versus metformin or glitazones, which are slower in onset of action.
Transfer from Insulin to an Oral Agent
Type 2 diabetics receiving insulin can be abruptly switched to an oral agent, provided that they do not need more than 40 units of insulin a day. Patients who require
P.1306
such large dosages are unlikely to respond well to an oral agent. Patients with a history of ketoacidosis are ordinarily not candidates for a transfer from insulin. If the patient has manifested ketosis in the past (e.g., during stress) but is otherwise thought to be a candidate for a switch to an oral agent, the dosage of insulin may be cut in half as the drug is started. Subsequent monitoring over the next few days will show whether the oral agent can control hyperglycemia or must be abandoned. A history of hyperosmolar nonketotic coma does not preclude a successful change from insulin. A patient with no tendency to ketosis but whose diabetes is so severe that it has produced weight loss might not respond to an oral agent given as initial therapy but might respond after hyperglycemia has been controlled for a short time with insulin.
Sulfonylureas
Within a few years after their introduction nearly 50 years ago, sulfonylureas came into wide use for the treatment of type 2 diabetes. The acute hypoglycemic effects of the sulfonylureas appear to be mediated through insulin release. However, in chronic administration, during which blood glucose has been lowered or even normalized, no increase of plasma insulin is apparent. Studies of the mechanisms of action of these drugs show both an increase in the number of insulin receptors and a potentiation of insulin action.
Although the University Group Diabetes Program (UGDP) study, a multicenter study published in 1970 suggested that tolbutamide was no more effective than placebo and might even increase the risk of death from cardiovascular disease, that study is now considered flawed (48). The recent UKPDS followed a much larger number of patients than did the UGDP and established unequivocally the value of sulfonylureas in the treatment of type 2 diabetics (42,44,45).
Effectiveness
In optimally selected patients, about one-half can be expected to experience normalization of fasting blood sugar and about one-third do not respond. In others, some drug effect is evident, perhaps to a degree that permits symptomatic relief. Maximal drug effect can be expected within a few days to a week. Those who do not respond during initial therapy are considered to be primary sulfonylurea failures. In other cases, after a month or more of good response, the drug seems to become ineffective (secondary sulfonylurea failure). The frequency of this response has been estimated at 3% to 10% per year. Some apparent secondary failures are in fact caused by noncompliance. Only rarely in secondary failure is a switch from a maximal dosage of one sulfonylurea to another successful.
Dosing
In initiating therapy, an average dosage is usually appropriate. Hypoglycemia can occur and may be both severe and protracted, especially in the elderly and in patients with decreased hepatic or renal function. A single adjustment upward or downward by a factor of 2 may then be made as indicated by the blood sugar response after a suitable interval. When switching from insulin to an oral agent, initial dosage can usually safely be at the maximum recommended level for the particular oral agent or at least in its mid-dosage range. Before switching from one sulfonylurea to another, or from one type of oral agent to another, the first should have been tried at a maximal recommended dosage for at least 1 week; trials of more than 2 weeks are not indicated. In changing from a first- to a second-generation sulfonylurea, similar time intervals pertain.
Choice of Sulfonylureas
For patients with normal hepatic and renal function, there is little to lead one to choose among the first- or second-generation agents (Table 79.5) except for cost and convenience of dosing in that the longer-acting drugs do not need to be taken as often. The frequency of toxicity with any of these drugs is very low. Chlorpropamide is less frequently used than it once was; it should never be taken at a dosage greater than 500 mg per day, above which hepatic toxicity becomes common and additional therapeutic effect is not seen. Because of the ability of chlorpropamide to produce a syndrome of drug-induced water intoxication, this drug should be avoided in the elderly, in whom this effect has been seen almost exclusively. Second-generation sulfonylureas have been in wide use in the United States and abroad for many years.Glyburide (Micronase, DiaBeta) and glipizide (Glucotrol) are both safe agents. Although all sulfonylureas improve the second phase of insulin secretion, claims have been made that only glipizide, in response to glucose stimulation, improves both first- and second-phase responses. The second-generation sulfonylureas are more potent on a per milligram basis. Slow-release forms of glipizide (Glucotrol-XL) and glyburide (Glynase PresTab) are available and may help with patient compliance. Glimepiride (Amaryl) is also available in generic form. Glibenclamide and other sulfonylureas are widely used outside the United States.
Comparative Cost
At present, the approximate monthly retail cost of therapy with these drugs has a wide range, depending on the dosage and the agent used. All the drugs are currently available in their generic forms at one-half to one-third the price of the trade name products.
Instruction to the Patient
The obese patient must realize that weight reduction is the mainstay of therapy and is not simply a general
P.1307
health measure; weight loss has a specific beneficial effect in diabetes. Drug therapy is an adjunct, not a substitute, for weight reduction. The possible risks and goals of therapy should be clearly outlined. Although hypoglycemia is uncommon with the sulfonylureas, when it does occur, it is likely to be both severe and prolonged. The symptoms of hypoglycemia should be clearly described to the patient and to whomever is in close contact with the patient, usually family or friends, and corrective measures outlined and understood. The possibility of drug interactions should be mentioned lest another clinician prescribe a drug that potentiates or decreases the effectiveness of the sulfonylureas, or vice versa. The sulfonylureas are most effective when administered about 30 minutes before breakfast or dinner.
|
TABLE 79.5 Oral Agents for Type 2 Diabetes |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
The elderly are especially prone to development of severe and prolonged hypoglycemia with use of the sulfonylureas, which may be related, in part, to the decrease of renal function that normally accompanies aging and may be worse in the diabetic. Decreased renal function (glomerular filtration rate, creatinine clearance) is often present in the elderly even when the serum creatinine is normal because creatinine production decreases with age as muscle mass decreases.
Drug Interactions
Various drugs enhance the hypoglycemic action of sulfonylureas, and others decrease their effect. Among the more commonly used drugs, salicylates, some sulfonamides, and warfarin all enhance the hypoglycemic action of the sulfonylureas. Nonspecific β-blockers may mask the hypoglycemia-induced release of epinephrine, thus prolonging and intensifying hypoglycemic reactions. β-blockers may also block insulin release. Clonidine (Catapres), like β-blockers, may mask the signs and symptoms of hypoglycemia. Acute ingestion of alcohol can
P.1308
enhance hypoglycemia; chronic alcohol use accelerates metabolic disposal of sulfonylureas and antagonizes their hypoglycemic action. Sulfonylureas interfere with the metabolism of alcohol and may produce a disulfiram (Antabuse)-like effect. Diuretics (the thiazides, chlorthalidone, and loop diuretics) may produce hyperglycemia even in normal people and antagonize the sulfonylureas. The anticonvulsant phenytoin (Dilantin), another commonly used drug, also has an antagonist action. Numerous other drugs may enhance or negate the effect of the sulfonylureas; equally important, the sulfonylureas themselves produce numerous alterations of drug action. These problems should not be overstated, but the clinician should be aware of these possibilities and interactions, especially in the elderly, who may be receiving many drugs.
Biguanides
Metformin (Glucophage), a biguanide, was approved for use in the United States several years ago, after decades of use elsewhere (Table 79.5). The drug's lowering of blood sugar is probably the result of multiple actions; it does not enhance insulin secretion. Another biguanide, phenformin, was at one time in wide use in the United States, but was withdrawn because of occasional cases of fatal lactic acidosis and other problems. Metformin causes lactic acidosis much less often. The drug is effective as monotherapy, producing a 50- to 60-mg/dL fall of PG (decrease of HbA1c of 1% to 1.5%). Currently, metformin is preferred by many as the initial oral agent over a sulfonylurea, although no consensus exists on this issue. The main advantages over sulfonylureas are that metformin does not produce hypoglycemia and is less likely to be associated with weight gain. The main disadvantage is that nearly 25% of patients cannot tolerate the gastrointestinal side effects of metformin, whereas sulfonylureas are very well tolerated. Data on overall efficacy are scarce; primary failures occur in 10% to 15% of patients and secondary failures in about 5%. Metformin is often used in combination with a sulfonylurea, as an add-on after failure of the latter. If the combination is effective, an attempt to reduce or withdraw the sulfonylurea can be made after a month. After another month, the need for readministration of the sulfonylurea can be determined. Small decreases of LDL cholesterol and triglycerides are common, as is minimal weight loss (2.2 to 6.6 lb [1 to 3 kg]). Metformin is not well tolerated by all patients: Nausea, anorexia, and diarrhea are fairly common side effects. These symptoms can be minimized by starting at a once-daily dose (500 mg) and increasing the dosage at weekly intervals until a maximum dosage of 2,500 mg in divided doses is reached. Lactic acidosis is very rare in younger patients, but if the patient has renal or cardiopulmonary disease with hypoxia, a significant risk is present. Metformin is not metabolized and is disposed of by renal excretion. It should not be used if renal function is decreased (serum creatinine above 1.5 mg/dL). The drug should be used with great caution in the elderly because their renal function is often compromised even when the serum creatinine is normal, the result of age- or disease-related diminution of muscle mass that causes decreased endogenous creatinine production. It is recommended that metformin be discontinued temporarily before procedures that may result in hypotension or impaired renal function, such as surgery or radiologic studies with iodinated contrast agents.
Thiazolidinediones (“Glitazones”)
Pioglitazone (Actos) and rosiglitazone (Avandia) are the currently available drugs of this class (Table 79.5). The term insulin sensitizers has been applied to these drugs, but the term is misleading in that it implies specificity of action. Although their effect on glucose metabolism is to facilitate insulin's action, they also affect many other cellular processes. These drugs appear to be safe, although the first one marketed, troglitazone (Rezulin), was withdrawn from the market because of hepatic toxicity that resulted in some deaths. At present, the two agents marketed in the United States are probably less effective than sulfonylureas or metformin when used as monotherapy. Side effects include fluid retention and plasma volume expansion, a concern in patients with cardiac disease (49). In patients who are at risk for congestive heart failure (CHF), or who have diagnosed but mild or compensated CHF (class I or II symptoms), these drugs can be prescribed in lower doses with careful monitoring for increased symptoms. Thiazolidinediones should not be given to patients with class III or IV symptoms. Other disadvantages include weight gain, a delayed onset of action (1 to 3 weeks), a prolonged time to reach a full effect (4 to 12 weeks), and increases in LDL cholesterol. The glitazones can be used in combination with other hypoglycemic agents, including insulin. Side effects, such as pedal edema, are more apparent when high doses of the glitazones are used in combination with insulin therapy. This does not necessarily preclude their usage with insulin, but should serve as a reminder to discuss this potential problem with patients.
α-Glucosidase Inhibitors
Acarbose (Precose) is a nonabsorbable α-glucosidase inhibitor that acts by inhibition of the enzymes in the mucosal cells of the small intestine that digest complex carbohydrates. As monotherapy it can be expected to produce only a minimal effect on the FPG (15 to 20 mg/dL) and then only in patients with no more than modest hyperglycemia. A greater effect is seen on postprandial than on fasting glucose (a reduction of 30 to 60 mg/dL). Abdominal fullness,
P.1309
flatulence, and, less commonly, diarrhea are the side effects, all of which tend to abate with time. The drug (50 to 100 mg) is taken at the beginning of each meal.
Meglitinides
Repaglinide (Prandin) and nateglinide (Starlix) act by rapidly stimulating the secretion of insulin from pancreatic beta cells by a mechanism different from that of the sulfonylureas (Table 79.5). The drugs are given shortly before meals, have a short duration of action, and are not prone to producing hypoglycemia. Repaglinide and nateglinide have been recommended for treatment of hyperglycemia in the elderly, a group in whom hypoglycemia is especially hazardous. In patients with moderate hyperglycemia, these agents are reported to reduce, on average, FPG by about 60 mg/dL and HbA1c by 1.7%. When substituted for metformin in patients whose response to metformin as monotherapy had become unsatisfactory, repaglinide produced only a minimal response, but when given in combination, the mean decrease in glucose was nearly 40 mg/dL and in HbA1c, 1.4%.
Combinations of Oral Agents
The increasing number of available oral agents and the relative ineffectiveness of monotherapy for control of glycemia have led in the last few years to a proliferation of combination therapies. A significant number of type 2 patients do initially reach glycemic targets with monotherapy but within several years require more than a single drug to maintain satisfactory levels of PG.
Metformin plus a sulfonylurea is probably the most commonly used combination therapy at present. This combination of a “sensitizer” and an insulin secretagogue target the two defects in the pathophysiology of type 2 diabetes mellitus (insulin resistance and defective insulin secretion). Various fixed-dosage combinations of metformin with glyburide (Glucovance) and glipizide (Metaglip) are available. A glitazone plus either a sulfonylurea or metformin (e.g., the rosiglitazone and metformin combination, Avandamet) may also be used. In one study, patients received metformin (2,500 mg) plus 4 or 8 mg of rosiglitazone, all given together once daily. At the end of 26 weeks, patients receiving the combination had a mean FPG of about 180 mg/dL (50). Whereas this was 40 to 50 mg/dL lower than the metformin-only group, only 28% of these patients reached an HbA1c level of 7% or less. These suboptimal results are likely is attributable to the lack of insulin action, and emphasize the need to use a combination of medications that include an insulin secretagogue (a sulfonylurea or either repaglinide or nateglinide) or insulin itself.
Insulin Therapy
Table 79.6 list the insulins sold in the United States today. Until about 10 years ago, insulins were prepared from the pancreas of animals (cattle and pigs). Only a few of these preparations are still available. Most insulin is now made by recombinant methodology and has the amino acid sequence of the human or is modified from that sequence. Thus, so-called human insulin is actually produced in bacteria and purified for clinical use.
Except for the rapid-acting insulins, which are clear solutions, most preparations now in use are suspensions of insulin that have been modified by complexing the insulin with the protein protamine (neutral protamine Hagedorn [NPH]) or precipitated from solution (Lente) to prolong their action by delayed absorption after subcutaneous injection. However, some of the newest preparations have modified sequences so that they are clear solutions with altered durations of action related to their physicochemical properties, which, in turn, has been tailored to ensure delayed absorption (e.g., insulin glargine; see Long-Acting Insulins). The characteristics and uses of these insulins are summarized below.
Rapid-Acting Insulins
Regular Insulin (Crystalline Zinc Insulin)
Regular insulin is a completely dissolved (clear) preparation that has long been used intravenously in hospitalized patients for acute therapy of ketoacidosis. In the treatment of ambulatory patients, regular insulin is used subcutaneously, often in mixtures with other insulins. The onset of action of subcutaneously injected regular insulin is 20 minutes; peak action is at 2 to 4 hours, and the duration of action is 4 to 6 hours. Regular insulin also is used for continuous subcutaneous injection with portable infusion pumps and is increasingly used in combination with Ultralente insulin or insulin glargine in intensive control schemes; the long-acting component provides the equivalent of background activity provided by the basal infusion rate of a pump, whereas additional subcutaneous injections before meals are equivalent to the bolus injections of the pump (Fig. 79.1). Regular insulin is available as mixtures with NPH to provide a more rapid-acting component (Table 79.6).
Insulin Lispro
Insulin lispro, an amino acid-modified recombinant human insulin, when given subcutaneously has a more rapid onset of action (5 to 10 minutes) than ordinary regular insulin and a somewhat shorter duration of action, both resulting from its more rapid absorption. Lispro's main usefulness is in multidose programs involving intensive therapy and in insulin pumps.
|
TABLE 79.6 Insulins Sold in the United States |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
P.1310
Insulins Aspart and Glulisine
These rapid-acting insulins have a similar mode of action as the other rapid-acting insulin analogues. Aspart (NovoLog, Novo-Nordisk) and glulisine (Apidra, Sanofi-Aventis) insulins should be administered subcutaneously 15 minutes before meals, have an onset of action within 60 minutes and duration of action of approximately 2 hours. They are approved for use in insulin pumps.
Intermediate-Acting Insulins
Neutral Protamine Hagedorn Insulin
NPH insulin is a standardized neutral crystalline suspension prepared from an excess of regular insulin and protamine zinc insulin. NPH is the most commonly used intermediate-acting insulin in the United States. NPH exhibits a relatively rapid onset of action and a duration of action that begins to wane after about 12 hours but may last up to 20 hours. When a single injection of NPH is given in the morning, its onset of action usually occurs in the early afternoon, which essentially provides coverage for the midday meal. NPH can be given in the evenings and in multidose schemes described in Initiation of Insulin Therapy. It should be remembered that if NPH is given early in the evening (around 5 p.m.), some patients may experience nocturnal hypoglycemia (especially if the NPH dose is too high) or, alternatively, they may experience fasting hyperglycemia because the effects of NPH have worn off by the next morning. For most patients the achieved effect of a single dose of NPH is inadequate. Nonetheless, in the United States, many physicians continue to prescribe a single daily injection of NPH, a practice that has long been discontinued in Europe, where NPH is almost always given in two doses. Recently, NPH has been increasingly used in combination with sulfonylureas. Mixtures of NPH and regular insulin (e.g., 70:30; Table 79.6) are now marketed and are useful in some patients in controlling the postprandial increase in blood sugar.
|
FIGURE 79.1. Plasma glucose and free insulin levels in patients with type 1 diabetes treated by three methods: closed-loop intravenous infusion (plasma glucose sensor-controlled apparatus), open-loop subcutaneous injections (insulin pump), and multiple subcutaneous injections (intensive conventional therapy). B, Breakfast; L, lunch; S, supper; HS, bedtime snacks. Note that the results are essentially the same with all methods used. (Modified from Rizza R, Gerich JE, Haymond MD, et al. Control of blood sugar in insulin-dependent diabetes; comparison of an artificial endocrine pancreas, continuous subcutaneous insulin infusion, and intensified conventional insulin therapy. N Engl J Med 1980;303:1313 , and Schade DS, Santiago JV, Skyler JS, et al. Intensive insulin therapy. Garden City, NY: Medica Examination Publishing, 1983:138 , with permission.) |
P.1311
Lente Insulin
The Lente insulin series was devised to avoid the use of the foreign protein protamine. Controlled addition of zinc is used to prepare Semilente insulin, an amorphous, rapidly absorbed, and rapidly acting material, and Ultralente insulin, a crystalline product with much slower absorption and longer action. Lente insulin is a mixture (30:70) of amorphous (semilente) and crystalline (ultralente) insulins. Although commonly thought to be equivalent to NPH, Lente is slower in onset and its duration of action is significantly longer, usually exceeding 24 hours. If a single daily dose of insulin is the treatment goal, Lente may be appropriate and a better choice than NPH.
Long-Acting Insulins
Ultralente Insulin
Ultralente insulin of beef origin has a duration of action exceeding 24 hours but is no longer marketed in the United States and has been replaced by the human form. It may occasionally be used alone. Until recently, Ultralente insulin (beef or human) was the backbone of intensive therapy by the basal-bolus technique. The prolonged effect of this preparation provides the basal (background) activity equivalent to that of an insulin pump. It is important to note that unlike the beef product, Humulin U (human recombinant Ultralente) has a duration of action shorter than 24 hours and may not be the best choice for use as a source of basal activity in multidose intensive-therapy schemes, although human Ultralente has been successfully used for this purpose.
Insulin Glargine
Glargine insulin (Lantus), a recombinant long-acting insulin analogue, is a completely soluble preparation that lasts approximately 24 hours. Glargine is comparable in its duration of action to beef Ultralente, but it is more predictable in its absorption than either beef or human Ultralente. Glargine is becoming the agent of choice for basal-bolus regimens. However, it is twice as expensive as Ultralente insulin.
Mixtures of Insulins
A frequent goal for patients receiving conventional insulin therapy is a single injection once daily. This goal is inferior to that of a multidose program for control of glycemia and should only be used if the patient cannot take insulin at least twice daily. If a single dose is to be used, insulin effect must be prolonged sufficiently to produce normoglycemia in the morning and at the same time provide adequate daytime control of the increases of blood glucose that occur
P.1312
postprandially. However, the duration of action of NPH is usually too short for this goal, and Lente, although sometimes a better choice, is often unsatisfactory as well. One of two scenarios is observed. First, the excessive daytime hyperglycemia dictates the need for additional rapid-acting insulin. Thus, regular insulin is added to NPH or Lente. Second, the single injection of NPH or Lente controls daytime hyperglycemia but the total duration of action is inadequate, resulting in hyperglycemia at the beginning of the next day. A predinner or bedtime dose of NPH can also be added to a sulfonylurea regimen (see Insulin Therapy for Type 2 Diabetes in Combination with Other Oral Agents).
Regular and NPH insulins can be mixed in the same syringe in all proportions without affecting the onset and duration of action of the separate components. In the United States, 70:30 and 50:50 mixtures of NPH and regular are now sold; in Europe a series of such mixtures (e.g., 90:10, 80:20) have long been available. Regular and Lente insulins cannot be mixed and allowed to stand for more than a few minutes before injection; delay of injection results in blunting of the action of the rapid-acting regular insulin.
Commercial Insulin Preparations
A number of products are available; Table 79.6 lists the various types, species of origin, and the producers. Clearly, no clinician needs to memorize this ever-changing list. However, even though most clinicians write prescriptions for insulin without specifying the brand, it is important to recognize the product that the pharmacist has dispensed. Not all preparations are available in a given region; pharmacies often supply the products of particular manufacturers according to local profit considerations. At present, two companies market most of the insulin used in the United States: Lilly and Novo-Nordisk. Both produce reliable, clinically comparable products within a particular category. A third company, Sanofi-Aventis, introduced insulin glargine and glulisine insulins.
Most problems caused by impurities in insulin in the past disappeared with the introduction of the highly purified animal insulins several decades ago. Allergic reactions and lipoatrophy were the two most troublesome events; both appear to have been related to impurities and now are rarely seen, except that lispro insulin used in a pump was recently reported to produce lipoatrophy. Purified animal insulin of pig origin (the only animal insulin available now in the United States) is about as expensive as human insulin and is of a comparable degree of purity.
In the past, insulin resistance was often ascribed to antibody formation, but little evidence exists to document significant immunogenic differences or clinical improvement as the result of switching insulins. Human insulin is certainly immunogenic in humans. Some types of human insulin do differ from the products of animal origin by having somewhat more rapid onset and peak of action and shorter duration of effect. Given the variability of onset and duration shown under clinical conditions, these differences may not be particularly important, especially in multidose programs.
Trade Names, Unit Designations, and Syringes
All insulin (Table 79.6), regardless of type or source, is standardized at a specific concentration per milliliter. The symbol U refers to the insulin concentration in units per milliliter. All insulins are marketed at a concentration of 100 U/mL (U100). Human regular (Humulin R, concentrated, Lilly) also is marketed in a preparation that contains 500 U/mL. Although long-term storage is best done by refrigeration, opened vials of insulin may be kept unrefrigerated for up to 1 month. When insulin is used during travel, extremes of temperature should be avoided, as in a sun-exposed automobile or next to a stove or heating element.
Several sizes of syringes are available for use with U100. A 1-mL syringe can be used for all doses up to 100 units, but most accurate dispensing of less than 30 units is made when syringes of 0.5-mL (50-unit) capacity are used. The bores of these syringes are smaller and the scales are consequently expanded. Some patients require more than 100 units for a single injection. For such use, 2-mL syringes (200-unit capacity) are manufactured, but these are in short supply and are difficult to obtain. No syringe is calibrated for use with U500.
The use of disposable plastic syringes with attached needles has greatly simplified use of insulin and is preferred by almost all patients. Many patients reuse disposable syringes without obvious harm, but the practice should be discouraged. Special syringes are available for use by patients with severe impairment of vision that prevents them from accurately measuring a dose. However, a simple solution to this problem is often possible. Disposable syringes can be prefilled with ordinary sterile precautions by an able person (relative, friend, pharmacist) and safely stored in a refrigerator for at least a week.
Insulin Injection Technique
After initial instruction the patient should be observed during self-administration of insulin to be certain that the correct volume is being drawn into the syringe and that the proper injection technique is used. Sterilization of the skin with an alcohol wipe is not necessary, although the injection site should be clean. If the injection is made through skin that is wet with alcohol, unnecessary burning discomfort is produced. Injections with disposable needles are essentially painless. Repeated punctures of the rubber diaphragm of vials of insulin with the same needle dulls the point and leads to painful injections.
P.1313
In ambulatory patients, insulin preparations should always be given subcutaneously. Most needles in present use are one-half inch in length. Unless the patient is very thin, the best technique involves insertion of the needle at an angle of 90 degrees to the skin surface. If the patient is very thin, the needle is 5/8-inch in length, or if the site is covered by thin skin, the needle may be inserted at an angle of approximately 45 degrees so as to avoid intramuscular injection. After injection, the area should not be massaged because that may accelerate absorption.
The choice of injection region is important because the rate of insulin absorption, and, hence, the duration and magnitude of insulin effect, varies considerably between anatomic locations. Absorption is slowest from the thigh, fastest from the anterior abdominal wall, and intermediate from the arm. In addition, absorption from an exercising extremity is accelerated. The long-used technique of rotation of sites is unwise and may contribute to erratic control. On the other hand, the repeated use of precisely the same spot within a region should be avoided. Because of the variability of insulin absorption, injections in the abdomen should be encouraged. The patient should also avoid injecting insulin prior to showering or entering a hot tub or spa. The heat in these settings can accelerate insulin absorption, which can lead to serious hypoglycemic episodes.
Insulin Injection Devices
A variety of devices are available to facilitate the injection of insulin. Button-like injection ports are devices that can be left in place, usually over the abdomen, all day and decrease the number of skin punctures when multiple injections are being given. Needleless injectors that use a high-pressure jet are used by some patients, but they are not always painless, and absorption may be more rapid than with ordinary injections. Fountain pen-shaped injectors use a cartridge containing insulin; the needle does not need to be changed for several days. Delivery is with a push button or a preset dial. These devices are especially useful for diabetic patients taking more than one injection daily, who are eating meals in a restaurant, or who are traveling. The cost of the insulin in the cartridges used with these devices is high.
Initiation of Insulin Therapy
Type 3 Versus Type 2 Patients
Typical type 1 patients are almost always started on insulin during an initial acute episode of ketoacidosis that was treated during a hospitalization, and when encountered on an ambulatory basis, most will be receiving at least two injections of NPH insulin and a total dose of between 30 and 50 units per day. Adjustments of dosage must be made on knowledge of at least premeal and prebedtime SMBG. Type 2 patients who require insulin will usually have received oral agents that have failed to control the blood sugar. At the time of initiation of therapy with insulin, the clinician should establish clear targets for the degree of control of glycemia. Especially in patients with type 1 diabetes, but also in patients with type 2 diabetes in whom tight control is sought, it should be understood that both basal and bolus insulin will be required (see Intensive Therapy).
Initiation of Insulin
There is not a single, standard approach to initiating insulin therapy. If cost is an issue, the older NPH and regular insulins can achieve excellent glycemic control, but will require more education and intervention by the health care team. For instance, NPH needs to be resuspended before injection and should be given twice daily; regular insulin should be injected 30 minutes prior to each meal. If cost is not a major issue, then newer agents such as glargine and the rapid-acting insulin analogues lispro, aspart or glulisine can be used. Glargine does not require resuspension and the rapid-acting analogues can be injected within 10 minutes of eating a meal. Various algorithms are available to calculate the dose of insulin, but one of the simplest is to begin with glargine 10 units subcutaneously once daily, at any time of day. The patient can increase the dose by 2 to 4 units every 3 days as long as the fasting glucose levels remain above 120 mg/dL. Once the patient achieves blood sugars below 120 mg/dL, without experiencing hypoglycemia, the current dose of glargine is maintained. The patient may be instructed to increase the initial dosage of basal insulin by 2 to 4 units every 3 days until satisfactory control is approached. Usually such a program brings the patient under control within a few weeks. Increments of 10 units every 3 days are also safe as long as the patient is not markedly symptomatic or if no effect is apparent within a week. This aggressive approach can be used in patients with suspected insulin resistance. During this time the patient should monitor blood glucose at least twice daily (prebreakfast and predinner). A telephone call to the physician, nurse, physician's assistant, or diabetes educator should be made every week (more often if the patient is insecure), but at least when a single-dose program is being established, the patient should be encouraged to proceed with the dosage adjustments as planned and should not require or expect a physician's instructions at every dosage increment. Unnecessary dependence is thus discouraged, and the patient's involvement in management is enhanced.
Use of NPH Insulin Regimens
Most patients cannot be controlled with a single morning dose of intermediate-acting insulin (NPH or Lente; see Table 79.5). A single dose in the morning will improve hyperglycemia and glucosuria; the late morning or afternoon
P.1314
glucose measurements are the first to show a tendency to normalize. However, the effect of the morning insulin is insufficient to ensure normoglycemia in the fasting state (i.e., early in the next morning). Several maneuvers can be tried: changing the timing of a single dose to the evening, using twice daily dosing of NPH, changing to another type of insulin with longer action (such as glargine), or adding an oral agent.
In the two-dose approach, a predinner or bedtime dose of the same intermediate-acting insulin can be added, sometimes requiring a concomitant reduction of the morning dose. For example, if such a patient is receiving 60 units of NPH prebreakfast daily, up to 15 units may instead be given in the evening—usually before dinner—and the morning dose can be reduced to 50 to 55 units. Additional increments of 5 units may then be made to either dose, depending on whether the fasting or postprandial glucose is too high. The evening dose will have its greatest effect on the fasting glucose. Patients using such a split-dose schedule often receive 40 units or more daily, with 50% to 70% of total daily dose in the morning and the remainder in the evening. A two-thirds morning–one-third evening split is common.
A number of investigators report improved control of hyperglycemia using intermediate-acting insulin (NPH) given in the evening (before dinner or at bedtime) rather than, as conventionally used, in the morning. The insulin can be given alone or added to a sulfonylurea if insulin alone is inadequate. Part of the rationale includes a need for increased insulin action during the night to suppress the normal tendency for hepatic glucose output to increase during the early morning. Even patients with severe obesity are candidates for this type of therapy. Improved metabolic control has been claimed for this approach (51).
Insulin Therapy for Type 2 Diabetes in Combination with Oral Agents
There is not a standard approach to adding an oral agent to insulin therapy with regard to which drug should be used preferentially in combination with insulin. Because the pathophysiology of type 2 diabetes entails two problems, insulin deficiency and resistance, it is now commonplace to use combination therapy in the early phase of drug treatment. If the patient is insulin naive and is already receiving an oral agent at maximal dose, NPH or glargine is generally given at bedtime with the aim of achieving a normal FPG; improved daytime control will usually follow. In one study, the addition of a once-nightly dose of either NPH or glargine to oral therapy was effective in reducing fasting plasma glucose and HbA1c (52). However, there were fewer episodes of nocturnal hypoglycemia with glargine.
Intensive Therapy
The term intensive therapy has had different meanings over the years. At one time, two doses of NPH was considered “intensive.” Currently, the intensive approach is an attempt to normalize not only FPG but also preprandial and postprandial levels, using a multiple-dose basal-bolus insulin regimen. Much evidence suggests that the elevation of HbA1c is a function not only of the FPG or the height of postprandial glucose excursions, but also of the mean 24-hour glucose or the integrated glucose concentration (“area under the curve”) over the entire 24 hours of the day. Intensive therapy requires a maximal effort by the patient, the caregiver, and a team of support personnel (trained diabetes nurse and dietitian).
Intensive therapy uses an insulin pump or, alternatively, a long-acting form of insulin is given to provide background insulin activity in addition to three or four daily doses of rapid-acting insulin given preprandially (basal-bolus). Approximately 50% to 60% of the total daily dosage of insulin is given as the long-acting depot injection in the morning or before bedtime; the remainder is divided and given before meals as bolus injections of rapid-acting insulin. Regular insulin can be mixed with the long-acting form (except when used with glargine) before breakfast. The term basal-bolus is often used to describe multidose intensive therapy.
Other regimens have been developed that use three or four doses of short-acting insulin in combination with one or more doses of intermediate-acting NPH and sometimes an oral agent (sulfonylurea). Regardless of the exact regimen, frequent smaller doses of insulin at equivalent or somewhat lower total daily amounts appear to produce better overall control than single larger doses, as evidenced by HbA1c(44,53).
Lispro (or the other rapid-acting insulins described above), although more expensive than regular insulin, can be taken immediately before eating instead of 20 to 40 minutes before a meal. Also, because of its shorter duration of action, it is less likely to cause postprandial and nocturnal hypoglycemia. Initiation of bolus insulin may be started by asking the patient to administer 1 unit of rapid-acting insulin for every 15 g of carbohydrates ingested (this usually requires education by a nutritionist) or approximately 4 units of insulin before each meal. It is important for the health care provider to review the patient's dietary history to be sure that the patient is consuming some carbohydrates in each meal. If there is no carbohydrate in the meal, the patient should not inject a rapid-acting insulin.
The goal for preprandial sugars should be 70 to 120 mg/dL. Occasional postprandial blood sugars are measured and should not exceed 180 mg. Weekly, three morning levels should be determined to detect nocturnal hypoglycemia. Diet must be optimized and contain as close to a constant total of calories from carbohydrate at each meal
P.1315
as possible. SMBG must be practiced before any program of intensive therapy with insulin is initiated.
The treatment group in the DCCT demonstrated that on such a program a daily mean glucose of 155 mg/dL and a HbA1c of 7.2% can be achieved. These values compare with a normal daily mean blood glucose of 110 mg/dL and HbA1c of less than 5.0%. Thus, even with the best possible motivation and supervision, the mean blood glucose will still be, on average, 40% greater than normal, yet severe hypoglycemic attacks can be expected to occur about three times more often than during conventional therapy. Intensively treated patients lose their early awareness of symptoms of hypoglycemia and develop symptoms at a lower level of blood sugar. Indeed, the 1980s saw an epidemic of frequent and severe hypoglycemia, a byproduct of the then-new popularity of intensive therapy. So far, intensively treated adults subject to repeated episodes of hypoglycemia have not shown decreased cognitive or psychomotor function, although this was a problem in children and in young adults in earlier studies of frequent, severe, hypoglycemic episodes.
Some patients achieve better glycemic control with an intensive regimen, whereas others do not do as well; the same is true of the frequency of hypoglycemia. For patients who can achieve results comparable to those in the DCCT, the benefits in terms of prevention of complications are significant. However, these patients must decide whether the effort is personally worthwhile; the caregiver's role is that of a supportive advisor. A significant number of patients who try intensive therapy are not able to achieve a satisfactory degree of control because of lack of motivation or skill, but for others, despite maximal cooperation, the goal is not attainable, presumably because of their particular metabolic makeup.
Insulin Pumps for Intensive Therapy
Continuous subcutaneous insulin infusion using a pump was first reported in 1978. In the past two decades, many reports have attested to the efficacy and advantages of this approach and have defined the complications and risks of this method of therapy. Not as generally appreciated is the demonstration that identical success at normalization can be achieved by multiple-dose programs (Fig. 79.1 and Table 79.6).
If a clinician decides to institute therapy with a pump, referral to a specialist familiar with one of these devices is usually necessary. Selection of a suitable current model of pump and initiation of therapy are best made by a team active in this specialized field, although if necessary, the continuation of therapy can be supervised by the general clinician. The generalist should, however, have a working knowledge of how insulin therapy is used via the pump. There are two settings for insulin infusion: basal and bolus. The basal infusion represents steady-state insulin levels that are separate from the insulin requirements needed to cover meals (bolus insulin). Typically, the basal insulin dose represents one-half of the daily insulin requirement. As an example, if a patient uses 48 units of insulin per day (total includes short- or rapid- and intermediate- or long-acting insulins) then half the daily dose would equal 24 units as the basal insulin dose. The basal rate in the pump would then be programmed as 1.0 unit per hour.
An advantage of using pump therapy is the ability to manipulate the basal infusion depending on the patient's particular needs. For instance, the basal rate can be temporarily reduced or discontinued (usually for no more than 2 hours) while the patient is engaged in exercise activity. The basal rate can also be adjusted at specific times to improve glycemic control. For instance, the patient might manifest fasting hyperglycemia but the early morning blood sugars are in an acceptable range (this would be an example of the dawn phenomenon described in Nocturnal or Early Morning Hypoglycemia in Patients Receiving Insulin). In this case, an adjustment to the basal rate would require a higher basal rate between the hours of 5 and 8 a.m. and then a lower basal rate for the remainder of the day.
The bolus infusions are similar to those used with syringe or pen injections. Typically, patients need to estimate the carbohydrate portion of the meal (usually using 1 unit for every 15 g of carbohydrates) plus the premeal plasma glucose value (usually using 1 unit for every 50 mg/dL above 100 mg/dL) as the dose of bolus insulin needed prior to the meal. One feature of the pump is the ability to deliver the bolus insulin in a square wave pattern compared to the more rapid, 30-minute peak with traditional injections.
Current pumps have fail-safe devices and alarms to guard against runaway pump action, power (battery) failure, empty insulin reservoir, and inadvertent turnoff. The insulin is administered through a 25-gauge butterfly needle attached to the pump via a piece of plastic tubing and inserted into the subcutaneous tissue of the abdomen. The needle is replaced every 1 or 2 days. Insulin reservoirs vary greatly in size and can accommodate 1 to several days’ supply. Pumps must be worn almost continuously, being removed for only short periods (15 to 30 minutes) to allow showering, bathing, or swimming. Some patients need a respite from the pump (e.g., in anticipation of sexual activity); in that case, a dose of intermediate-acting insulin can be used at bedtime and the pump used during the day. Other physical activities, including sports, are often performed with the pump in operation.
Despite the apparent inconvenience of wearing the device, acceptance of the pump is remarkable. Many patients have continued pump therapy for 10 years or more. Patients, observing the improved blood glucose levels, are gratified by a sense of control of their disease. In addition,
P.1316
they experience a normalization of activities by being freed from the need for meals to be taken at fixed times. However, hypoglycemia does occur even in the best-managed cases.
A number of problems with pump therapy have become obvious. Failure of the pump or clogging of its infusion line sometimes occurs. Diabetic ketoacidosis rapidly ensues, often overnight, when the insulin infusion is interrupted. Local infection at the needle site also predisposes to ketoacidosis caused by poor absorption of insulin from the infected area, and it may be severe enough to require antibiotic therapy and hospitalization. The approximate frequency of diabetic ketoacidosis has been estimated at 1 episode per 100 patient months, infection at 1 episode per 40 patient months, and severe hypoglycemia at 1 episode per 30 patient months.
Factors Affecting Insulin Requirement
Insulin Resistance
Classically, the term insulin resistance referred to a state in which the requirement for insulin exceeds 200 units daily. This extreme type of insulin resistance is only occasionally caused by the development of antibodies to insulin. Ordinarily, resistance to insulin occurs independently of antibodies to insulin. This resistance—or decreased sensitivity to insulin—is most apparent in type 2 diabetes that is associated with obesity. Even nondiabetic (often destined to eventually become overtly diabetic) obese patients who maintain normal levels of blood sugar do so by secreting supranormal amounts of insulin (i.e., they are in a state of compensated insulin resistance). Acanthosis nigricans, velvety hyperpigmented, hyperkeratotic lesions, often in skin folds, is often associated with insulin resistance.
Although obese diabetic patients are insulin resistant in terms of their glucose homeostasis and certain aspects of lipoprotein metabolism, their metabolic state is not so deranged that it allows ketoacidosis to develop. In most insulin-resistant diabetic patients, weight reduction at least partially reverses the insulin resistance. Glucose tolerance often improves to (or toward) normal, and the need for insulin to control hyperglycemia may decrease or disappear, only to reappear when weight is regained, as is almost always the case.
Mechanisms of Insulin Resistance in Type 2 Diabetes Mellitus
Insulin resistance is a hallmark of type 2 diabetes, but the mechanisms are still being elucidated. Free fatty acids (FFAs) are known to interfere with glucose utilization. Because obesity and diabetes are often associated with elevations of plasma FFA, a major role in insulin resistance for FFA was suspected. Indeed, a specific role for increased FFA flux from the central body (intra-abdominal) fat depots has been postulated to result in increased hepatic glucose output and elevation of blood sugar (54).
In addition to the FFA hypothesis, work linking obesity and insulin resistance has focused on the secretion by adipose tissue of the cytokine tumor necrosis factor-α as a metabolic messenger that induces insulin resistance. The inhibitory effects of tumor necrosis factor-α on insulin action and experiments in animals support this concept, but when obese diabetic patients were given antagonists to tumor necrosis factor-α, no effect was seen (54). Nonetheless, the secretion of anti-insulin factors from fat cells remains an active area of research.
Exercise During Insulin Therapy
Exercise-related hypoglycemia is a potential problem, because exercise has an insulin-like but insulin-independent effect on blood glucose that can be as strong as the maximal effect produced by insulin. The glycemic response to exercise is dependent on the exercise's intensity and duration. Exercise-related hypoglycemia can range from minimal and asymptomatic, requiring little or no therapy, to severe symptoms that need to be treated with intravenous glucose or injected glucagon. Some patients require an anticipatory reduction of insulin dose to prevent hypoglycemia. Others may need or prefer only additional food before moderate activity. In some patients, especially those who exercise sporadically, prolonged exercise may result in severe hypoglycemia some 6 to 15 hours after cessation of exercise. Modest exercise (walking 3.5 miles in 1 hour) may use 350 extra calories, but only 10 to 20 g of carbohydrates, a fraction of these extra calories, may be sufficient to prevent hypoglycemia. Similar considerations guide management of more vigorous exercise. Recent guidelines for type 1 diabetes are based on the assumption that the patient will be receiving a basal-bolus (e.g., glargine-lispro) insulin regimen (56). Such an insulin program provides about as much flexibility of dosing as one can presently achieve without the use of an insulin pump. At 30 minutes of exercise at levels 25%, 50%, and 75% of maximal oxygen consumption (VO2max), reductions of preprandial (breakfast) doses of lispro of 25%, 50%, and 75% were needed to avoid most episodes of hypoglycemia. Obviously, if a patient is to avoid hypoglycemia, any regular exercise program must be designed on an individual basis and must be reproducibly performed.
The hypoglycemic effect of exercise may be greater if the insulin has been injected into an extremity that is being exercised. Many patients ordinarily prefer to inject insulin into the thigh, but some who engage in vigorous exercise (running, other sports, manual labor) may have to use abdominal or arm injection sites to avoid excessive insulin effect caused by exercise-induced overly rapid absorption.
Changes in Insulin Requirement Caused by Stress and Other Factors
The stress of infection (or another inflammatory disorder) or trauma may increase insulin requirements quickly.
P.1317
Usually the site of any infection that is severe enough to produce this effect is obvious, or at least there is good evidence of an infectious process (fever, leukocytosis). The mechanisms by which infections increase insulin requirement are not understood completely but appear to relate to the interference of insulin action by cytokines (e.g., tumor necrosis factor-α). In addition, stress hormones, such as cortisol, can also increase insulin requirement by increasing insulin resistance.
Only rarely does a search for a hidden focus of infection provide an explanation for changing insulin requirements or even for the development of ketoacidosis. Although most type 2 diabetic patients are not prone to the development of overt ketoacidosis during metabolic stress, occasional patients, usually African American type 2 diabetics, seem especially prone to this type of response. After the acute episode is over, they may return to their baseline state. These patients may have been controlled by diet, oral agents, or insulin.
In type 1 patients many episodes of ketoacidosis that are not related to obvious stress are not caused by an increased insulin requirement. Rather, such episodes are usually related to noncompliance, although this may be unintentional, as when the patient mistakenly omits insulin.
Decreases of Insulin Requirement
Diabetic patients who develop nephropathy often show a decreased insulin requirement. Patients receiving long-term dialysis may develop severe hypoglycemia, even after insulin is discontinued, an ominous prognostic sign, presumably caused by failure of gluconeogenesis. A tendency to normoglycemia or even hypoglycemia develops occasionally in patients previously requiring insulin who develop chronic congestive heart failure. Development of adrenal or pituitary insufficiency in a diabetic patient results in a decreased insulin requirement, but such cases are rare.
Effect of Anorexia
When a patient with type 1 diabetes develops anorexia because of mild short-term illness (cold, flu, gastroenteritis), insulin should not be discontinued, but a reduction of the normal dosage by one-third to one-half may be needed. More severe illness (e.g., a marked febrile state) may require continuation of the usual dosage or even an increase in the dosage despite decreased intake of food. Every effort should be made to ensure intake of 50 g of carbohydrate in every 8-hour period to prevent starvation ketosis and hypoglycemia. Careful monitoring of urine or blood ketones (and blood glucose) during such periods, with prompt adjustment of insulin dosage, may prevent a hospitalization for ketoacidosis.
Allergic Reactions to Insulins
Allergy to insulin is rare. When it occurs, it is most commonly a local reaction at the site of injection. Local redness, swelling, heat, and itching occur within minutes to an hour after injection and persist for a few hours to a day, often with formation of an area of induration. Such reactions, no longer common, occur during the first few weeks of therapy and usually disappear as therapy is continued. Similar reactions can develop many hours or up to a day after injection (delayed hypersensitivity). Local reactions, misinterpreted as allergic, may also be caused by improper injection technique, the presence of preservatives in a particular brand, or even the injection of cold insulin.
Systemic allergic reactions, with or without a local reaction, are vanishingly rare; they are manifest by urticaria, angioedema, and even anaphylactic shock (immunoglobulin [Ig]E mediated; see Chapter 30). Such reactions seem to occur most often in patients who have previously received insulin and appear during reinstitution of therapy after a lapse of months or years. Local reactions may progress to systemic ones; if this seems to be occurring, one should treat the patient before anaphylaxis occurs. The first maneuver involves a trial of highly purified insulin. If this approach fails, drugs such as antihistamines and glucocorticoids are helpful, but persistent insulin allergy is best treated by desensitization in consultation with an allergist.
Lipoatrophy and Lipohypertrophy
Insulin lipoatrophy is now an uncommon event. Harmless but disfiguring localized atrophy of subcutaneous fatty tissue occurs around the site of insulin injections and is sometimes seen simultaneously with insulin allergy. The process may be related to impurities in insulin preparations rather than to insulin itself because preparations of high purity are much less likely to produce this problem. However, even insulin lispro, used in an insulin pump, has been reported to produce this phenomenon.
Insulin lipohypertrophy is even less common than insulin atrophy. This phenomenon is probably caused by an intrinsic action of insulin and has not been improved by use of purer insulins. Repeated injections into the same area do appear to predispose to lipohypertrophy.
Newer Agents
Pramlintide acetate injection (Symlin) represents a new therapeutic class for control of hyperglycemia (57). Pramlintide is a synthetic analogue of human amylin that, under normal circumstances, is cosecreted with insulin in the postfed state. It is indicated for patients with type 1 or type 2 diabetes (who must be insulin requiring) and is injected separately prior to each meal. It is contraindicated in patients with known or suspected gastroparesis,
P.1318
as it delays gastric emptying, and in patients with hypoglycemic unawareness. Initial trials suggested that it may decrease HbA1c 0.3% to 0.6%.
Exenatide injection (Byetta) represents another new therapeutic class for control of hyperglycemia, for use in patients with type 2 diabetes mellitus (58). Exenatide is an incretin mimetic agent that enhances insulin release from pancreatic beta cells in response to a meal and thus lowers postprandial blood glucose. It is indicated in patients with type 2 diabetes (who do not require insulin) and is contraindicated in patients with known hypersensitivity to the drug. Exenatide is injected subcutaneously twice daily (starting dose 5 µg), before breakfast and before dinner. Average reductions in HbA1c have been 0.5% to 1.0%. The most common side effect is nausea, so caution should be exercised in patients with known or suspected gastroparesis. It is not recommended in patients with severe renal impairment or end-stage renal disease.
Pancreas and Islet Cell Transplantation
Pancreas Transplants
With the advent of more effective immunosuppressive regimens, the survival of transplanted patients and of the transplants themselves has improved considerably. Currently, graft survival approaches 95% at 1 year (59). Most transplantations have been simultaneous cadaveric pancreas–kidney transplantations in type 1 diabetics with end-stage renal disease. Alternatively, many patients have received renal transplants, either from cadavers or living related donors, followed at a later time with a cadaveric pancreas transplant. Pancreas transplantation alone, in patients without renal failure, has been done much less often because survival of the graft is considerably worse (60), perhaps because of a protective effect of the transplanted kidney.
The best candidates for transplantation are unstable diabetics with end-stage renal disease. There is no role currently for pancreas transplantation in type 2 diabetics. The best results are seen in relatively young patients (age 45 years or younger) with no cardiac risk factors. Both older age and heart disease are associated with poorer graft and patient survival.
When successful, pancreatic transplantation removes the need for exogenous insulin administration, protects the transplanted kidney from the noxious effects of hyperglycemia, and improves the recipient's quality of life. It does not appear to reverse diabetic retinopathy. Evaluation of patients for transplantation requires the collaborative efforts of an endocrinologist, usually a nephrologist, and a transplant surgeon.
Although islet cell transplantation has been reported to be successful in small studies (61), the American Diabetes Association recommends that this procedure should still be considered to be experimental and should be performed only within the context of a research study (60).
Treatment of Other Types of Diabetes Mellitus (Secondary Diabetes)
Drug-induced diabetes (e.g., diabetes induced by high-dose thiazides) and diabetes associated with the use of glucocorticoids are usually not characterized by ketosis and ordinarily resemble type 2 diabetes. Treatment with a sulfonylurea may be tried, but insulin is often necessary. Withdrawal of the offending agent does not always ameliorate the diabetic state. The possibility of precipitating diabetes in patients with a strong family history should not deter the physician from the judicious use of glucocorticoids when these agents are clinically indicated. Similarly, a diabetic who is already receiving insulin should not be denied glucocorticoids for fear of aggravating the diabetes. If such aggravation occurs, usually only an increase of insulin dosage is necessary to reestablish the previous state of glycemic control.
Diabetes secondary to chronic pancreatitis or pancreatectomy should be treated with insulin. The insulin requirement is usually 20 to 40 units per day. The patients should be instructed to follow the dietary strategy outlined in Table 79.4. Alcoholic patients with this form of diabetes are particularly difficult to manage if they continue to drink heavily and eat erratically.
Monitoring Glycemic Control
The rational approach to day-to-day monitoring depends on whether the treated patient has type 1 or type 2 diabetes and whether conventional or intensive therapy is being used. Unless insulin dosage is being adjusted as a function of the results of SMBG, multiple daily measurements are unnecessary.
Urine Glucose Monitoring
Urine glucose monitoring is no longer advocated and is even disdained by most diabetologists. Yet it is a simple, inexpensive procedure and in some socioeconomic circumstances, it may be all that is available.
In ketosis-prone patients with type 1 diabetes, the urine should also be monitored for ketonuria. This can be accomplished using Acetest tablets (which incorporate the nitroprusside reaction for detecting ketones) or one of the combination “stix.” Ordinarily, monitoring for ketones is unnecessary as a routine procedure, even in type 1 patients. Some experts recommend monitoring for ketones whenever the PG exceeds 350 mg/100 mL or remains persistently elevated. Type 2 patients do not require monitoring for ketonuria unless a severe intercurrent illness develops.
P.1319
Self-Monitoring of Blood Glucose
If glycemic control to a degree that approaches normoglycemia is the goal, SMBG is mandatory. In the patient with type 2 diabetes, complete absence of glucosuria can often be achieved safely (i.e., with avoidance of most episodes of hypoglycemia) even without the need for frequent determinations of fasting blood glucose, but this is not possible in patients with type 1 diabetes or with intensive insulin therapy. Other indications for SMBG include patients with an unusually low or high renal threshold for glucose, many patients with type 1 diabetes treated conventionally, all patients prone to hypoglycemic episodes, pregnant patients, and some patients with type 2 diabetes who, despite an inability to master effective therapy, seem to find SMBG more satisfying.
The basis for all SMBG methods is a paper strip impregnated with an enzyme reagent (glucose oxidase) and suitable dyes. When placed in contact with a drop of capillary blood, the change of color intensity indicates the glucose concentration. Some strips are read only visually, that is, without a reflectance meter (e.g., Chemstrip bG), others are read either visually or with a reflectance photometer (e.g., Glucostix), and still others are read only with a photometer (e.g., Glucofilm). Accuracy of strips properly examined visually is adequate for monitoring control, except when the goal is intensive therapy with normalization of the blood sugar. For many patients a meter is unnecessary, but most feel more secure with machine readings.
In the United States, the meters in widest use are Accu-Check, which uses Chemstrip bG, and LifeScan. All machines are reliable, portable, battery operated, and relatively inexpensive. The manufacturers have reduced the prices to promote sale of the matching strips, the retail cost of which is approximately 60 to 80 cents each. For patients who check their home blood glucoses several times daily, the monthly cost may be considerable. Medicare now pays all costs of monitoring in eligible patients (older than age 65 years), and many companies now provide all necessary paraphernalia through online orders.
Capillary blood is most commonly obtained from the tip of the finger. Disposable lancets are used to produce the puncture. The required drop of blood may be obtained almost painlessly using a spring-triggered device (e.g., Autolet, about $30).
Blood flow from the finger can be enhanced before puncture by holding the hand in warm (not hot) water for 30 seconds. The skin should be quickly dried. Puncturing the thumb is least painful, but the ring finger has the best blood supply. Puncturing the lateral aspect of the fingertip (distal phalanx) is less painful than puncturing the ball. Pain is also less when sufficient pressure to produce erythema is applied to the palmar surface (ball) of the distal phalanx; an opposing digit of the same hand is used to apply the pressure. The first drop of blood produced suffices; the presence of extravascular fluid does not affect the result. The finger is inverted and the drop of a size recommended by the manufacturer is transferred to the strip according to the manufacturer's directions; then timing is begun and the glucose level is read.
Glycosylated Proteins, Hemoglobin A1c
Chronic elevation of blood glucose results in an increase in the concentration of glycosylated hemoglobins, a major component of which is HbA1c. Determination of the level of HbA1c gives an integrated estimate of the degree of hyperglycemia over 5 weeks to 2 months (Fig. 79.2). The normal range of HbA1c is 3.8% to 6.3% of total hemoglobin (normal range of total glycosylated hemoglobin is slightly higher, e.g., 5.3% to 7.9%) and may rise to 15% with chronic hyperglycemia postprandially. Values of less than 7.5% suggest good control with fasting and 1-hour postprandial sugars in the range of 70 to 120 and 100 to 140 mg/100 mL, respectively. With 120 to 140 mg/100 mL fasting and 141 to 160 mg/ 100 mL postprandial, one might see HbA1c at 7.5% to 9%. At 140 to 160 mg/100 mL fasting and 160 to 200 mg/100 mL postprandially, values of 9.1% to 11% are common, whereas minimal control gives values of greater than 11%. Glycosylated hemoglobin levels fall slowly with
P.1320
reduction of mean glucose because circulating red blood cells containing high levels of glycosylated hemoglobin disappear normally in approximately 120 days. If euglycemia is established, glycosylated hemoglobins subsequently normalize in 4 to 6 weeks. Conversely, persistent hyperglycemia must be present for 1 to 4 weeks before elevated levels of glycosylated hemoglobins are seen. Short periods of hyperglycemia (6 to 24 hours’ duration) may result in disproportionate elevations because some methods include measurement of unstable glycosylated derivatives. Other conditions render interpretations of glycosylated hemoglobin values uncertain, including any in which red cell life span is low (bleeding, hemolysis, sickle diseases) or in which hemoglobin F is increased (some hemoglobinopathies).
|
FIGURE 79.2. Relationship between hemoglobin A1c and mean blood glucose. Twenty-one subjects performed self-monitoring of blood glucose four to six times per day for 8 weeks. The arithmetic mean of those values as compared with the hemoglobin A1c value determined at the end of the 8-week period. (Modified from Nathan DM, Singer DE, Hurxthal K, et al. The clinical information value of the glycosylated hemoglobin assay. N Engl J Med 1984;310:341. ) |
HbA1c should, in theory, be usable not only for monitoring glycemic control but in the diagnosis of diabetes (chronic hyperglycemia) and in screening. The objections to such uses have been based on lack of standardization of methodology. Although this is no longer the situation, sensitivity and specificity barriers remain that are population (ethnicity) based. HbA1c has also been found to be remarkably nonreproducible in healthy adults, in contrast to the situation in diabetics. To explain the variability of the HbA1c in normal subjects, it has been postulated that the process of hemoglobin glycation varies from one erythrocyte generation to the next (62). The limitation on the use of HbA1c for diagnosis and screening, therefore, is not technical or analytical but rather biological variability.
Fructosamine refers to the ketoamines formed from glycosylated proteins other than hemoglobin. It can be measured accurately, quickly, and relatively cheaply, but because it varies with the concentration of albumin in the serum and because the turnover of albumin is much shorter than that of hemoglobin, measurement of HbA1c is the preferred test in the evaluation of long-term control of blood glucose. Fructosamine could be used to monitor glycemic control in patients with hemoglobinopathies that lead to erroneous HbA1c levels.
Monitoring Glycemic Control in Patients Receiving Insulin Therapy
Efficacy of treatment in the conventionally treated ambulatory patient should be monitored, if possible, by measuring fasting glucose levels. Near normalization of the fasting (overnight) PG represents the basic or coarse adjustment of insulin dosage. Preprandial and postprandial normalization can be viewed as fine adjustments, both of which are difficult to attain. No useful purpose is served by attempts to adjust preprandial glucose levels before normalization of the fasting level is achieved; only thereafter should blood glucose be monitored at midafternoon or before the evening meal. The availability of techniques for self-monitoring of blood glucose (SMBG) have made blood glucose much easier to track.
The process of SMBG should be initiated as a prelude to tight control because unless the patient is able to master the technique and accept it as an ongoing necessity, the effort at tight control will fail. SMBG does not eliminate the need for dietary compliance. Recent studies indicate that within the wide range of what is grossly considered normal, neither intelligence, socioeconomic status, nor personality type has any predictive value for success with intensive therapy. Patients of limited financial means may be challenged by the high cost of such a program.
Monitoring Glycemic Control with Oral Therapy
The frequency and type of monitoring should be determined by the severity of the diabetes and the goal of treatment. For patients whose FPG becomes normal or reaches an acceptable level, monitoring can be simple because patients receiving oral therapy are not ketosis prone and have fairly stable diabetes. Similar considerations apply to patients being treated with diet alone. These patients must be taught that if they develop symptoms and signs of uncontrolled hyperglycemia (heavy glucosuria, polyuria, polydipsia, blurred vision), prompt advice from a clinician is absolutely necessary. Routine testing for urinary acetone is unnecessary unless the patient has new onset of persistent glucosuria or at some earlier time had an episode of ketoacidosis, perhaps during stress.
FPG should be determined every few months in most patients, but the best means of monitoring patients who respond to oral treatment with normalization of blood sugar is by determination of HbA1c. Development of frank hypoglycemia or excessive lowering of FPG below 70 mg/dL (which may be detected before symptoms develop) is an indication for downward adjustment of drug dosage. When dosages are changed or when drugs are added or removed, monitoring should be done more often, perhaps daily, with SMBG.
Hypoglycemia
When severe, hypoglycemia causes central nervous system symptoms (neuroglycopenic symptoms) ranging from headache or subtle disturbances of mental function, to confusion, visual disturbances, and personality change, or, rarely, to seizures, unconsciousness, and transient hemiparesis. More commonly, when hypoglycemia occurs during waking hours and is accompanied by the usual symptoms of epinephrine release (tremor, sweating, tachycardia, and palpitations), there is no problem in recognizing the condition. However, in some poorly controlled diabetics, as in some normal people, even mild reductions of blood glucose to levels (50 to 70 mg/100 mL) not
P.1321
clearly identifiable as hypoglycemia can sometimes produce epinephrine release with its resulting symptoms (see Chapter 81). Under these circumstances, documentable hypoglycemia is not present and the clinical situation may be confusing.
Diabetic patients often develop defects in mechanisms that normally counterregulate hypoglycemia (63,64). This pathophysiologic state may occur within a few years of onset of the disease. Absolute deficiency of glucagon secretion is common in type 1 diabetes and sometimes occurs in type 2 diabetes as well. Defective endogenous glucagon responsiveness to hypoglycemia in type 1 diabetes is not normalized (reversed) by establishment of tight control (62). Defective counterregulation caused by impaired secretion of epinephrine is also common early in type 1 disease and may become marked in patients with autonomic (adrenergic) neuropathy late in the course of the illness. Other patients may have defective counterregulation caused by impairment of epinephrine action as a result of treatment with β-adrenergic-blocking drugs. In addition, such agents may mask many of the symptoms of epinephrine excess. Regardless of their precise mechanisms, these defective counterregulatory responses undoubtedly contribute in many diabetic patients to their high risk of developing severe hypoglycemia during therapy with insulin.
The most common cause of hypoglycemia in the diabetic patient receiving conventional insulin therapy is failure of the patient to eat at normal times. Skillful questioning usually reveals the problem. Many intensively treated patients appear to develop tolerance to hypoglycemia and remain asymptomatic despite markedly subnormal concentrations of glucose, a state called hypoglycemia unawareness(65).
Nocturnal or Early Morning Hypoglycemia in Patients Receiving Insulin
Excessive insulin action often occurs during the night or early morning hours. The hypoglycemia-induced release of epinephrine and other counterregulatory hormones (cortisol, growth hormone, glucagon) then causes rebound hyperglycemia, glucosuria, and ketonuria, an effect known as the Somogyi phenomenon. If the clinician notes an elevated blood sugar and prescribes still more insulin, the result is further hypoglycemia, perpetuation of the cycle, and possible serious consequences. Although the existence of the Somogyi phenomenon has been repeatedly challenged, other evidence convincingly points to its contribution to the problem of glucose regulation (66).
To detect this phenomenon, all insulin-receiving patients should be questioned carefully for clues to the presence of nocturnal hypoglycemia (e.g., nightmares, night sweats, and headache during the night or on arising), although these symptoms may not be present and hypoglycemia is revealed only by routine SMBG during the night. The point at which epinephrine release is secreted and produces sweating and other symptoms is quite variable; some diabetic patients trigger secretion at glucose concentrations as high as 50 mg/dL, others do not have counterregulatory release until the blood sugar falls to as low as 30 to 40 mg/dL, and still others have defective counterregulation (see above) and only neuroglycopenic symptoms.
Increasing the intake of carbohydrate in the late evening or reducing insulin dosage by 10% in type 1 diabetes and up to 20% to 30% in type 2 diabetes often corrects the situation. In the latter patients, such a brief and substantial reduction in insulin dosage can be made with impunity.
The classic Somogyi phenomenon must be distinguished from two other possibilities: waning of insulin action and the dawn phenomenon.Waning of insulin action occurs when the patient is receiving an insufficient amount of intermediate- or long-acting insulin; either a single morning dose is not carrying into the next day or the second dose, given before dinner or at bedtime, is inadequate. The dawn phenomenonis an increase of blood sugar between 3 and 7 a.m. that occurs despite continuous subcutaneous infusion or background insulin action from a long-acting insulin. An increased amount of insulin is necessary to overcome the glucose raising action of growth hormone, which is secreted in pulsatile fashion during the night with considerable interindividual variation and, unfortunately, variation from day to day as well. Because of this variation, an amount of insulin that is sufficient one day may be inadequate or excessive on the next.
Obviously, patients with waning insulin action or the dawn effect need more insulin, whereas the Somogyi effect requires that less be given or dosing times adjusted. The simplest way to distinguish these is by SMBG, often for several nights, with samples at 9 p.m., midnight, 3 a.m., and 7 a.m. More frequent sampling may be needed. Figure 79.3 shows the different patterns. Constantly rising glucose indicates waning insulin. A plateau followed by a rise indicates the dawn phenomenon. A drop during the night to a clearly hypoglycemic level points to the Somogyi effect. If SMBG cannot be done, cautious reduction of the dosage of (evening) insulin should be attempted.
Treatment of Hypoglycemia
The immediate therapy of daytime hypoglycemia in a conscious patient is ingestion of food, preferably sugar. Patients should carry a ready carbohydrate source, such as candy, and must realize that a tiny piece of such material will not suffice. Five or six Life Savers provide the minimum necessary 10 g of carbohydrate, as does a piece of fruit. Glucose tablets (5 to 10 g glucose/tablet) are now available. Also, 4 to 6 oz of sweetened fruit juice or of a nondiet soft drink are satisfactory. A tablespoon of ordinary table sugar (sucrose) may be added to fruit juice
P.1322
or dissolved in one-half cup of water. Relief of symptoms should be seen in 10 to 20 minutes. Family members or friends should be instructed in the treatment of such an emergency and should not waste time attempting to reach medical assistance before administering sugar. Emergency medical care may be sought after sugar is given, but the problem is usually resolved by the time medical assistance can be obtained. If no obvious cause is apparent for the episode of hypoglycemia—such as a missed meal that is subsequently eaten—the patient should be on guard for recurrence over the next few hours, during which time repeated ingestion of sugar, at hourly intervals, may be advisable.
|
FIGURE 79.3. Idealized patterns of blood glucose concentrations during the night. The three patterns represent waning insulin action, the dawn phenomenon, and the classic Somogyi effect. All result in fasting hyperglycemia but are distinguished by the patterns of blood glucose concentration in the preceding hours. |
Occasional patients cannot be treated by the simple means described. Either because of a hypoglycemia-related alteration of mental status resulting in an uncooperative state or because of unconsciousness, some patients cannot take oral sugar. A safe and effective emergency therapy is administration of 1 mg of glucagon subcutaneously by a person instructed in this technique (glucagon promotes the breakdown of hepatic glycogen to glucose). Glucagon is readily available in single-dose form (1-mg vial) and should be kept available during initiation of insulin therapy and in hypoglycemia-prone patients. About 10 to 15 minutes are required for an obvious effect on the sensorium. As soon as possible, oral sugar should then be given. An effort should always be made to identify the cause of the hypoglycemic episode and to reduce insulin dosage or take other appropriate action to prevent recurrence.
Miscellaneous Factors Contributing to Hypoglycemia
Although defects of counterregulatory responses undoubtedly contribute to recurrent episodes of hypoglycemia in many patients, other factors in their daily lives also contribute. Noncompliance with diet may be deliberate or accidental. The need to consume small snacks can be unappreciated or forgotten. Meals are often taken off schedule, upsetting the effort to adjust insulin dosage to preferred time of meals. Amounts of food, if greatly varied, adversely affect insulin dosage. Emotional upset, difficult to evaluate as a cause of varying control, is nonetheless a significant factor in some circumstances. Injudicious use of alcohol is always a concern. Patients may have trouble measuring their insulin or may reverse the ratio of mixtures. Undocumented hypoglycemic reactions may be improperly treated and unreported. Techniques of blood monitoring are often at fault; patients misread directions or introduce variations that lead to errors of measurements. Not to be ignored is the effect of exercise and accelerated absorption of insulin by factors that raise skin temperature (hot shower, tub or spa).
Complications of Diabetes Mellitus
Diabetes mellitus is associated after many years with two distinct types of vascular damage in various organs (Fig. 79.4). Hyperglycemia produces microvascular (capillary) disease that affects the eyes, kidneys, and nerves. The blood lipid abnormalities of diabetes produce accelerated and extensive atherosclerosis of the cardiovascular and peripheral vascular systems. Coronary artery atherosclerosis and renal failure account for most of the deaths attributable to diabetes. Although considerable progress has been made over the past four decades in the understanding and treatment of this disease, diabetes remains a leading and increasing cause of death and enormous morbidity. In the United States and worldwide, the disease is showing a relentless increase.
Cardiovascular Problems
Hypertension and Its Therapy
Hypertension is a common complication of diabetes. Type 1 diabetics have an increasing incidence of hypertension with time (5% by 10 years, 33% by 20 years) (67). In type 2 diabetes, hypertension is often part of the metabolic syndrome at onset.
The risk of microvascular and macrovascular complications is nearly doubled in hypertensive compared with normotensive diabetics (67,68), independent of the risk of
P.1323
microvascular complications imposed by hyperglycemia. In both type 1 and type 2 diabetics, therefore, control of hypertension is critical. A number of sizable prospective controlled studies support this statement (69,70).
|
FIGURE 79.4. Complications of diabetes mellitus as a function of duration of the disease. (From Davidson MB. The continually changing “natural history” of diabetes mellitus. J Chronic Dis 1981;34:5 , with permission.) |
In type 2 diabetes, the same general measures that are important in the treatment of the diabetes itself—exercise and weight reduction—also are important in the control of hypertension. In addition, sodium restriction is important to avoid volume expansion that is part of the syndrome.
If blood pressure cannot be reduced to 130/85 mm Hg or lower, antihypertensive drugs should be prescribed for both type 1 and type 2 diabetics. If patients already have evidence of renal disease, the target blood pressure should probably be 125/80 mm Hg or lower. Detailed recommendations for the pharmacologic treatment of hypertension in diabetic patients are provided in Chapter 67.
There has been concern that previously used, higher doses of thiazide diuretics increase morbidity and mortality in this population and decrease glucose tolerance (71). It now seems clear that a low dose of a diuretic (e.g., hydrochlorothiazide 12.5 to 25 mg per day) is as effective as a higher dose in lowering blood pressure and that the development of vascular disease is retarded and the incidence of adverse vascular events is reduced (72).
Although β-adrenergic blocking agents provide effective blood pressure control, nonselective agents (propranolol, pindolol, nadolol, timolol) are best avoided because of possible worsening of glucose tolerance (inhibition of insulin release in type 2 diabetes), interference with recovery from hypoglycemia, masking of hypoglycemic symptoms, and occasional promotion of hyperkalemia. However, the selective β1-blockers (atenolol, metoprolol) can be used because they are unlikely to cause these problems at conventional dosages. Concern over aggravation of hyperlipidemia by selective β-blockers is unwarranted because, unlike the nonspecific propranolol, which gave rise to these concerns, the more selective agents actually have only trivial effects on blood lipids.
As with diuretics there has been concern about the adverse effects of calcium channel blockers in the treatment of hypertensive diabetics (73). However, a number of studies have shown that these drugs, too, effectively reduce the risk of macrovascular disease in this population (73,74). Both nondihydropyridine blockers (e.g., verapamil) and dihydropyridine blockers (e.g., nifedipine) appear to be effective. Whether calcium channel inhibitors retard the development of microvascular disease remains to be seen (74).
Angiotensin-converting enzyme (ACE) inhibitors have become the antihypertensive agents of choice in the treatment of hypertensive diabetics. They reduce the risk of macrovascular complications at least as well as other classes of hypotensive drugs (75,76) and have the best risk-to-benefit ratio (see Chapter 67). In type 1 diabetics, ACE inhibitors have the added advantage of retarding the development of renal microvascular disease (77). In type 2 diabetics, the effect of ACE inhibitors on the development of nephropathy has not been as clearly demonstrated, but recent studies show that angiotensin receptor blockers do have such an effect (78,79). The accumulated evidence, therefore, supports the preferential use of drugs that inhibit the renin–angiotensin system as primary agents in the treatment of both type 1 and type 2 hypertensive diabetics. As for now, the major difference between ACE inhibitors and angiotensin receptor blockers is cost. (ACE inhibitors are considerably less expensive than receptor blockers and will become even less expensive as the patents on them expire.)
Because of the need to lower blood pressure aggressively in hypertensive diabetics with even early evidence of renal disease, combinationsof hypotensive agents often need to be prescribed. In such circumstances, adding a diuretic to an ACE inhibitor is probably the first thing to do. If optimum pressures are not reached, a β-blocker and/or a calcium channel blocker can be added sequentially.
Other problems with antihypertensive drugs that may be especially troublesome in diabetic patients are erectile
P.1324
dysfunction (diuretics, α- or β-adrenergic blockers) and hyperkalemia (triamterene, ACE inhibitors), possibly in relation to the subclinical hyporeninemic hypoaldosteronism often present in these patients.
Atherosclerotic Vascular Disease
Diabetics experience two to four times the coronary artery disease of nondiabetics, and when myocardial infarction does occur, the morbidity and mortality in diabetics is twice that of nondiabetics (80). Early attention should be paid to complaints of chest pain and, even in its absence, diaphoresis, dyspnea, or nausea, because these may be anginal equivalents. In addition, diabetics have silent myocardial ischemia more frequently than nondiabetics. These considerations have led to guidelines for screening for cardiac disease in diabetics in the hope of improving outcomes (81). Because coexistent hypertension, diabetic cardiomyopathy, and autonomic neuropathy may alter the sensitivity and specificity of screening tests, more sophisticated studies, such as sestamibi imaging, may be necessary (see Chapter 62).
Because of the prevalence and severity of cardiac disease in diabetics, aggressive management of the risk factors (e.g., smoking cessation, blood pressure control, management of hyperlipidemia) is important (see Hypertension and Its Therapy; Hyperlipidemia; Chapters 67 and82).
Cardiovascular thrombotic events contribute to the morbidity and mortality of diabetes, no doubt in part because of increased atherosclerosis, but probably also because of a variety of demonstrated alterations of platelet function and of the blood fibrinolytic system. Plasma plasminogen activator inhibitor type 1 and fibrinogenase are often elevated in type 2 patients, especially in the metabolic syndrome, and are correlated with elevated triglycerides and hyperinsulinemia (82). Although diabetes is considered to be a state of hypercoagulability, the relationship of these abnormalities to practical therapy is unclear.
Given the high cardiovascular risk of diabetic patients and the proven usefulness of aspirin therapy in nondiabetic patients, it is unsurprising that antiplatelet therapy is beneficial. The routine use of aspirin in type 2 diabetes is supported by at least six studies that attest to its safety and effectiveness (83). The recommended dose is 81 to 325 mg/day as an enteric-coated preparation, a recommendation sanctioned by the ADA (83). Aspirin does not increase intraocular bleeding in diabetic retinopathy and can be safely used in this condition. Studies of other antiplatelet agents (e.g., clopidogrel [Plavix]) in diabetics are not available, but their use may be considered in aspirin-allergic patients.
Diabetics are also more likely than nondiabetics to develop peripheral vascular disease (up to 30-fold increase in incidence) and cerebrovascular disease (a 2- to 6-fold increase in the incidence of stroke and cerebrovascular deaths).
Diabetic Cardiomyopathy
The Framingham study demonstrated that the risk of heart failure is increased 2.4-fold in diabetic men and 5-fold in diabetic women (84). This risk is associated with an increased mortality rate, even higher than the increased mortality rate of nondiabetics in heart failure. Although heart failure in diabetics is often associated with ischemic cardiomyopathy, a consequence of the accelerated atherosclerosis that is so often associated with diabetes (see above), it also may occur without clear-cut evidence of coronary artery disease. Increased left ventricular mass and primarily diastolic dysfunction are characteristic of this condition. Treatment should include an ACE inhibitor (seeHypertension and Its Therapy, Chapter 66).
Hyperlipidemia
Hypercholesterolemia in diabetics, as in nondiabetics, is a major risk factor for atherosclerotic disease. The kinds of lipid (lipoprotein) abnormalities in type 1 and type 2 diabetic patients differ. The term diabetic dyslipidemia embraces both of these patterns, but it is usually used to denote the pattern seen in type 2 diabetes. In patients with type 1 diabetes whose blood sugar is well controlled and who have normal renal function, the serum lipoproteins are not very different from those in normal subjects: LDL is usually normal and HDL may be even higher than normal. Oxidized LDL, the formation of which is facilitated by glycation, is not ordinarily measured but is more atherogenic than normal LDL and is often increased. In contrast, patients with type 2 disease have increased levels of intermediate-density lipoproteins (β-very-low-density lipoprotein, intermediate-density lipoprotein), and HDL is often low. Glycated and oxidized LDL are increased. A direct contribution of hyperinsulinemia to the atherogenic process is also suspected but remains controversial. Current guidelines (see Chapter 82) consider diabetes mellitus equivalent to a cardiac risk factor and recommend treatment should be instituted to lower serum LDL if it exceeds 100 mg/dL, with the goal of achieving levels less than 70 mg/dL.
Management
The hyperlipidemia of type 1 diabetes usually responds to conventional control of hyperglycemia. LDL cholesterol also is usually normalized. Triglycerides are usually not elevated after hyperglycemia is controlled. In patients with type 2 diabetes, as part of the metabolic syndrome, hyperlipidemia may improve somewhat with effective lowering
P.1325
of hyperglycemia, but even euglycemia does not abolish the problem.
The serum lipid abnormalities that persist after the best achievable control of hyperglycemia in both type 1 and type 2 diabetes should be treated aggressively (85,86). At any given level of cholesterol, the type 2 diabetic has two to four times the risk of coronary artery disease of the nondiabetic. It is currently recommended that diabetics with LDL levels above 100 mg/dL should be treated with lipid-lowering agents (see Chapter 82). Using this criterion, 80% of diabetics will be candidates for treatment of hyperlipidemia.
Neuromuscular Disease
Neuropathy
The precise prevalence of neuropathy in diabetic patients is unknown, although it is clear that neuropathy is common and that in most patients the occurrence and severity of involvement are related to duration of the disease. Usually, many years pass before the process becomes obvious, but occasionally even severe neuropathy can have an early onset (see Chapter 92 for a general discussion of peripheral neuropathy).
The most commonly appreciated abnormality is that which affects peripheral sensory nerves. Several types of sensation are involved (pain, proprioception, vibration, light touch) and can lead to unsteadiness, ataxic gait, and such uncommon but striking disorders as neuropathic arthropathy. Less-well appreciated are the autonomic disorders that give rise to disturbances of cardiovascular function (postural hypotension, resting tachycardia), genitourinary function (impotence, bladder dysfunction), and gastrointestinal function (nocturnal diarrhea, fecal incontinence). Motor deficits are much less common but may occur with striking suddenness. Weakness is distal (neuropathic) rather than proximal (myopathic), although a specific type of myopathy also occurs in diabetic patients (see Amyotrophy [Proximal Asymmetric Motor Neuropathy]). Some authors use the term distal symmetric sensorimotor polyneuropathy to describe the most common form of diabetic polyneuropathy.
Peripheral Sensory Neuropathy
Classically, the deficit is distal, with the lower extremities affected first, followed by the upper extremities. The term stocking–glove distribution is appropriate. The disorder is a symmetric polyneuropathy with a proximal–distal gradient of dysfunction. In severe cases, even the sensory innervation of the trunk is involved; in this instance, the most distal fibers are those of the anterior abdomen and lower thorax. Rarely, even the distal portions of the cranial nerves are affected (e.g., the distal sensory portion of the trigeminal nerve). The patterns of loss are not specific for diabetes mellitus and can be seen in such diverse states as amyloid neuropathy and toxic (e.g., lead) neuropathies.
The nerve damage at first may be asymptomatic, although subtle symptoms may be revealed with careful questioning of the patient. Alternatively, the patient may first complain of hyperesthesia and dysesthesia, including tingling and burning sensations. Later, various symptoms are experienced, including sensations of numbness or heaviness. Patients often complain that their feet feel dead or that they have a sensation of walking on a soft or nonexistent surface. Loss of ability to perceive temperature and firmness gives rise to these complaints. Severe, spontaneous, short-lived, stabbing leg pains and cramps are common. Often, these pains are most troublesome at night.
On neurologic testing, skin hyperesthesia is the most common finding (pinprick, two-point discrimination, light touch). Sensory testing with a standardized (10-gm) monofilament is recommended (information on obtaining monofilaments is found athttp://www.diabetesmonitor.com/footresources.htm). The hyperesthesia and loss of temperature perception lead to unappreciated skin trauma and predispose to infection. Sensory loss in the fingertips can prevent the blind diabetic from learning Braille. Deep tendon reflexes, especially that of the Achilles tendon, are lost, often in the early stages of the neuropathy.
Peripheral Motor Neuropathy
Much less common and less-well recognized are the motor function abnormalities that occur as part of diabetic neuropathy. The intrinsic muscles of the feet are those most commonly involved. Interosseous atrophy produces inability to separate toes but, more important, allows the foot to assume abnormal positions. When claw or hammer toe develops, new pressure points appear at the tips of the toes and along the dorsal aspects; hyperkeratosis, callus formation, and ulceration follow. The interosseous atrophy that may affect the hands does not lead to total loss of function but does result in weakness of grip. Diffuse weakness of the legs and upper extremities may also occur.
Therapy of Painful Peripheral Neuropathies
A recent report from the American Diabetes Association provides an evidence-based algorithm for the management of painful peripheral neuropathy in diabetic patients (87). Nondiabetic etiologies should be considered and excluded when appropriate, and glycemic control stabilized. Drug therapy usually starts with tricyclic drugs, for example, amitriptyline 25 to 150 mg at bedtime. Amitriptyline may have more side effects than desipramine and may be less desirable for use in elderly diabetic patients. Alternative or additional drugs include anticonvulsants (e.g., gabapentin [Neurontin] (88), typically 1,800 mg daily in
P.1326
divided doses) and opioid or opioidlike drugs (e.g., tramadol [Ultram] or oxycodone). Nonpharmacologic therapies, such as acupuncture, and topical treatments, including capsaicin (Zostrix; applied to the skin as a cream three to four times daily) also may be helpful for some patients in the treatment of painful diabetic neuropathy. Referral to a specialized pain clinic should be considered if the patient has bothersome symptoms and does not respond to or tolerate initial therapeutic interventions.
Mononeuropathies
Mononeuropathy (mononeuritis simplex and multiplex) may occur in any superficial nerve (simplex) or asymmetric simultaneous combination (multiplex). The lower extremities are more commonly involved (femoral, lateral femoral cutaneous, sciatic, peroneal) than are the upper (e.g., ulnar, radial). Onset is usually sudden with intense often cramping and lancinating pain (see Chapter 92). Typically, the pain is worse at night and, when the lower extremities are involved, may be relieved by pacing about. When the pain is radicular (trunk or abdomen), intrathoracic or intra-abdominal disease may be misdiagnosed.
At onset, diagnosis can only be surmised, although tenderness along a nerve trunk is suggestive. Herpes zoster may be suspected, especially when hyperesthesia occurs, but when no vesicles appear and muscle weakness and atrophy are eventually evident, the diagnosis becomes obvious. The prognosis is good; complete recovery within a few months is the rule.
Cranial and Oculomotor Neuropathies
Cranial and oculomotor neuropathies are distinguished from other mononeuropathies mainly by their location. Pain and headache may be present. The most common nerves involved are III (palpebral ptosis, pupillary function undisturbed), VI (inward deviation of eye, diplopia), and IV (inward and upward deviations, diplopia). Recovery within 3 months is almost universal. When the facial nerve is involved, distinction from Bell palsy is impossible (see Chapter 92), although the diabetic variety tends to be less severe and recovery is usually complete.
Autonomic Neuropathy
Abnormal Sweat Production
Almost always associated with other evidence of diabetic autonomic neuropathy, this complication in its typical form produces heat intolerance and increased sweating (hyperhidrosis) of the upper half of the body with decreased or absent sweating (anhidrosis) below the midtrunk. In other cases, anhidrosis is generalized, and recognition of the complication may be difficult. In women the condition may be confused with menopausal sweats.
Affected patients have decreased thermoregulatory reserve and are predisposed to hyperthermia and heat stroke. Another consequence of impaired sweating includes failure to recognize hypoglycemia (see Hypoglycemia During Insulin Therapy). This is a serious problem because one of the warning signals of insulin reaction is lost. Many elderly patients, including those without diabetes mellitus, already have impaired sympathetic responses as a result of aging rather than diabetes.
Cardiovascular Autonomic Neuropathies
In addition to abnormalities of innervation that result in abnormal cardiovascular reflexes, diabetic cardiac denervation apparently accounts for the phenomenon of painless myocardial infarction, which is said to occur in more than 30% of diabetic patients who experience an acute event. Diagnosis is difficult unless acute electrocardiographic changes are present. Precipitation of unexplained ketoacidosis or myocardial failure may direct attention to these secondary events.
Resting Tachycardia
Heart rates of 90 to 100 beats/min are common in patients with autonomic neuropathy; occasionally even higher rates are observed. Normal sleep-related bradycardia is absent. Parasympathetic damage is the apparent explanation; the sympathetics appear to be less affected. A β-blocker is useful if therapy is needed. In severe cases, the tachycardia “improves” over the years as denervation becomes more complete and the sympathetics are also lost. Once these abnormal cardiovascular reflexes have developed, there is a marked decrease in 5-year survival. Sudden death, not attributable to myocardial infarction, has been described in many such patients (89).
Postural Hypotension
The most readily recognized troublesome cardiovascular abnormality is postural hypotension (see Chapter 89). The patient may complain merely of dizziness or faintness on standing, or the problem may be more severe, with visual disturbances and syncope. These symptoms may be confused with episodes of hypoglycemia. Remarkably, some patients with fairly marked postural hypotension are asymptomatic.
On initial examination, every diabetic patient should be checked for a postural decrease in blood pressure. In addition, a check for postural hypotension should be made whenever a potentially aggravating condition occurs. The onset or aggravation of postural hypotension is often associated with the beginning of therapy with a variety of drugs often used in diabetic patients, such as antihypertensive drugs, including diuretics, vasodilators such as nitrates, and antidepressants. Occasional diabetic patients may be unable to tolerate effective dosages of these drugs because of this problem.
P.1327
The mechanism of this disorder is thought to reside in the efferent limb of the baroreceptor arc secondary to damaged sympathetic vasoconstrictor fibers in the splanchnic bed, muscles, and skin. Diminished plasma renin responses to postural change have been noted in such patients, as have abnormalities of plasma norepinephrine, but the role of these defects is unclear.
For patients with severe postural hypotension, the most useful drug has been the mineralocorticoid fludrocortisone (Florinef). In dosages of 0.1 to 1.0 mg/day, the drug is often helpful, but because one of its actions is to expand fluid volume, it can precipitate cardiac failure or produce severe hypertension in the recumbent state. Hypokalemia is common, requiring monitoring of electrolytes and appropriate management (K+-sparing agents, supplements). Refractoriness may eventually occur. The β-adrenergic agonist midodrine (ProAmatine) has been marketed for the treatment of orthostatic hypotension caused by a variety of causes, including diabetes. The drug requires careful dosing (starting at 2.5 mg twice a day) but seems useful (90). In mild cases, the simple advice that the patient assume upright positions slowly by sitting on the edge of the bed after recumbency may help avoid syncopal episodes; continuing postural hypotension, although readily documented, may not be especially symptomatic and may not warrant therapy (also see Chapter 89).
Amyotrophy (Proximal Asymmetric Motor Neuropathy)
Amyotrophy is a rare, but devastating, complication of diabetes mellitus that is probably a proximal motor neuropathy. Severe asymmetric proximal muscle weakness and pain usually affect the pelvic girdle and thigh muscles, although upper truncal musculature can also be involved. The typical patient is a middle-aged or elderly type 2 diabetic with mild disease. Men are affected more often than women. Onset may be fairly rapid, and a low-grade fever and an elevated erythrocyte sedimentation rate may be present. Cerebrospinal fluid protein concentration may be very high. Muscle biopsy shows fiber degeneration. Electromyography shows a pattern typical of motor denervation. Weight loss of moderate degree is common but may be severe (44 to 66 lb [20 to 30 kg]). Prognosis for spontaneous improvement over 2 to 6 months is good, but complete recovery often requires 24 months or longer and significant residual effects are common.
Infections
Diabetic patients are more prone to infections than are nondiabetics. Clinicians often encounter patients who have experienced repeated bacterial or fungal skin infections (carbuncles, furuncles, external otitis, moniliasis) or gastrointestinal moniliasis at some time before the diagnosis of diabetes was made or in association with uncontrolled hyperglycemia. Once established, infections in the diabetic are difficult to treat and patients are prone to develop complications. Elevated blood glucose appears to be a major cause of this problem. Experimentally, hyperglycemia (blood glucose levels greater than 250 mg/ 100 mL) inhibits the phagocytic activity of granulocytes and the immune function of lymphocytes, factors that may contribute to lowered host resistance. Consequently, control of blood sugar should be part of any treatment program for an infection. This conclusion is supported by studies of perioperative infections in diabetic patients undergoing coronary artery bypass surgery (91).
Urinary tract infections are an especially troublesome problem in diabetic patients. Although infections are not clearly increased in incidence, a greater prevalence of complications is obvious. Half of all cases of papillary necrosis occur in diabetics. Diabetic patients also seem prone to develop infections with unusual pathogens. However, there is no evidence that treatment of asymptomatic bacteriuria in diabetics is worthwhile. Development of pyelonephritis is an indication for immediate hospitalization and vigorous antibiotic treatment; the risk of a renal abscess is a special hazard for the diabetic.
Skin infections caused by Candida are common in diabetics, especially those with type 2 disease who are obese. These infections require therapy with a local antifungal agent (see Chapter 117) and control of hyperglycemia.
Major soft-tissue infections in diabetic patients require prompt hospitalization and treatment with parenteral antibiotics. However, most infections encountered involve the lower extremities, usually the feet, and are not complicated by serious conditions such as osteomyelitis or gangrene (see Neuropathic Foot Ulcers). Until several years ago such uncomplicated infections were treated by hospitalization and a combination of systemic antibiotics, often for several weeks. It is now clear that most of these less-severe infections can be treated with an oral antibiotic and without hospitalization (92).
More than half of infections of the feet are acute, with concomitant skin ulceration; others are acute infection of a previously uninfected chronic foot ulcer, whereas a lesser number are abscesses (often paronychias) or cellulitis. Although cultures of such lesions are of limited usefulness, they can be obtained by swab, aspiration, or curettage, the latter being more likely to reveal anaerobes. Aerobic gram-positive organisms are present in more than 50% of lesions, and aerobic gram-negative organisms are present in 15% to 25% of lesions. On average, slightly more than two organisms are isolated, and anaerobes are present in 15% of cases. Staphylococcus is the most common organism, followed by Streptococcus and the common gram-negative organisms Klebsiella, Pseudomonas, and Proteus. Corynebacteriashould not be considered as
P.1328
merely contaminants; they are the only organism present in many cases.
Clindamycin or cephalexin given orally cures 75% of these infections clinically and bacteriologically, even when treatment is without reference to culture results; another 15% to 20% of infections are greatly improved, and some 5% to 15% are treatment failures (92). Other antibiotic regimens yield similar results (93). The value of topical antibiotic preparations (povidone-iodine; silver sulfadiazine) is controversial (94).
It must be emphasized that such results can be expected only in these less-severe infections. Gangrene, severe ischemia, crepitus, and persistent fever are indications for hospitalization. Also to be emphasized is the need for proper adjunctive local care of the foot: elevation of the affected extremity, limitation of ambulation, proper foot hygiene, and, of course, surgical drainage, if appropriate.
Foot Problems
A number of common foot problems (e.g., bunions, calluses, corns, fungal infections, and ingrown toenails) occur in diabetic patients and can lead to devastating complications. Prevention through proper foot care and early recognition and treatment are important considerations in the long-term care of every diabetic patient. These problems are discussed in detail in Chapter 73.
Neuropathic Foot Ulcers
Epidemiology and Etiology
Diabetic foot ulcers and lower extremity amputations occur in 15% of diabetics. The problem is multifactorial. Enormous morbidity and expense are involved. A major technical review of preventive foot care and the ADA position paper that followed have been published (95,96), as have the conclusions of a consensus conference on Diabetic Foot Wound Care (97).
Foot ulcers are typically plantar and occur at the point where weight-bearing is greatest. They are now considered to be primarily the manifestations of diabetic neuropathy, but high plantar pressures and low tissue oxygen tension and blood pressure at the toe are other independent factors (98). Additional risk factors include hammer/claw toe and other acquired foot deformities. Although the diabetic is certainly prone to vascular (arterial) insufficiency and neuropathy and although the presence of vessel disease often contributes, the origin of the problem is primarily the sensory deficit. Neuropathy often is asymptomatic until an ulcer develops. Because the patient does not perceive pain normally, unappreciated trauma occurs, for example, from new or poorly fitting shoes that produce pressure points that go unrelieved and end in penetrating abrasions. Wounds can also result from skin penetration by foreign materials or from accidents during self-trimming of toenails. In addition to the sensory deficit, simultaneous motor weakness of extensor or flexor muscles together with proprioceptive defects can also contribute to anatomic deformity that in turn produces pressure points and ulceration.
Altered motor nerve function leads to muscle atrophy and tendon shortenings, which result in chronic toe flexion and finally hammertoe deformity. This anatomic change shifts weight from the padded ball of the foot to the metatarsal heads, where calluses form and contribute to the formation of new pressure points. The calluses themselves may develop fissures, which further promote ulceration. A contributor to the problem is diabetes or age-related loss of adipose tissue from the submetatarsal “padded” areas of the foot. This adipose padding normally distributes weight of the foot; loss of this tissue is a major anatomic destabilizer.
Management
Prevention is the mainstay of management. A standardized but simple nylon filament touch perception test allows detection of the sensory impairment, which is a major predictor of the risk of foot ulceration. Patients who cannot feel the filament are at about 10 times greater risk of ulceration and 17 times greater risk of amputation. Identification of such patients and institution of preventive measures may be useful in avoiding amputations. When an ulcer is present, decreased weight-bearing (“off loading”) is an essential part of treatment. This is best accomplished with a total contact cast, the exact nature of which is best left to the treating practitioner's preference based on experience and technical expertise. Infection is not invariably present but when present is almost always a mixture of aerobic, facultatively anaerobic, and anaerobic organisms (98). In the presence of infection, a total contact cast is contraindicated. Antibiotics of choice and their method of administration are discussed under Infections. Intravenous antibiotics and hospitalization have been standard therapy for many years, but recently oral antibiotics in a home-based setting have been shown to be effective for most patients. Because callus formation aggravates the tendency to increase local pressure and worsens ulceration, regular debridement is essential. Some patients can be taught debridement techniques, which may at least delay the intervals between visits to the caregiver for this purpose, but usually periodic professional assistance is essential. Such care is often best provided by podiatrists (see Chapter 73). Fitting of custom-made molded shoes is helpful and is essential in some cases for prevention of ulceration or its recurrence. Remarkably, with proper treatment the ulcer may heal completely and not recur. However, recurrence is likely as long as the anatomic distortion or continued point pressure is left unmodified. Plantar injection of silicone
P.1329
reduces risk factors for ulceration but has yet to be shown in clinical trials to be effective in preventing actual ulceration (99).
Neuropathic Arthropathy (Diabetic Charcot Foot)
Neuropathic arthropathy, a complication of diabetes, is often unrecognized or misdiagnosed. The disorder, preceded by a sensory neuropathy, is a progressive degenerative change of the bony structure of the foot, most often involving the tarsal and tarsometatarsal joints (60%) but also the metatarsophalangeal joints (30%) and the ankle (10%). The prevalence has been estimated at 1 in 680 cases, but the disorder is probably more common. The patient presents with a swollen foot, often attributed to or associated with recent trauma. The foot may be painful or may be remarkably free of pain, considering the appearance. Examination shows moderate to gross deformity of the foot with rocker-bottom subluxation of the midtarsal region or subluxation of the metatarsophalangeal joints. Usually, the foot is erythematous and warm to the touch. An infected neuropathic ulcer may be present. More often than not, the pulses are intact. Clinicians unfamiliar with this presentation are likely to diagnose some other type of inflammatory arthritis or osteomyelitis, and their impression may be apparently verified by radiographic findings. In these early stages, the radiographs often show severe osteoarthritis, but as the disease progresses, there is complete destruction of the involved joints with resorption of the metatarsal heads and phalangeal diaphyses. Various other bony changes occur, including fractures, joint effusions, and subluxations. When these changes are at the maximal stage (i.e., when soft-tissue involvement is most prominent), the diagnosis of osteomyelitis is often entertained, especially when there is an associated, often infected, ulcer. Synovial biopsy showing a thickened synovium containing osseous debris may provide the correct diagnosis and avoid the necessity of embarking on a prolonged and difficult course of antibiotic therapy for suspected osteomyelitis.
Diabetic Charcot foot may also be confused with the changes associated with osteoarthritis and gouty, rheumatoid, and psoriatic arthritis. Consultation by an orthopedic surgeon, rheumatologist, or podiatrist to confirm the diagnosis and assist with therapy is almost always indicated.
Treatment is based on the cessation of further trauma to the affected area, which is best accomplished by elimination of weight bearing. Hospitalization may be necessary for this purpose. Reduction of edema and signs of inflammation may take several weeks. Immobilization with a cast may be helpful but should not be undertaken in the acute stage, and if used, should be done with great care to ensure the integrity of the areas covered by the cast. Simpler bootlike devices may also be used. Crutches can be used at this point, followed eventually by a walking cast. Up to 4 months of treatment may be required. Thereafter, molded or contoured shoes are essential to proper long-term management. Surgical intervention is inadvisable, although occasionally a stabilization procedure may be required if conservative therapy fails. Amputation is not indicated unless osteomyelitis unequivocally coexists or the entire process fails to respond to prolonged conservative efforts. Despite the discouraging appearance of the foot at its worst stages, sufficient healing and stabilization to produce a useful foot can be anticipated.
Gastrointestinal Dysfunction
Most disorders of the gastrointestinal tract in diabetes are related to disturbances of motility. Esophageal motor dysfunction can be demonstrated on testing but is usually not a clinical problem.
Gastroparesis (atony of the stomach) is often asymptomatic but may be troublesome. Symptoms include anorexia, early satiety, postprandial fullness and bloating, and, occasionally, vomiting. Delayed and unpredictable emptying of the stomach may produce irregular diabetic control in already difficult to manage patients with either type 1 or 2 diabetes. Diagnosis is apparent—sometimes as an incidental finding—on barium radiograph of the upper gastrointestinal tract. Nuclear scintigraphy is the most sensitive method to detect the presence of atony. Gastroparesis has long been thought to be a late complication associated with poor prognosis for survival, but there is no evidence to support this notion (100). Metoclopramide (Reglan) 10 mg three times daily may be helpful (see Chapter 43). In intractable cases, gastrectomy may be needed. Alternatively, insertion of a gastric pacemaker may also ameliorate symptoms and improve quality of life (101).
Small-bowel dysfunction is common and symptomatic, leading to diabetic diarrhea. Typically, the diarrhea is nocturnal. Fecal incontinence, a result of impaired sensation of rectal distention, may occur and is very distressing. The disorder tends to be episodic, with attacks lasting from a few days to weeks or rarely months. Watery brown diarrhea, usually without steatorrhea, is typical. On barium radiograph of the small bowel, the findings are those of disturbed motility. Despite the distressing symptoms, the patient appears well; weight loss is uncommon. When steatorrhea occurs, pancreatic exocrine insufficiency and sprue syndrome, more common in diabetic patients than in the general population, should be considered. Fully developed sprue is associated with gross evidence of malabsorption. A trial of antibiotic therapy (e.g., tetracycline, 250 mg four times a day for 2 weeks) may improve the diarrhea and the malabsorption if the latter is caused by small-bowel stasis and bacterial colonization. Symptomatic treatment
P.1330
with antispasmodics (e.g., loperamide [Imodium]) may be useful, especially when attacks of diarrhea are short lived.
Large-bowel complaints, especially of constipation, are common in the elderly. It does not appear that diabetic patients are especially prone to any additional problems in this regard.
Patients with poorly controlled diabetes may develop fatty changes of the liver. Nonalcoholic steatohepatitis, with hepatomegaly or elevations of serum aminotransferases, may occur. Effective control of blood sugar results in disappearance of these abnormalities if diabetes is the cause. However, diabetes and obesity often coexist; fatty liver is also common in uncomplicated obesity.
Genitourinary Problems
Bladder Dysfunction (Neurogenic Bladder)
The symptoms of bladder dysfunction in diabetic patients are often overlooked. Onset is insidious and occurs over many years. Eighty percent of patients have clinical evidence of neuropathy affecting other systems. The first clinical manifestation of bladder dysfunction is an increase in the interval between voiding until urine is passed only twice or even once daily. A need to strain, slow stream, dribbling, and sensation of incomplete voiding may be present. These symptoms should be routinely solicited from diabetic patients, especially when there are symptoms or signs of peripheral neuropathy.
Demonstration of residual urine is the hallmark of clinically symptomatic cystopathy, but many diabetic patients, when studied by cystometric techniques, have objective evidence of neurogenic involvement and a grossly enlarged bladder well before symptoms are evident. At this stage, residual urine is not present and other urinary tract abnormalities (recurrent infections) are not evident. If large volumes of residual urine do develop, patients become prone to infection (see Chapter 36) and incontinence (see Chapter 12). Patients suspected to have cystopathy should be referred to a urologist for evaluation and recommendations about treatment.
Erectile and Sexual Dysfunction
The frequency of erectile dysfunction, formerly termed impotence, is high in diabetic men, perhaps 50% to 60% overall. This complication, like many others in diabetic patients, is related to duration of disease. The problem is usually caused by a type of autonomic neuropathy involving the pelvic parasympathetic nerves, but impaired blood flow is the cause in some cases. Several studies report that sexual function in diabetic women appears to be unaffected by the disease. Others assert that many diabetic women lose the ability to achieve orgasm (102).
The evaluation and management of erectile and sexual dysfunction are described in Chapters 6 and 85. Psychogenic factors probably account for a significant fraction of cases, but there is no evidence that psychogenic problems are more common in diabetic patients. The onset of erectile dysfunction in diabetics usually is slow (6 months to several years), often associated with retrograde ejaculation; impotence eventually becomes complete. Despite this, libido is characteristically retained. Although patients with psychogenic impotence often report nocturnal erections and emissions, these are absent in patients with diabetic impotence. An important point on clinical examination is that testicular sensitivity to pressure and pain is retained in men with psychogenic impotence but is often greatly diminished or lost in the diabetic in whom accompanying sensory neuropathy is common.
A variety of drugs, especially ones that are often used in diabetics, may cause erectile dysfunction. The most common offenders are antihypertensives and antidepressants.
Nephropathy
Progressive renal failure is a major life-threatening complication of diabetes. The relationship between hyperglycemia and the development of microangiopathy with eventual nodular glomerulosclerosis (Kimmelstiel-Wilson disease) is unequivocally established, as strong clinical and experimental evidence has accumulated in favor of such a relationship. One of the earliest indications of incipient diabetic nephropathy is proteinuria, first manifest as microalbuminuria. The detection and significance of proteinuria and of microalbuminuria in diabetic patients are discussed in Chapter 48.
Diabetic patients with even minimal elevations of serum creatinine greater than 1.1 mg/dL, but not those with normal renal function, are at increased risk of acute renal failure from contrast media used in various radiographic procedures. Although the risk is only moderately increased (approximately 10% to 15% in the highest risk patients versus 5% in low-risk patients and less than 2% in those with no renal disease and in nondiabetic subjects), these procedures can often be replaced by others with less risk (magnetic resonance imaging, sonography). If contrast media are to be used, the dosage should be minimal and the patient well hydrated. Diuretics and ACE inhibitors should be reduced or briefly withheld, and metformin should be stopped temporarily. The clinical course of diabetic nephropathy and the impact of failing renal function on insulin requirement and on the dosage of oral hypoglycemic drugs are discussed in Chapter 52.
Skin Lesions
An uncommon abnormality known as necrobiosis lipoidica diabeticorum occurs in perhaps 0.3% of all diabetics,
P.1331
seemingly unrelated to glycemic control. Typically, lesions occur on the anterior legs, but also can occur on the arms, trunk, and face. The lesions, described as atrophic plaques, are sharply demarcated, orange or red, and often ulcerate. On biopsy, diagnostic findings include destruction of collagen fibers. Until recently no agent seemed to be useful, but several reports indicate that these lesions respond rapidly to tretinoin-containing creams; pentoxifylline (Trental), 400 mg twice daily, also seems to be effective.
Retinopathy
Diabetic retinopathy has become one of the leading causes of blindness in the United States. In type 1 diabetic patients, some degree of retinopathy can be detected by angiography, the most sensitive technique, after as few as 1 to 2 years in 10% of patients. By 10 years, retinopathy is evident in 50% of cases by ophthalmoscopy, with a 70% prevalence by angiography. At 25 years, nearly all patients can be shown to have some degree of retinopathy. By the time diabetes has been present for 15 to 20 years, approximately 33% of patients have severe disease and another 50% have obvious, but lesser degrees of, progressive retinal involvement. Remarkably, not all cases of early retinopathy are progressive. These figures represent the era prior to the use of intensive therapy to control blood sugar (45).
Type 2 diabetic patients also develop retinopathy, apparently with less frequency, but clearly related to duration and severity of the hyperglycemia. Hyperglycemia, as indirectly measured by glycosylated hemoglobin, clearly predicts the incidence and progression of diabetic retinopathy in type 2 diabetes (103,104).
In patients with retinopathy in whom intensive therapy is initiated, a paradoxical worsening of the retinopathy can be seen. Patients whose retinopathy is mild to moderate usually have no associated visual loss, and long-term intensive treatment of glycemia counterbalances the early worsening (105).
Blindness in Diabetic Retinopathy
The visual loss in diabetic retinopathy is potentially even more severe than is blindness as a result of other causes. Many people who are legally blind (defined as visual acuity less than 20/200 in both eyes) from causes other than diabetes have slow onset of visual loss, thus allowing time for adaptation. In addition, they often retain reasonably full visual fields and visual acuity at or close to the legal limit. Such patients can see well enough to ambulate independently and perform a variety of common activities (self-care, housework). With the aid of special devices they may even be able to read newsprint and to engage in some occupations. In contrast, the visual loss from diabetic retinopathy is often caused by sudden hemorrhage or retinal detachment and often leaves the patient with only light perception. In addition, the diabetic often already has other complications of the disease when blindness develops. Although total blindness afflicts only some diabetic patients, a larger number have some degree of loss of visual acuity caused by macular edema, the most common cause of visual loss in diabetics.
Ocular Symptoms
Diabetic patients who experience symptoms of visual disturbance are not necessarily experiencing a catastrophic complication and deserve reassurance along these lines. Like others, diabetic patients develop changes of visual acuity such as a change in refractive error and astigmatism. In addition, they can experience decreased visual acuity as a result of marked changes in blood sugar (e.g., as the lens swells during acute normalization of blood sugar after prolonged hyperglycemia). However, sudden or persistent change of visual acuity requires examination by an ophthalmologist, especially when advanced retinal disease, proliferative diabetic retinopathy (PDR), is present. A common cause of visual loss in PDR, macular edema, is not readily detectable by the direct ophthalmoscopy available to internists and requires binocular slit lamp or stereoscopic fundus photography by an ophthalmologist.
Sudden painless loss of vision in PDR demands urgent ophthalmologic consultation. This symptom is often caused by hemorrhage from proliferating new vessels or from retinal detachment. Lesser degrees of hemorrhage may cause floaters or cobwebs. Another complication in PDR is outflow obstruction of the aqueous humor produced by fibrous scar tissue extending into the angle of the anterior chamber, causing a marked rise in intraocular pressure and acute (neovascular) glaucoma with severe pain. Loss of vision occurs unless emergency therapy is given.
Types of Retinal Disease in Diabetic Patients
The current classification of diabetic retinal disease is nonproliferative (or background), preproliferative, and PDR. The latter more advanced stage is the point at which sudden and massive visual loss becomes a problem.
Nonproliferative and Preproliferative Retinopathy
The earliest lesions—readily visible with an ordinary ophthalmoscope—are in the region of the macula: microaneurysms, punctate retinal hemorrhages, hard exudates, soft exudates (cotton wool), and so-called intraretinal microvascular anomalies (IRMAs). Both microaneurysmsand small-dot intraretinal hemorrhages appear as red dots, and both tend to fade within months. Blot hemorrhages are larger. Distinction can best be made by fluorescein angiograms, in which only microaneurysms light up. This
P.1332
procedure, performed by an ophthalmologist, often identifies extensive intraretinal disease when only a few abnormalities are evident by ophthalmoscopy.
Hard exudates are glistening yellow or white lipid deposits located in the outer retinal layers. If they are greater than 1 disk diameter from the macula, they are not ominous. Soft exudates are areas of ischemia or infarction of the nerve fiber layer; they disappear within a few months. IRMAs are dilated hypercellular vessels that are thought to represent either dilated capillaries or intraretinal vascular proliferation. These abnormal vessels are identifiable using the green filters of an ordinary ophthalmoscope, but are best seen in the secondary phase of fluorescein angiograms, during which they leak dye into the retina. IRMAs occur adjacent to areas of capillary closure. Their identification is not essential in the routine examination.
Whereas changes of background retinopathy indicate capillary damage and leakage, preproliferative changes (many soft exudates, extensive hemorrhages, IRMAs, and venous bleeding from enlarged dilated [beaded, sausage-link] retinal veins) indicate areas of intraretinal vascular occlusion with resulting nonperfusion. Eyes with retinal ischemia and moderate to severe preproliferative changes have a 50% chance of developing new vessel proliferation (neovascularization) within 1 year.
Macular Edema
It is not usually appreciated by the general clinician that even without proliferative disease, macular edema may result in visual loss as severe as the 20/200 level. Spontaneous improvement is not common but may occur. Visual acuity can also become poor because of lack of proper perfusion of the perifoveal capillaries. In this instance, visual acuity may be as low as 20/200 in the absence of macular edema. Fluorescein angiography reveals the cause of poor visual acuity to be a result of the lack of perfusion of the perifoveal capillaries. In the absence of accompanying proliferative disease, patients at this stage can usually ambulate freely and can engage in some occupations. Ability to read a newspaper, except with a vision aid, is unlikely, and the patient will have to stop driving.
Proliferative Retinopathy
At this stage of retinal disease, new vessels and accompanying fibrous tissue extend from the retinal substance and grow along the inner retinal surface and the posterior surface of the vitreous gel, often causing contraction of the gel and traction on the vessels and the retina. This process creates the conditions for retinal detachment and hemorrhage into the vitreous.
In addition, in advanced PDR, growth of new vessels and scar tissue into the angle of the eye may cause acute glaucoma (see Ocular Symptoms). When the new blood vessels grow out of the surface of the optic nerve heads, they are called new vessels on the disk; elsewhere in the retina, usually extending from large vessels, they are called new vessels elsewhere.
The National Diabetic Retinopathy Study (106) not only established the efficacy of photocoagulation therapy (see Photocoagulation Therapy) but also defined high-risk characteristics as follows: new vessels on the disk greater than 25% of the optic disk area, any new vessels on the disk with preretinal or vitreous hemorrhage, or new vessels elsewhere equal to or exceeding 50% of the disk area with preretinal or vitreous hemorrhage. The presence of high-risk characteristics increases the chance of blindness to 30% to 50% within 3 to 5 years unless appropriate photocoagulation therapy is given.
Treatment of Diabetic Retinopathy
Two forms of surgical therapy are now available for the treatment of proliferative retinopathy and its complications. Photocoagulation is of proven value in the prevention of visual loss caused by proliferative disease, and vitrectomy restores and appears to stabilize vision after hemorrhage and/or retinal detachment.
Photocoagulation Therapy
Argon laser therapy reduces severe visual loss by nearly 60% over 5 years in patients with proliferative disease. Multiple (1,200 to 1,600) 500-µm burns are placed in the retinal periphery. The Early Treatment Diabetic Retinopathy Study (107) evaluated panretinal photocoagulation (and laser photocoagulation using 450- to 650-µm widely spaced burns) to determine whether such therapy affects the course of disease in eyes with high-risk characteristics or preproliferative diabetic retinopathy. Considerable benefit was shown.
Diabetic macular edema is also treated with photocoagulation therapy. Leaking microaneurysms and other lesions in the macula are treated with 50- to 100-µm burns. The Early Treatment Diabetic Retinopathy Study showed a reduction of visual loss caused by macular edema of 50% over 3 years.
Patient Experience
Photocoagulation therapy is an office procedure, usually performed in several sessions. Ordinarily only topical (corneal) anesthesia is necessary. Occasionally, some pain is experienced, in which case a local anesthetic is injected into the retro-orbital tissues to allow a completed pain-free procedure.
Vitrectomy
Hemorrhage into the vitreous is the usual indication for vitrectomy, a procedure that removes old blood and opaque vitreous and can be combined with cataract extraction. Retinal detachment resulting from traction bands that are formed in the vitreous is another indication for vitrectomy.
P.1333
Other repair procedures can be attempted. The results of vitrectomy can be dramatic in restoring sight after vitreous hemorrhage. Currently, however, vitrectomy is used only in severely diseased eyes. Recovery of near-normal vision is the exception rather than the rule, but the results are definitely worthwhile in many patients.
|
TABLE 79.7 Reasons for Referral of Patients with Diabetes Mellitus to an Ophthalmologist |
||
|
Every effort should be made to control hypertension in an attempt to prevent retinal hemorrhages. Lifting of heavy objects, jarring exercise, and exposures to high altitudes may also increase the risk of hemorrhages and should be avoided in patients with PDR, but commercial airline travel presents no increased risk. The use of aspirin therapy does not increase the risk or severity of intraocular hemorrhage and isnot contraindicated in PDR (105).
Role of the Primary Care Physician/Internist Versus That of the Ophthalmologist: Indications for Referral
If there is no evidence of retinal disease, diabetics should have yearly examinations of their eyegrounds, with their pupils dilated. If that cannot be done for any reason or if there is concern about the validity of the examination, referral to an optometrist or ophthalmologist is indicated (Table 79.7). Because most diabetic retinopathy occurs within several disk diameters of the macula, most lesions are visible by examination with the direct ophthalmoscope after dilation of the pupils. Nonophthalmologists often defer this examination out of concern over precipitating acute angle-closure glaucoma. Such reluctance is not warranted because this complication is rare at any age and is hardly ever seen before age 40 years. A drop of dilating solution (2.5% phenylephrine or 1% tropicamide) in each eye is sufficient and causes only sensitivity to bright light (requiring dark glasses) that lasts but a few hours.
If only minimal nonproliferative diabetic retinopathy is present, the patient does not need to be referred immediately if visual acuity is normal. Early preproliferative changes warrant prompt referral to a general ophthalmologist, whereas more extensive changes and proliferative changes should be followed by an ophthalmologist who is also expert in photocoagulation. In addition to examining the retina, the condition of the lens should be evaluated during the examination of the eye because senile cataracts occur prematurely in diabetic patients and metabolic cataracts result from chronic elevation of blood glucose levels.
Periodic checks of intraocular pressure to detect glaucoma, which is more prevalent in diabetics, should also be part of routine health maintenance. These evaluations can be made either by an optometrist or an ophthalmologist (see Chapter 108). Recommendations for followup are given in Table 79.7 and in detail elsewhere (108).
Tracking and Coordinating Care in the Diabetic Patient
Because management of the diabetic patient is complex and multidimensional and because numerous interventions improve healthcare outcomes, it is important for the primary care practitioner to have a systematic approach to monitoring and coordinating the various aspects of longitudinal care of the diabetic patient. A flow sheet or computerized tracking and reminder system that targets treatment and treatment changes, critical laboratory and physical examination parameters, and key referrals is highly recommended (109; and Fig. 1.2). Items that should be tracked include serum glucose and HbA1c; serum lipids; urine protein and renal function; foot examinations; and eye examinations. Prominent display of a primary care front sheet that includes a problem list (see Fig. 1.1) is particularly important in diabetic patients, because diabetics are likely to have a plethora of associated conditions (see Complications of Diabetes Mellitus). Finally, a preventive care profile and flow sheet (see Fig. 14.3) is as important in care of the diabetic as it is in the care of other patients. Some preventive care recommendations are more specific for diabetic patients, such as recommendations for pneumococcal vaccination and yearly influenza vaccinations at any age and more aggressive lipid management. Knowledge of smoking status and smoking cessation counseling (see Chapter 27) are particularly important in the diabetic patient, whose risk of atherosclerotic disease is very high.
P.1334
Specific References*
For annotated General References and resources related to this chapter, visit http://www.hopkinsbayview.org/PAMreferences.
P.1335