Screening for Type 2 Diabetes Mellitus

Diabetes mellitus, a major cardiovascular risk factor as well as an important cause of chronic kidney disease, proliferative retinopathy, and neuropathy, is becoming a problem of epidemic proportions in the United States and worldwide. Currently, over 9% of the US adult population is estimated to have some form of the condition, and the proportion is predicted to double or triple by 2050. Despite the emerging enormity of the problem, the availability of convenient and accurate diagnostic testing, and effective means of treatment during the asymptomatic phase, there remains considerable disagreement as to the value of population-wide screening for diabetes. The primary care physician needs to understand the elements of the debate and the nature of available evidence so that an effective, personalized screening decision can be made for individual patients and one’s patient panel as they present for preventive care. This chapter focuses on type 2 (formerly referred to as adult-onset) diabetes, by far the most prevalent form of the disease (type 1 disease and gestational diabetes are dealt with in the chapter’s Appendix 93-1-93-93).

EPIDEMIOLOGY, PATHOPHYSIOLOGY, AND RISK FACTORS (1, 2, 3 and 4)

Depending on the diagnostic criteria used, the prevalence of all types of diabetes in the United States is estimated to be in excess of 9% and increasing rapidly, largely as a consequence of obesity. Type 2 diabetes is nearly 10 times more common than type 1 disease; incidence is especially high among Latino and black populations and among the poor, due in part to high prevalences of obesity in these communities. Pathophysiologic mechanisms of type 2 disease include a blunted β-cell response to glucose, a defect at the insulin receptor, and a defect in hepatic uptake of glucose that contributes to glucose intolerance; tissue insulin resistance is characteristic, especially in the obese. Type 2 diabetes has a much stronger genetic component than does type 1 diabetes; concordance in identical twins is greater than 90%.

The overwhelming risk factor for type 2 diabetes is obesity resulting from overnutrition and lack of exercise. Eighty percent of diabetic adults are obese or have a history of obesity. Among adults who are at least 25% greater than their ideal body weight, one of every five has elevated fasting blood sugar levels, and three of every five have abnormal results on glucose tolerance tests. Increasing waist circumference confers increased risk (see later discussion). Such truncal obesity increases insulin levels and decreases the concentration of insulin receptors in tissue, including skeletal muscle and fat. The relationship between the concentration of insulin receptors and glucose tolerance is modified, however, by intracellular sequences following insulin binding that are poorly understood. Exercise increases the concentration of insulin receptors, and a sedentary lifestyle is associated with glucose intolerance. Regular exercise has been shown to be associated with a decreased incidence of type 2 diabetes after adjustments for body mass index (BMI).

Steroids reduce receptor affinity for insulin, as do uremia and hepatic failure. Other drugs can impair glucose tolerance by further diminishing the sluggish response of β-cells to glucose. These include thiazide diuretics, β-adrenergic blockers, α-adrenergic stimulants, and phenytoin. Prostaglandin inhibitors, including indomethacin and salicylates, may increase β-cell release of insulin.

Persons with impaired glucose tolerance (IGT) are also at greater risk for frank type 2 diabetes (see later discussion). Although long-term follow-up studies indicate that as many as half of those with IGT will have normal glucose tolerance test results 5 to 10 years later, somewhere between 1% and 5% of these patients become diabetic each year. Gestational diabetes confers increased risk of developing type 2 diabetes, estimated to be as much as seven times that for women who remain normoglycemic during pregnancy.

NATURAL HISTORY OF DIABETES AND EFFECTIVENESS OF THERAPY (4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19 and 20)

Natural History

The natural history of diabetes mellitus has been difficult to define because the condition is so heterogeneous, especially as regards type 2 disease, with a much more subtle and drawn-out clinical course. This problem has historically been confounded by studies that have defined diabetes according to varying degrees of glucose intolerance. Nonetheless, some patterns emerge.

Prediabetes: Impaired Glucose Tolerance/Impaired Fasting Glucose

In type 2 disease, the prediabetic state is quite variable and characterized by asymptomatic impaired fasting glucose (IFG) or impaired glucose tolerance (IGT—see quantitative definitions in later section). Only 15% to 30% of persons with prediabetes progress to frank symptomatic diabetes during 10 to 15 years, delaying diagnosis while increasing the risk of macrovascular complications (coronary heart disease [CHD], peripheral vascular disease, stroke), which can develop during this time as a consequence of accelerated atherogenesis. Characteristic microangiopathic complications of diabetes (retinopathy, nephropathy,
neuropathy) usually do not appear until much later in the course of illness subsequent to the development of frank clinical diabetes, being a function of the severity and duration of hyperglycemia. However, evidence of microvascular complications (e.g., retinopathy, microalbuminuria) is found at time of initial diagnosis in 10% to 20% of persons with type 2 disease, reflecting several years of disease progression in otherwise asymptomatic persons and delay in diagnosis. Intensive lifestyle interventions in people with IFG or IGT have been shown to significantly reduce the rate of progression to frank diabetes, often by as much as 42% to 58%.

Type 2 Diabetes

Typically, there is an asymptomatic detectable period for type 2 diabetes, previously estimated to be as long as 4 to 7 years, but more recently found in community-based screening studies to be considerably shorter (˜3 years), purportedly reflecting heightened attention to screening for major cardiovascular risk factors. Classically, onset of symptomatic diabetes may be heralded by manifestations of marked hyperglycemia (polyuria/polydipsia/polyphagia), but presentations may be more subtle and nonspecific, with fatigue or a manifestation of a complication (e.g., erectile dysfunction, paresthesias, microalbuminuria) being the first symptom or sign.

Cardiovascular disease, especially coronary artery disease, is the most common cause of death among diabetics. Autopsy studies indicate that coronary disease is two to three times more common among diabetics than among nondiabetics. Peripheral vascular disease is also very common—prevalence about 60%. Diabetics in the Framingham Study were shown to be four to five times more likely to have intermittent claudication and two to three times more likely to suffer the morbid consequences of stroke than were nondiabetics. Conversely, the presence of peripheral vascular disease in a nondiabetic person increases the risk for the eventual development of type 2 diabetes.

Retinopathy is the most common specific complication of diabetes; risk increases with the duration of diabetes, regardless of age at onset. Prevalence ranges from 40% to 80% among patients with known diabetes of 20 to 30 years’ duration. Because undiagnosed diabetes may be asymptomatic for many years while microvascular changes progress, approximately 20% of newly diagnosed patients with type 2 diabetes are found to have diabetic retinopathy. In one study of patients whose diabetes began before age 30 years, nearly all had retinopathy 25 to 30 years later, and about 50% had proliferative changes. Retinopathy was less prevalent among those whose diabetes began later in life and those not receiving insulin.

Renal disease has been reported in 15% to 80% of diabetic patients at autopsy. Renal failure is the cause of death in 6% to 12% of diabetic patients. Prevalence estimates vary widely from study to study, but it is clear that the risk for glomerulosclerosis with clinically evident functional impairment increases dramatically with the duration of disease.

Effectiveness of Therapy (see also Chapter 102)

Therapeutic efforts have focused on correcting glucose intolerance and addressing cardiovascular risk factors.

Treating Glucose Intolerance

The effect of glycemic control on important outcomes has also been the subject of intense study (see Chapter 102). Landmark randomized trials of patients with type 2 disease (e.g., United Kingdom Prospective Diabetes Study [UKPDS]) and meta-analyses of randomized trials involving patients with type 2 diabetes find that tight glycemic control (e.g., HbA1c <7.0) significantly reduces microvascular disease risk and produces a trend toward reduction in macrovascular complications, most notably with use of metformin. However, intensive glucose lowering (HbA1c <6.0%) may reduce risk of nonfatal myocardial infarction but at a cost of increased all-cause mortality (ACCORD study).

Prevention of diabetes among patients with impaired glucose tolerance is also an area of great interest. Interventions designed to prevent or delay diabetes have included pharmacologic and lifestyle approaches. In the Diabetes Prevention Program trial, intensive lifestyle intervention reduced diabetes incidence by 58%, whereas metformin reduced the incidence by 31%. A meta-analysis of 17 trials with more than 8,000 participants concluded that both lifestyle and pharmacologic interventions reduced the rate of progression to diabetes, with the former at least as effective as the latter. The number needed to treat for benefit was a low 6.4 for lifestyle.

Treating Cardiovascular Risk Factors

Of greater significance for reducing cardiovascular mortality among persons with type 2 diabetes is intensive multifactorial cardiovascular intervention. In a major randomized trial of tight glycemic control, aspirin, angiotensin blockade, and statin therapy in type 2 diabetics with microalbuminuria (Steno-2 trial), all-cause and cardiovascular mortalities were reduced by nearly 50%. Rates of end-stage renal disease and treatment-requiring proliferative retinopathy were also significantly lower.

As demonstrated in numerous controlled trials of persons at high risk of developing type 2 disease (e.g., those with gestational diabetes or IGT), a program of lifestyle modification (diet and exercise) significantly and often dramatically reduces risk of developing full-blown diabetes (in some cases by up to 50% or more). Similarly, in persons with established type 2 disease, lifestyle changes significantly reduce the risks of cardiovascular and microvascular complications, especially in persons at high risk.

Other effective means of improving outcomes include laser photocoagulation for reducing the risk of bleeding from proliferative retinopathy (see Chapter 209) and angiotensin blockade to preserve glomerular function, especially when instituted at the first sign of glomerular compromise (microalbuminuria), even in the absence of hypertension (see Chapters 26 and 102).

Effectiveness of Screening

The rising prevalence of type 2 diabetes, a long presymptomatic period, and the potential to reduce both microvascular and macrovascular morbidity and mortality have stimulated interest in mass screening for diabetes that would enable earlier implementation of outcome-modifying therapy. While a number of simulation studies and meta-analyses of observational studies have suggested potential benefit, others have not, resulting in widely differing recommendations from various authorities. The dearth of randomized, prospective studies is glaring and a major impediment to resolving the question of screening for diabetes.

Evidence from Randomized Trials

Helping to clarify these issues are results emerging from the first and only large-scale, long-term randomized controlled trial of screening high-risk persons for diabetes in primary care practices—the ADDITION study. It engaged hundreds of primary care practices and tens of thousands of their patients in the United Kingdom, Denmark, and the Netherlands in a study of screening high-risk persons for type 2 disease followed by randomization of those testing positive to usual or intensive disease-modifying therapy. The principal outcomes were cardiovascular and all-cause mortality. Being the only available
rigorously designed, large-scale study, it is worthwhile reviewing some of its important findings.

In a United Kingdom arm of the study (ADDITION— Cambridge) involving over 20,000 mostly white, at-risk adult patients ages 40 to 69 years in 33 primary care practices randomized to a single screening followed by treatment versus no screening or specific treatment, the diagnostic yield was 3%, and there was no significant difference found in all-cause, cardiovascular, diabetes-related mortality rates at 10 years of follow-up. The only significant difference in cardiovascular mortality between screened and unscreened populations in this intention-to-treat analysis was found among the 27% randomized to screening who did not show up (thought to reflect a healthy volunteer bias for screening and less interest in screening the less healthy).

Explanations put forward by the investigators for the lack of observed benefit include the low rate of undiagnosed diabetes in the study population (3%), the low cardiovascular mortality rate in the control population (50% below expected and believed reflective of successful cardiac preventive care efforts in that part of England), the much shorter than expected lead time for diagnosis of diabetes in the unscreened populations (3 years— found in another arm of the study), a single round of screening rather than multiple rounds, and an insufficient follow-up period (the mortality curves were only starting to diverge at 8 to 10 years of follow-up).

Reflecting on their results and taking into account those of randomized trials of intensive multifactorial cardiovascular interventions (which achieved significant reductions in all-cause and cardiovascular mortality), the ADDITION investigators concluded that “benefits to the population could be increased by detection and management of related cardiovascular risk factors alongside assessment of diabetes risk, repeated rounds of screening, and identification of non-attendees and strategies to maximize uptake of screening.” They also recognized that their findings did not rule out potential benefit from population-wide screening in settings where the incidence of type 2 diabetes is likely to be higher than that in their largely all-white, middleclass study population.

SCREENING TESTS AND CLASSIFICATION (3,11,15,21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31 and 32)

Diagnostic criteria are somewhat arbitrary for quantitative tests such as those used for screening and diagnosis of type 2 diabetes. Recommendations made by internationally recognized authorities such as the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus represent a balancing of practical considerations, correlations with disease pathophysiology, and trade-offs between sensitivity and specificity (see Chapter 2). Over the past two decades, thresholds for diagnosis have been reset, initially at higher levels to maximize diagnostic specificity and avoid overdiagnosis, but more recently, at lower levels as evidence accumulates of benefit from early diagnosis and treatment. In addition, a state of “prediabetes” has been defined to encourage lifestyle modification and prevention.

In diabetes, many of the same tests used for diagnosis are well-suited for screening. The most recent expert consensus recommendations adopted and promulgated by the American Diabetes Association (ADA) reflect important advances in diabetes detection and diagnosis, particularly the addition of a nonfasting glycosylated hemoglobin determination (hemoglobin A1c [HbA1c]) as a validated if not preferred diagnostic test for type 2 diabetes. These recommendations also include use of results from either a fasting plasma glucose (FPG) determination or a 2-hour plasma glucose measurement after a standard 75-g glucose load (oral glucose tolerance test [2-hour OGTT]).

In addition to defining diagnostic criteria, which have been widely accepted in the United States, the ADA suggests using the term impaired glucose tolerance to designate persons at risk for developing diabetes (i.e., having “prediabetes”), but not yet meeting the threshold for formal diagnosis. It is recommended that persons tested for type 2 diabetes be placed into one of three categories: “normal,” “increased risk for diabetes,” and “diabetes.”

As regards age of onset and frequency of screening, there are no data from prospective trials. Typical recommendations for onset and frequency of screening derive from modeling studies that seek to maximize yield while minimizing cost; they suggest every 3 to 5 years after age 45 years for persons at normal risk and starting earlier for persons at heightened risk. The ADDITION study chose to screen only persons deemed high risk.

Only gold members can continue reading. Log In or Register to continue