Canadian Task Force on Preventive Health Care

Full Text Review

Screening for Diabetes Mellitus in the Non-Pregnant Adult

Adapted by Marie-Dominique Beaulieu, MD, MSc, FCFP, Associate Professor, University of Montreal, Quebec, from materials prepared for the U.S. Preventive Services Task Force

These recommendations were finalized by the Task Force in January 1994

Up Contents

Overview

In 1979 the Canadian Task Force on the Periodic Health Examination concluded that screening for diabetes mellitus (DM) in non-pregnant asymptomatic adults without risk factors was not part of the periodic health examination.< 1> Review of further published evidence does not alter this conclusion. The screening test for DM in the absence of symptoms or risk factors has poor sensitivity and has never been associated with secondary prevention of diabetic complications (D Recommendation). Test characteristics are enhanced when applied to high risk populations, namely native Canadians, obese individuals and those with a strong family history of diabetes. Based on some evidence that treatment of overt diabetes may retard micro-vascular complications, a decision to screen in those populations could be considered. Screening pregnant women is discussed in Chapter 2 on Gestational diabetes.

Burden of Suffering

The prevalence of DM in the general Canadian population is estimated to vary between 2% and 2.7%.<2> Prevalence increases with age. Estimates from the Manitoba data base yielded rates of 0.8% in adults less than 45 years, 3.5% among those between 45 and 64 years of age, and 7.6% among those 65 and or older.<3> Prevalence is higher in native Canadian populations. A study conducted in Northern Quebec revealed that 6.2% of the Cree population suffered from DM, with a prevalence rate of 20% in women over 50 years of age.<4> Other risk factors for diabetes include family history, obesity, impaired glucose tolerance, and a previous history of gestational diabetes.

Diabetes may cause life threatening metabolic complications and is an important risk factor for other leading causes of death such as coronary heart disease, congestive heart failure and cerebrovascular disease. Diabetes is the most common cause of polyneuropathy with approximately 50% of diabetics developing neuropathic disease within 25 years after diagnosis.<5> Diabetes is responsible for 40% of all nontraumatic amputations in Canada.<6> Diabetic nephropathy has now become the leading cause of end-stage renal disease.<7> Diabetes is also the leading cause of blindness in adults.<7> (Screening for diabetic retinopathy in the elderly is reviewed in Chapter 78, screening for pancreatic cancer in Chapter 71 and screening for end-stage renal disease in Chapter 28). Individuals with this disease have higher hospitalization rates, longer hospital stays and more ambulatory care visits than persons without diabetes.

Most of this burden of illness is due to type II, or non-insulin dependent DM (NIDDM), which comprises approximately 80-90% of all cases. Type II diabetes generally occurs after age 30. Type I, or insulin dependent DM (IDDM), generally begins before this age. This discussion of the efficacy of screening and detection of diabetes mellitus does not include type I DM. The low prevalence of this disease and the fact that it has a brief presymptomatic period are reasons against screening for type I DM.

Some authors have proposed to use immunopathologic markers, such as islet cell antibodies, to identify presymptomatic type I DM. These studies, which have been restricted to relatives of patients with type I DM, have shown asymptomatic periods ranging from months to years prior to the onset of clinical disease.<8> Many issues remain to be resolved before recommending screening for type I DM in the general population. The immunoassays for islet cell antibodies remain difficult to standardize.<8> Screening for these markers has largely been restricted to first degree relatives of patients with type I DM. Finally, firm evidence that treatment of individuals with immune markers prevents progression of disease or reduces complications is lacking.

Maneuver

There are two gold standard diagnostic criteria for DM depending on the presence or the absence of symptoms. In patients who have symptoms and signs of diabetes (polyuria, polydipsia, polyphagia, weight loss, fatigue, blurred vision etc.), the diagnosis of diabetes is made on the basis of either a random venous plasma glucose above 11.1 mmol/L (200 mg/dl) or at least two fasting venous plasma glucose levels over 7.8 mmol/L (140 mg/dl). The oral glucose tolerance test (OGTT) is unnecessary if the patient meets these criteria. In patients without symptoms, biochemical hyperglycemia must be confirmed by the 75 gr 2 hour OGTT.<7> The most commonly used diagnostic criterion for the OGTT is that of the National Diabetes Data Group (NDDG).<9> The test is considered abnormal when the 2-hour venous plasma glucose level is >11.1mmol/L (200 mg/dl) and one glucose value before two hours is >11.1 mmol/L (200 mg/dl).

The National Diabetes Data Group recognizes an intermediate form of disordered glucose metabolism, i.e. impaired glucose tolerance (IGT). The diagnosis of IGT requires OGTT test results in which the 2-hour glucose value is between 7.8 mmol/L (140 mg/dl) and 11/1 mmpl/L (200 mg/dl), and one intervening value is greater than ll.1 mmol/L (200 mg/dl). Thus, IGT is a point on a continuum between normal glucose tolerance and a diabetic state.

The OGTT must be positive on more than one occasion to establish a diagnosis of diabetes mellitus. Indeed, the OGTT is known for its test-retest variability. In an attempt to decrease some of the test-retest variability, the American Diabetes Association (ADA) recommends that the OGTT only be performed in patients who have had an unrestricted diet for three days preceding the test and that the test should be administered after an overnight fast.<10>

Several screening tests for DM have been proposed (fasting, random, or 2-hour post-prandial glucose). ADA states that the best screening test is the fasting plasma glucose, in which sampling occurs after the patient has refrained from any food or beverage for at least a 3 hour period.<7,10>.

The fasting plasma glucose has a wide range of reported diagnostic sensitivity varying between 21% and 73%.<11-13> The wide variability in test performance reflects the unimodal distribution of glucose levels in most populations. Test sensitivity and specificity will vary with the cutoff used to define a positive result on the index test and the gold standard test. Study populations that differ in disease prevalence and severity may also contribute.

A number of studies have evaluated measurement of glycosylated proteins, primarily HA1c and serum fructosamine, as screening tools for diabetes. Test characteristics compare with the fasting plasma glucose, with sensitivities ranging from 15-91% and specificities reported as 84-89%.<14>

Effectiveness of Screening and Treatment

Screening for diabetes mellitus in asymptomatic individuals can only be justified if treatment during the asymptomatic phase prevents complications more effectively than treatment instituted after symptoms have developed. Studies over the preceding 20 years have suggested an association between the degree of glycemic control, duration of disease and complication rates. However, there is no conclusive evidence supporting the increased effectiveness of early intervention.

Microvascular complications (e.g. diabetic nephropathy and retinopathy) are correlated with level of glucose control in longitudinal studies. Studies examining this question have utilized almost exclusively type I IDDM patients. The relevance of these findings to type II NIDDM patients is unknown. The Diabetes Control and Complications Trial (DCCT), involving over 1,400 subjects, is the largest prospective randomized clinical trial to date and was designed to determine whether tight glycemic control in type I diabetic patients retarded the progression of preexisting diabetic complications (secondary prevention). The trial also evaluated whether tight glycemic control prevented long term sequelae in diabetic subjects without preexisting complications (primary prevention). Intensive therapy reduced progression of microvascular disease in both groups.<15> People randomized to intensive therapy had a three-fold increase in severe hypoglycemic events. The study findings provide supportive, but not conclusive, evidence that hyperglycemia mediates microvascular disease in diabetic subjects. Directly translating the DCCT results to the management of type II diabetic patients remains controversial at this time. While improved glycemic control may result in decreasing microvascular complications in the type II diabetic population, studies providing firm evidence for this have not been published yet. The U.K. Prospective Diabetes Study is designed to address this question and data is anticipated in the near future.

Patients with diabetes are also at significantly increased risk of developing atherosclerotic and peripheral vascular disease. There is evidence that disease duration and degree of glycemic control do not affect macrovascular complications in individuals treated for type II diabetes. The rate of increase in risk of CHD events in type II diabetic patients increases with disease duration but at the same rate as people without diabetes. A recent review of the World Health Organization (WHO) multinational study of vascular disease in diabetes found no increased risk of cardiovascular events in type II diabetic subjects after 8 years of follow-up.<16> There was no significant correlation between cerebrovascular and peripheral vascular events and diabetes duration. Difficulty in accurately ascertaining the time of disease onset has complicated the study of possible associations between macrovascular complications of type II diabetes mellitus and disease duration. However, the estimation of disease duration is quite unreliable, since some data suggest that in the United States, as many as 50% of diabetics ignore their condition. Some studies have shown that up to 21% of newly diagnosed diabetics have signs of retinopathy.<17>

The majority of individuals with disordered glucose metabolism are classified as having impaired glucose tolerance (IGT). There is little direct evidence that asymptomatic persons benefit from the detection and treatment of IGT<16,18> Most untreated asymptomatic persons with IGT do not develop diabetes. The rate of progression from IGT to diabetes ranges from 1% to 6% per year, with a higher rate found in some minority populations.<16> Thus, although IGT is an important risk factor for diabetes, it is not by itself an established indication for treatment.

Recommendation of Others

In 1989, the U.S. Preventive Services Task Force recommended against routine screening for DM in the general asymptomatic non-pregnant adult population but stated that screening may be appropriate in selected high-risk groups.<19>

The American Diabetes Association (ADA) recommends screening for type II DM in individuals with one or more risk factors for this condition.<10> The screening test recommended is the fasting blood glucose assay. Routine screening of patients without risk factors is not recommended by the ADA.

Conclusions and Recommendations

Screening for DM in asymptomatic non-pregnant adults suffers from two important limitations: the lack of a screening test that combines accuracy with practicality, and the absence of adequate evidence that early detection and treatment improve outcome in asymptomatic persons. Possible, but as yet unproven benefits of early treatment of asymptomatic persons must be weighed against the potential adverse effects of screening (e.g. false positives, labelling) and treatment (e.g. dietary restrictions, medications, risk of hypoglycemic events). In the general population there is fair evidence not to screen for DM (D Recommendation).

In high-risk groups, in which the prevalence is higher, false positive results are likely to be fewer. It may be clinically prudent to consider selective case-finding in these groups, since unrecognized overt diabetes carries the same risks as diagnosed type II diabetes.

Periodic testing of individuals with risk factors for diabetes mellitus, (obesity, older age, family history of diabetes, belonging to a high risk ethnic group) could be a reasonable middle course between screening no one and universal screening. However, there is no evidence that early detection improves outcomes in high-risk groups.

In non-pregnant adults, primary prevention rather than screening may be an important means of preventing diabetes and its complications. Among the many benefits of exercise and weight reduction, for example, are improved glucose tolerance and reduced obesity, important risk factors for diabetes as well as for other serious chronic diseases. Since these healthy behaviours are widely recommended even in the absence of diabetes, patients should be encouraged to adopt these behaviours independent of diabetes screening (see Chapter 47).

Unanswered Questions (Research Agenda)

Evaluating the benefits of primary prevention and screening in high-risk populations is an important research question to resolve.

Evidence

The literature was identified by a MEDLINE search for the English language from 1989 to 1993 using as key words: diabetes non-insulin- dependent or diabetes-gestational and diagnosis, epidemiology and prevention-control. The review was initiated in November 1993 and finalized by the Task Force in January 1994.

 Full Citation

Link to Structured Abstract of this review

Link to Summary Table of this review

Link to Selected References list of this review

Top of Page

Home PageCTFPHC Home Page

Reprinted in modified format by the Canadian Task Force on Preventive Health Care
with permission.
For any technical issues please contact: webmaster@ctfphc.org
Original Copyright © 1994 Minister of Supply and Services Canada.
Last modified March 27, 1998.