[Postgraduate Medicine]
[an error occurred while processing this directive]

Update on screening for type 2 diabetes

The why, who, how, and what of testing and diagnosing

Nasir M. Khan, MD; Charles O. Hershey, MD

VOL 109 / NO 2 / FEBRUARY 2001 / POSTGRADUATE MEDICINE

CME learning objectives

  • To review the need for screening for type 2 diabetes
  • To discuss the risk factors for and methods of screening for type 2 diabetes
  • To review the new diagnostic guidelines

The authors disclose no financial interests in this article.

This page is best viewed with a browser that supports tables.

Preview: Diabetes mellitus, a leading cause of death in the United States, remains undiagnosed in more than 5 million people. While screening for this disease is key to reducing its complications, morbidity and mortality, the recommendations for identifying and testing at-risk populations are debatable. This article summarizes the competing guidelines for screening for type 2 diabetes and offers insight on the different diagnostic methods and tools available.
Khan NM, Hershey CO. Update on screening for type 2 diabetes: the why, who, how, and what of testing and diagnosing. Postgrad Med 2001;109(2):27-34

The prevalence of diabetes mellitus in US adults aged 20 and older is estimated to be 7.8%, according to the Third National Health and Nutrition Examination Survey. Of the estimated 15.6 million Americans with diabetes, 5.4 million (35%) are unaware they have the disease (1). Undiagnosed diabetes is not a benign condition, however, and many patients have serious complications by the time the disease is diagnosed (2). In the face of these findings, the case for diabetes screening is compelling.

Generally, screening for a particular disease is likely to be beneficial if the disease is serious enough to warrant medical attention; effective treatment is available; and the prognosis for the disease is better if it is diagnosed and treated early. Further, the screening and diagnostic methods should have high accuracy and be available at low cost and with little inconvenience and minimal discomfort relative to the benefits of diagnosis (3).

The case of diabetes satisfies almost all these criteria. The disease is unquestionably serious, and effective therapy is available. Whether or not prognosis is improved by early diagnosis has been debated, but evidence now indicates that treatment may improve the outcome (4-6).

In this review, we will address screening for type 2 diabetes in nonpregnant adults. Currently, the American Diabetes Association (ADA) does not recommend screening for type 1 diabetes by testing for the presence of autoantibodies related to the disease. The reason for this decision is threefold. First, reference values for some of the markers have not been established. Second, there is no consensus as to what to do if autoantibodies are present. And third, the incidence of type 1 diabetes is low. Further, because of its acute onset, type 1 diabetes is usually detected soon after symptoms appear (7).

The following questions need to be addressed when a physician is deciding to screen a patient for type 2 diabetes:

  • Why screen for type 2 diabetes?
  • Who should be screened?
  • How should patients be screened?
  • What are the diagnostic criteria?

Why screen for type 2 diabetes?

Diabetes is the seventh leading cause of death in the United States (8) and is the leading cause of blindness, end-stage renal disease, and lower extremity amputations. The estimated annual cost of medical care for diabetic patients is $100 billion, which accounts for 15% of total healthcare costs in the United States and 25% of Medicare costs (9).

Among patients with undiagnosed type 2 diabetes, 10% to 29% have retinopathy, 10% to 37% have proteinuria, and 9% have neuropathy at the time of diagnosis. About 22% have abnormal heart findings, 19% have coronary artery disease, and 10% have peripheral vascular disease. In addition, about 61% have hypertension, 49% have hypercholesterolemia, and 40% have elevated levels of low-density lipoprotein (LDL) cholesterol (2). Neuropathy and abnormal heart findings are significantly more common in patients with undiagnosed diabetes than in nondiabetic persons, and the prevalence of cardiovascular events has been found to be as frequent in undiagnosed diabetes as in diagnosed diabetes (2). In fact, it is estimated that onset of type 2 diabetes precedes clinical diagnosis by 10 to 12 years (9). Thus, undiagnosed diabetes takes a toll, and a lack of symptoms should not be considered indicative of the absence of complications.

Now that we have evidence that the risk for microvascular complications (eg, retinopathy, nephropathy) can be reduced by intensive glucose control (4-6), earlier detection of diabetes--and hence the rationale to screen for it--becomes important. Although data demonstrating improved prognosis for macrovascular diseases with aggressive glucose control are lacking, diabetes detection may become a potent source of motivation (for both the patient and the physician) to pay serious attention to other modifiable risk factors for atherosclerosis (10). The United Kingdom Prospective Diabetes Study (11) showed that tight blood pressure control in type 2 diabetes resulted in a statistically significant risk reduction of any diabetes-related end point, death related to diabetes, fatal and nonfatal strokes, and heart failure. The number of patients needed to treat was 6.1 to prevent any complication and 15 to prevent a death related to diabetes.

Who should be screened?

Many authorities (including those mentioned subsequently) recommend screening only in patients who have risk factors for type 2 diabetes and not for the population in general. The main challenge, however, is in defining what constitutes a risk factor. A number of authoritative sources provide screening guidelines, but each identifies a slightly different group of risk factors. This article discusses the risk factors proposed by the ADA (7), the American College of Physicians, the Canadian Task Force on the Periodic Health Examination, and the US Preventive Services Task Force (12). The most exhaustive list of risk factors is from the ADA (table 1).

Table 1. Major risk factors for type 2 diabetes mellitus

Family history of diabetes (ie, parents or siblings with diabetes)

Obesity (>20% over desired body weight or body mass index >27 kg/m2)

Race or ethnicity with high risk of diabetes (eg, African American, Hispanic American, Native American, Asian American, Pacific Islander)

Age >45 years

Previously identified impaired fasting glucose or impaired glucose tolerance

Hypertension (>140/90 mmHg)

Hyperlipidemia (HDL cholesterol level <35 mg/dL [0.90 mmol/L] or triglyceride level >250 mg/dL [2.82 mmol/L], or both)

History of gestational diabetes or delivery of a baby over 9 lb (4.1 kg)

HDL, high-density lipoprotein.

Reproduced, with permission, from the American Diabetes Association (7).

The American College of Physicians recommends screening in patients who are 18 years of age or older and have one or more of the following characteristics: family history of type 2 diabetes, age over 50, weight 25% over ideal body weight, gestational diabetes, and ethnicity known to be at high risk of diabetes.

The Canadian Task Force agrees with screening in patients who are 18 years of age or older and have gestational diabetes or a family history of type 2 diabetes, and it presents peripheral vascular disease as an additional risk factor. The task force does not mention as risk factors obesity, greater age, and ethnicity at high risk of diabetes.

The US Preventive Services Task Force concurs with initiating screening at age 18 in the presence of risk factors that include family history, gestational diabetes, and high-risk ethnicity. They also agree with screening when obesity is present in patients over 40 years of age.

Of interest is that only the ADA lists hypertension, hyperlipidemia, fetal macrosomia, and impaired glucose tolerance as important risk factors. Macrosomia is a complication known to be associated with gestational diabetes in 15% to 47% of such pregnancies (13,14). Is macrosomia an independent risk factor for abnormal postpartum glucose tolerance? Coustan and colleagues (15) did not find an association when they conducted a logistic regression analysis of suspected risk factors for postpartum glucose intolerance. Also of interest is that in 1997 the ADA deleted total cholesterol value from its list of risk factors; the value was included in its position statement in 1996. The ADA does not present the rationale for excluding total cholesterol in its position statement (7).

How should patients be screened?

For a physician to establish a diagnosis of type 2 diabetes, the patient must undergo a screening test. If the result is abnormal, a confirmatory test should be performed to establish the diagnosis conclusively.

Several methods can be used to screen for type 2 diabetes.

Questionnaire
Various questionnaires have been developed to aid in screening for type 2 diabetes (7,16). Their differences demonstrate the lack of agreement about the importance of individual risk factors, and some questionnaires may not be applicable to all patient populations. For example, reluctance to use a bicycle for transportation, presented in a questionnaire from the Netherlands, may not be considered a risk factor in a society that does not use the bicycle as a common mode of travel (16).

The ADA recommends the use of its questionnaire for community screening programs to identify high-risk persons. The questionnaire uses a scoring system and assigns different numeric values to the presence of different risk factors; a sum of 10 points or more indicates a high risk for diabetes (7). A MEDLINE search using the key words "diabetes mellitus," "mass screening," and "questionnaire" did not reveal evidence that the current ADA questionnaire has been validated.

Others have developed similar questionnaires with sensitivities of 59% to 72%, specificities of 55% to 57%, positive predictive values of 5.6% to 6.5%, and negative predictive values of about 98% (16).

Questionnaires are helpful as noninvasive tools to rule out diabetes with a good certainty. However, they have poor positive predictive values.

Fasting plasma glucose
The ADA recommends fasting plasma glucose (FPG) as the screening test of choice. For the purpose of screening, fasting is defined as no calorie intake for 8 hours before the test. The ADA considers an FPG level of 126 mg/dL (7 mmol/L) or more to be a positive result (7). While this methodology standardizes the testing procedure, logistically the FPG usually requires that the patient make an additional visit to the office or laboratory (unless the patient has fasted for 8 hours before the initial appointment).

Random plasma glucose
Random plasma glucose (RPG) refers to measuring plasma glucose without regard to the last food intake. The ADA considers an RPG value of 160 mg/dL (8.9 mmol/L) or above to be abnormal (7). The RPG is less standardized and has a lower sensitivity and specificity than FPG. However, it is clearly easier and more convenient to obtain because blood can be drawn during the same office visit.

Oral glucose tolerance test
The oral glucose tolerance test (OGTT) requires ingestion of a glucose load of 75 g anhydrous glucose dissolved in water. A plasma glucose value of 200 mg/dL (11.1 mmol/L) or more 2 hours after ingestion of the glucose load is abnormal. The ADA strongly recommends FPG measurement over OGTT because FPG is easier and faster to perform, more convenient and acceptable to patients, and less expensive (7). However, this opinion is not universally shared; some authorities still recommend the use of OGTT (17).

Glycosylated hemoglobin
The ADA considers glycosylated hemoglobin an important monitoring tool but does not recommend it for screening at this time. Peters and associates (18) suggested use of a glycosylated hemoglobin A1c (HbA1c) value of 7.0% as the cutoff for detecting "treatment-requiring diabetes." They argued that the clinical approach to a patient with an HbA1c below 7.0% would be the same regardless of OGTT results and that pharmacologic agents are rarely prescribed for such patients (18).

Urinalysis
Urinalysis should not be used as a screening tool for diabetes, mainly because of physiologic reasons. For most patients, the renal threshold for excretion of glucose is 180 mg/dL. Hence, glucose may remain undetectable in urine until levels exceed 180 mg/dL in plasma (19).

Finger-stick glucose
The ADA does not recommend that physicians use finger-stick glucose (FSG) for screening. In contrast, the association allows capillary blood testing by glucometer for screening in the community. However, the ADA recommends that a positive screen by FSG should be confirmed on two more occasions using venous plasma samples. Many glucometers measure glucose values in whole blood and not in plasma. Plasma glucose values are 10% to 15% higher than the corresponding whole blood glucose values. The cutoff for a positive screen of fasting whole blood glucose is a level of 110 mg/dL (6.1 mmol/L) or above; for a random whole blood glucose value, the cutoff is 140 mg/dL (7.8 mmol/L) or above (7).

The ADA advises screening of all persons aged 45 years and older. If results are normal, it recommends screening at 3-year intervals thereafter. Screening is recommended at a younger age and more frequently if additional risk factors are present (7). The actual age at and frequency of subsequent testing are left to the physician's discretion.

What are the diagnostic criteria?

The ADA recommends that almost all patients with a positive screening test result undergo a confirmatory test using FPG or OGTT. The RPG may be used in the presence of symptoms related to diabetes. Exceptions to confirmation include unequivocal hyperglycemia with acute metabolic decompensation (eg, a plasma glucose level of 700 mg/dL with ketoacidosis). The ADA proposed new diagnostic criteria for fasting glucose in 1997 (table 2). Its objective was to reduce the discrepancy between positive OGTT results when compared with those of the FPG and to facilitate the use of a simpler, equally accurate method for diagnosing diabetes (20). Controversy exists as to which diagnostic criterion--FPG or OGTT--is a better predictor of diabetic complications (17). The topic remains debatable.

Table 2. ADA 1997 diagnostic criteria for type 2 diabetes

Symptoms of diabetes and RPG > 200 mg/dL (11.1 mmol/L)

     or

FPG > 126 mg/dL (7.0 mmol/L)

     or

2-hour plasma glucose > 200 mg/dL (11.1 mmol/L) during an OGTT

FPG, fasting plasma glucose; OGTT, oral glucose tolerance test; RPG, random plasma glucose.

Adapted, with permission, from the American Diabetes Association (7).

Discussion

It is obvious that the recommendations of the leading diabetes authorities have some controversies, ambiguities, and uncertainties. Their guidelines agree on some components but disagree on others.

For instance, recommendations do agree on screening only those patients who have risk factors for diabetes. However, there is no consensus as to which risk factors are most important. If we follow the guidelines, all those who belong to a minority race or ethnicity should be screened at age 18 or older. In addition, all obese patients and persons who have a family history of diabetes or hypertension; have a personal history of gestational diabetes, macrovascular disease or hyperlipidemia; or are older than age 40, 45, or 50 (depending on the guideline preferred) should be screened. Many women who are pregnant will be screened as well. Thus, it would appear that a major portion of the US population would have one or more indications for screening.

The choice of screening test also merits discussion. The ADA recommends use of FPG as the test of choice for screening, and from a practical standpoint, the choice lies between FPG and RPG. The FPG, which requires a single blood draw, provides a standardized test with good accuracy. However, it usually requires that a person make a second visit to the office or laboratory for the blood draw. For some patients, it may not be acceptable to schedule this extra appointment. The RPG has less predictive value but requires a single blood draw that may be obtained during the same visit or at any time of day without regard to last food intake. Thus, for some patients, it may prove to be easier and more convenient than the FPG.

The OGTT may be used for screening, but it is not commonly used. Points to consider are test accuracy and the willingness of the patient to appear for the procedure. A question to answer is, How often will a patient with multiple risk factors appear for the 2-hour OGTT?

The predominant glycosylated hemoglobin, HbA1c, has been looked upon with interest for its use in diabetes screening. The formation of HbA1c is the result of slow, nonenzymatic binding of carbohydrates (mainly glucose) with hemoglobin. Its presence reflects the level of blood glucose concentration and the duration of its exposure to hemoglobin. Glycosylation occurs over the life span of an erythrocyte (mean, about 120 days). However, during this process, the greatest influence on HbA1c formation is recent hyperglycemia. It has been suggested that with stable glucose control, a patient forms 50% of the HbA1c in the month before the screening, 25% between 1 and 2 months before, and the remaining 25% between 2 and 4 months before screening (21). Hence, the rate of glycosylation is not linear. Further, it was found that among nondiabetic patients with the same level of glucose control, there was up to 2% variation in HbA1c values (21). The variation may be due to differences in rates of glycosylation or erythrocyte life span (21,22).

Factors other than glycemia are known to affect the HbA1c level as well. Hemolytic anemias (leading to shortened erythrocyte life span), cirrhosis, and high intake of vitamins C and E have been shown to lower the value of HbA1c. Hemoglobin variants (eg, hemoglobin F) can affect certain assays and produce falsely high results. There have been serious concerns about the standardization of assays to detect HbA1c, and reproducibility has been poor in healthy adults, which is in contrast to the stable results seen in patients with diabetes (22). Moreover, a meta-analysis found that many patients had a normal HbA1c value when they had either impaired glucose tolerance or frank diabetes diagnosed on the basis of World Health Organization criteria using OGTT (18). Because of these limitations, HbA1c is not currently recommended as a screening method in the United States, and recent European recommendations find a role for it only in the presence of simultaneous blood glucose testing (21).

Frequency of rescreening if the first screen result is normal depends upon the number of risk factors present. In a person with multiple risk factors and a normal initial screen, there are no data to provide us with a definite guideline. The authorities leave the physician to determine the frequency of rescreening but advise that it be based on the number of risk factors present.

Summary

Type 2 diabetes can cause serious complications even as it remains undiagnosed. Screening is recommended only in people with risk factors for the disease. The ADA recommends FPG as the test of choice, but RPG is also a practical alternative because it is easier and more convenient. Frequency of rescreening if the first screen result is normal depends on the number of risk factors present. Physicians should realize the importance of a confirmatory test and not base a diagnosis of diabetes on a single value unless the value is so high that the diagnosis is unequivocal.

References

  1. Harris MI, Flegal KM, Cowie CC, et al. Prevalence of diabetes, impaired fasting glucose, and impaired glucose tolerance in U.S. adults. The Third National Health and Nutrition Examination Survey, 1988-1994. Diabetes Care 1998;21(4):518-24
  2. Harris MI. Undiagnosed NIDDM: clinical and public health issues. Diabetes Care 1993;16(4):642-52
  3. Knowler WC. Screening for NIDDM: opportunities for detection, treatment, and prevention. Diabetes Care 1994;17(5):445-50
  4. Implications of the United Kingdom prospective diabetes study. American Diabetes Association. Diabetes Care 2000;23(Suppl 1):S27-31
  5. Ohkubo Y, Kishikawa H, Araki E, et al. Intensive insulin therapy prevents the progression of diabetic microvascular complications in Japanese patients with non-insulin-dependent diabetes mellitus: a randomized prospective 6-year study. Diabetes Res Clin Pract 1995;28(2):103-17
  6. Andersson DK, Svärdsudd K. Long-term glycemic control relates to mortality in type II diabetes. Diabetes Care 1995; 18(12):1534-43
  7. Screening for type 2 diabetes. American Diabetes Association. Diabetes Care 2000;23(Suppl 1):S20-3
  8. Centers for Disease Control and Prevention. Mortality patterns--United States, 1991. MMWR Morb Mortal Wkly Rep 1993;42(46):891-900
  9. Harris MI, Eastman RC. Early detection of undiagnosed non-insulin-dependent diabetes mellitus. (Editorial) JAMA 1996;276(15):1261-2
  10. Goyder E, Irwig L. Screening for diabetes. What are we really doing? BMJ 1998;317(7173):1644-6
  11. Tight blood pressure control and risk of macrovascular and microvascular complications in type 2 diabetes: UKPDS 38. UK Prospective Diabetes Study Group. BMJ 1998;317(7160):703-13
  12. Hayward RS, Steinberg EP, Ford DE, et al. Preventive care guidelines: 1991. American College of Physicians. Canadian Task Force on the Periodic Health Examination. United States Preventive Services Task Force. Ann Intern Med 1991;114(9):758-83
  13. Ryan EA. Pregnancy in diabetes. Med Clin North Am 1998;82(4):823-45
  14. Moore TR. Fetal growth in diabetic pregnancy. Clin Obstet Gynecol 1997;40(4):771-86
  15. Coustan DR, Carpenter MW, O'Sullivan PS, et al. Gestational diabetes: predictors of subsequent disordered glucose metabolism. Am J Obstet Gynecol 1993;168(4):1139-45
  16. Ruige JB, de Neeling JN, Kostense PJ, et al. Performance of an NIDDM screening questionnaire based on symptoms and risk factors. Diabetes Care 1997;20(4):491-6
  17. Davies M. New diagnostic criteria for diabetes. Are they doing what they should? (Commentary) Lancet 1999;354(9179):610-1
  18. Peters AL, Davidson MB, Schriger DL, et al. A clinical approach for the diagnosis of diabetes mellitus: an analysis using glycosated hemoglobin levels. Meta-analysis Research Group on the Diagnosis of Diabetes Using Glycated Hemoglobin Levels. JAMA 1996;276(15):1246-52
  19. Tests of glycemia in diabetes. American Diabetes Association. Diabetes Care 2000;23(Suppl 1):S80-2
  20. Report of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Diabetes Care 1997;20(7):1183-97
  21. Kilpatrick ES. Glycated haemoglobin in the year 2000. J Clin Pathol 2000;53(5):335-9
  22. Simon D, Senan C, Balkau B, et al. Reproducibility of HbA1c in a healthy adult population: the Telecom Study. Diabetes Care 1999;22(8):1361-3

Dr Khan is assistant professor of clinical medicine, department of medicine, State University of New York at Buffalo School of Medicine and Biomedical Sciences, and staff physician, division of general internal medicine, department of medicine, Erie County Medical Center, Buffalo. Dr Hershey is professor of medicine, State University of New York at Buffalo School of Medicine and Biomedical Sciences, and section head, division of general internal medicine, department of medicine, Erie County Medical Center. Correspondence: Nasir M. Khan, MD, Division of General Internal Medicine, Erie County Medical Center, 462 Grider St, Buffalo, NY 14215. E-mail: nkhan@acsu.buffalo.edu.

RETURN TO FEBRUARY 2001 TABLE OF CONTENTS

 

[an error occurred while processing this directive]

 

about us  |  cme  |   home  |  issue index  |   patient notes  |  pearls  |   ad services  |  

Please send technical questions related to the Web site to Ann Harste