Screening for Type II Diabetes Mellitus in the United States: The Present and the Future

Introduction

Type II diabetes mellitus can be diagnosed with an abnormal glycohemoglobin test (hemoglobin A1c%), oral glucose tolerance test, or fasting plasma glucose test. ¹ The 2014 National Diabetes Statistics report estimated that 37% of American adults, aged 20 years and older, were in the prediabetic range from 2009 to 2012 and 9.3% of the United States population had diabetes. ² Additionally, half of those who do present with type 2 diabetes have already developed clinical complications at the time of diagnosis. ¹ Diabetes causes more blindness, renal disease, and amputations than any other preventable disease. The estimated burden on the health-care system is around 245 billion dollars.^2,3 Therefore, it is imperative to diagnosis the disease when an individual transitions from being disease free to the asymptomatic state. This lead time can provide an opportunity to reduce long-term sequelae and implement lifestyle and pharmacological interventions to reduce health-care expenditures ultimately.

Screening: Current Recommended Practices

Screening is the process by which asymptomatic individuals who are at high risk of the disease are identified for further investigation. ⁴ An fasting blood glucose (FBG) ≤ 126 mg/dL is diagnostic for diabetes and warrants retesting. A 75-g oral glucose tolerance test is also suitable and screening is positive with a 2-hour postload value of ≤200 mg/dL. Values ≤200 mg/dL are repeated on a different day to confirm diagnosis of diabetes. The A1c test is also a valuable tool for diagnosis of diabetes and A1c > 5.6 indicates impaired glucose tolerance. ⁵ Genetic screening for diabetes is of little value in clinical practice. ⁶

In United States, different societies and task forces have recommended varying guidelines for screening for Type 2 diabetes mellitus as presented in Table 1. Despite these guidelines for earlier screening, there are individuals who are not clinically diagnosed until at least a decade after subclinical disease. ⁷ This is likely due to a combination of ineffective screening guidelines, inadequate implementation of the guidelines, and late presentation of disease. These guidelines, like all other screening tools in the medical community, are employed by many practitioners to target high-risk populations (Table 1). The United States Preventative Services Task Force (USPSTF) revised in its 2008 guidelines for screening in asymptomatic adults with sustained blood pressure > 135/80 or obese adults aged between 40 and 70 years.^8,9 The American Diabetes Association (ADA) recommends screening based on body mass index (BMI) ≤ 25 kg/m² in addition to risk factors. Individuals aged ≤45 years are recommended to be screened regardless of risk factors. Likewise, in children aged 10 years and older, ADA advocates for screening if obesity is present in addition to two risk factors. Criterion that is established as a risk factor includes family history, race/ethnicity, insulin resistance comorbidities, and maternal history of gestational diabetes during the child's gestation. ¹⁰ The American Association of Clinical Endocrinologists (AACE) expands their criteria to include individuals with singular risk factors for diabetes as screening criteria. ¹¹ The International Diabetes Federation (IDF) recommends against universal screening for diabetes. ¹²

OPEN IN VIEWER

Table 1.

Guidelines for screening for type II diabetes.

ORGANIZATION	SCREENING RECOMMENDATION
United States Preventative Services Task Force (USPSTF)	• Adults aged 40–70 years of age who are overweight or obese
American Diabetes Association (ADA)	• Adults aged ≤45 should be tested regardless of risk factors
Normal testing should be repeated every 3 years	• Adults of any age who are overweight or obese and who have 1 or more risk factors as below: • Physical inactivity • Hypertension • HDL < 35 mg/dL or TG > 250 mg/dL • First degree relative with diabetes • High risk race/ethnicity • Asian Americans with a BMI of ≤23 kg/m2 • Women who have delivered child >9 lbs or with gestational diabetes • A1c ≤ 5.7%, impaired fasting glucose (IFG) or impaired glucose tolerance (IGT) • Other: acanthosis nigricans, PCOS, history of coronary vascular disease (CVD)
American association of Clinical Endocrinologists (AACE)	• Age ≤ 45 without risk factors • CVD or family history of type 2 diabetes
Normal testing should be repeated every 3 years	• Overweight or obese
Consider annual screening for individuals with 2 or more risk factors	• Sedentary lifestyle • At risk racial/ethnic group: Asian, African American, Hispanic, Native American, Pacific Islander • HDL < 35 mg/dL or TG > 250 mg/dL • IGT or IFBG and/or metabolic syndrome • PCOS, acanthosis nigricans, or non-alcoholic fatty liver disease • Hypertension • History of gestational diabetes or baby > 9 lbs • Antipsychotic therapy of schizophrenia or bipolar, chronic glucocorticoid use, sleep disorders in presence of glucose intolerance (including obstructive sleep apnea, chronic sleep deprivation, and night-shift occupation)
international Diabetes Federation	• Universal screening for undiagnosed diabetes is not recommended • Screening should be based on prevalence of diabetes and available resources to detect and treat

Screening: Gaps in Current Practices

There is a growing concern that the current screening guidelines might be inadequate. Recently, Bullard et al suggested variabilities and differing recommendations by ADA and USPSTF 2008 guidelines. ¹³ Similarly, Sheehy et al found that the new USPSTF guidelines were inferior to ADA guidelines for screening and identifying diabetics. ¹⁴ Ochoa et al found that ADA risk factors for diabetic screening in the inpatient setting did not identify a significant number of diabetic patients. ¹⁵ Another study, conducted in the United Kingdom by Simmons et al examined individuals aged 40–69 years using a similar risk score as ADA and concluded similar findings. ¹⁶ Therefore, is it reasonable to assume that the current guidelines for the diagnosis of diabetes may not be ideal in the detection of diabetes? Better yet, does this call for revision of these guidelines to better characterize the population of diabetics to include a younger population? One way that this problem can be approached is to evaluate the risk groups included in these guidelines. The availability of evidence is skewed for different risk groups as not all populations have been studied with equal rigor.

Role of family history. It is estimated that an individual with one parent diagnosed with diabetes before the age of 50 years has a 1 in 7 chances of developing the disease. This number increases to 1 in 2 for those with both parents carrying the disease. ¹⁷ There is good evidence that universal screening for individuals with a positive family history should begin at a much earlier age prior to the onset of diabetes in their immediate family members. ¹⁸ However, this is rarely done. A report of the Expert Committee recommends earlier age of screening for individuals with first-degree relatives with diabetes, but the exact age is not mentioned. ¹⁹ A 1999 cross-sectional study of 972 diabetics of three ethnicities at the University of Singapore found an earlier age of diagnosis (mean of seven years) in individuals with positive family history when compared with patients without familial history (CI 95%, <0.01). It concluded that family history is an independent factor in predicting age at diagnosis in type II diabetics. ¹⁸ Therefore, in a given population, the mean sojourn time in order to gain lead time bias is roughly 8–12 years.^7,20 There are no guidelines that specify when earlier screening should be carried out in individuals with a family history of diabetes.

Role of ethnicity. The ADA notes the highest rate of diagnosed diabetes by ethnicity in minority populations, especially in the Native Americans, where the prevalence is 33%. ²¹ This is followed by non-Hispanic Blacks (13.2%) and Hispanics (12.8%). ¹ Recent research found that BMI-adjusted prevalence of diabetes in Asian Americans (Far East, South East, and Indian descent) was ∼60% higher than in non-Hispanic Whites. Therefore, Asian Americans are at a higher risk of diabetes at a lower BMI. This led to ADA adjusting its diabetic screening guidelines to include testing Asian Americans at a BMI of 23 kg/m² or greater. ²² While the ADA and AACE includes ethnicity, the USPSTF guidelines have neglected to include this high-risk population.

Role of genetics. As noted above, the aggregation of diabetes in familial and ethnic group patterns has a lot to do with genetic inheritance. However, the polygenic inheritance and lack of clear evidence limits the use of genetic testing in clinical practice. ⁶

Role of smoking. Smokers have increased insulin resistance. Smoking is believed to be linked to increased glycosylation of hemoglobin. One study showed an average decrease of hemoglobin A1c of 0.7% after smoking cessation. ²³ No current guidelines include smokers in the screening criteria.

Role of age. Approximately 0.35% of the diabetic population is of age 20 years or younger. ²⁴ In a recent press release, the ADA recognizes that in adults, 20–44 years of age, the rate of undiagnosed diabetes is 60% higher than the general adult population, yet it fails to include these individuals in its screening parameters. ²⁵ It is also important to note that the guidelines suggest screening in obese individuals who have other risk factors, but there are no recommendations in individuals with the sole risk factor of obesity. Likewise, if an individual has risk factors but is not obese, they are also not high-risk candidates to be screened routinely unless they are 10–18 years of age or ≤40 years of age. This provides gateway for individuals to be potentially missed by routine diabetic screening. As such, should we be initiating screening using a risk calculator to provide an estimate of early screening using familial history, race, ethnicity, and age among other risk factors? What would the implications of this, if any, be? This calls for further probability modeling and cost analysis.

Need for national benchmarks. A better system is needed to diagnose the 8.1 million asymptomatic diabetic Americans and the 76 million with prediabetes. ² USPSTF outlines screening that is focused on obesity and does not take into account the large array of risks, other than cardiovascular that predispose an individual to diabetes. It also does not take into account the population younger than 40 years of age or older than 70 years who continue to be diagnosed with diabetes. Similarly, the ADA outlines a strict parameter for screening, excluding the millions who are still at risk. There are no good studies which analyze the usage of the ADA and the USPSTF guidelines in the diagnosis of specific subpopulations. Data on physician use of these guidelines is also scarce. With growing emphasis on patient-centered medical home and pay for performance model of repayment, the need for national benchmarks of guideline usage and their pragmatic efficacy is paramount.

Where do we go from Here?

In summary, the studies to date suggest that evidence on the reliability of diabetic screening guidelines is poor. The guidelines fail to provide a broad umbrella for screening in those who are at highest risk. There is much debate regarding universal screening versus individualized screening. Earlier screening that is universal would increase the likelihood of diabetes to be detected, yet increasing anxiety and health-care expenditure.

On the other hand, the continued use of guidelines for recommending screening is coupled with lack of sensitivity and specificity. These recommendations do not provide coverage for the cost of screening in some individuals who are at greater risk.

The authors here recommend further equitable study of different high-risk populations as noted above. Meanwhile, prediabetes should also come to focus when screening for diabetes. Other than the ADA that recognizes prediabetic screening, the USPSTF, IDF, or the AACE advises screening for prediabetes. Physicians should consider the following:

Further high-quality studies that focus on earlier screening in otherwise asymptomatic patients in specific subpopulations are needed. In developing protocols for screening, diabetes can be detected at an earlier stage to decrease cardiovascular mortality and inevitably control this epidemic.

Author Contributions

Conceived and designed the experiments: AA. Analyzed the data: AA. Wrote the first draft of the manuscript: AA. Contributed to the writing of the manuscript: AA, SA, AW. Agree with manuscript results and conclusions: AA, SA, AW. Jointly developed the structure and arguments for the paper: AA, SA, AW. Made critical revisions and approved final version: AA, AW. All authors reviewed and approved of the final manuscript.