NEW ORLEANS -- Using the hemoglobin A1c (HbA1c) blood test alone missed many cases of diabetes and overestimated the prevalence of normal glucose tolerance, according to researchers here.
In an analysis of data from the National Health and Nutrition Examination Survey (NHANES), the HbA1c test underdiagnosed diabetes by 73.07% versus plasma glucose criteria, significantly underestimating the true prevalence of the diagnosis, reported Maria Mercedes Chang Villacreses, MD, of the City of Hope National Medical Center in Duarte, California, and colleagues.
The HbA1c test overdiagnosed glucose tolerance by 42.48%, they reported at ENDO 2019, The Endocrine Society annual meeting, adding that there would be a missed opportunity for early diabetic intervention in one case, and a missed opportunity for early intervention of diabetes prevention in the other.
"The [Hb]A1c shows a reflection of the average 3-month blood sugar and the oral [glucose tolerance test] will tell you how the patient is handling the load of sugar that you're giving him at the moment," Villacreses told 鶹ý. "There might be patients who don't have quite the elevated A1c, but it doesn't mean they're not handling the sugar appropriately."
Using an oral glucose tolerance test (OGTT) to confirm the findings of an A1c test could give way to earlier interventions in patients who aren't responding appropriately to glucose, preventing the progression of diabetes, Villacreses added.
The A1c test measures the percentage of glycated hemoglobin, with higher A1c percentages , while the OGTT is a 2-hour test that monitors blood glucose levels before and after drinking a sweetened beverage.
The state that OGTT and the HbA1c tests are "equally appropriate" for diagnosing prediabetes and type 2 diabetes. However, in certain conditions like sickle cell disease, anemia, or HIV, "only plasma blood glucose criteria should be used to diagnose diabetes," according to the guidelines.
Meanwhile, the recommend A1c tests only be used for screening prediabetes, but add that this diagnosis should still be confirmed with glucose testing. When screening for diabetes, they recommend using a glucose tolerance test as the primary diagnostic.
Susan Spratt, MD, of Duke University Medical Center in Durham, North Carolina, said HbA1c tests are often convenient because they use a single blood stick and do not require patients to fast like the OGTT does. She added that it would be important to determine more detailed characteristics of the patients in this study who were missed by HbA1c screening in order to better understand the discordance.
"Many of the patients we see who we screen for diabetes are overweight, obese, or have other cardiovascular risk factors," Spratt, who was not involved in the study, told 鶹ý in an email. "Even if patients in this category screen negative for diabetes, they should still be counseled on weight loss [and a] healthy diet, and other atherosclerotic cardiovascular disease risk factors should be treated."
"If you have a patient who has had abnormal glucose levels but their A1c is normal, I would strongly consider performing an oral glucose tolerance test," Spratt added.
The study collected data from 9,000 individuals (minimum age 20) who did not have established diabetes in NHANES from 2005 to 2014. The participants were an average age of 48 years and about half were female, Villacreses said.
The researchers used the American Diabetes Association definitions of normal glucose tolerance, prediabetes, and diabetes, considering hbA1c levels for normal glucose tolerance to be <5.7%, prediabetes to be between 5.7% and 6.4%, and diabetes to be >6.4%. Responses to the OGTT were classified as:
- Normal glucose tolerance: fasting plasma glucose <100 mg/dL and 2-hour post challenged plasma glucose <140 mg/dL
- Prediabetes: fasting plasma glucose 100-125 mg/dL or 2-hour post challenged plasma glucose 140-199 mg/dL but not diabetic
- Diabetes mellitus: fasting plasma glucose ≥126 mg/dL or 2-hour post challenged plasma glucose ≥200 mg/dL
The accuracy of the HbA1c test differed across various races and ethnicities, with the highest sensitivity for diagnosing diabetes in non-Hispanic black patients (41.25%), followed by Mexican Americans (30.71%), and non-Hispanic white patients (21.11%), Villacreses' group reported.
The HbA1c sensitivity for diagnosing normal glucose tolerance was lowest in non-Hispanic black patients (71.79%), followed by Mexican Americans (86.69%), and non-Hispanic white patients (89.40%), they noted.
Overall in diagnosing diabetes, the HbA1c sensitivity was 26.93% and the specificity was 99.39%, while for determining normal glucose tolerance, the A1c sensitivity was 84.91% and the specificity was 43.54%.
Villacreses said the primary limitation of the study was its retrospective nature, adding that future prospective studies would be necessary to confirm the findings.
Primary Source
ENDO 2019
Villacreses M, et al "Underestimation of the prevalence of diabetes and overestimation of the prevalence of glucose tolerance by using hemoglobin A1c criteria" ENDO 2019; Abstract SAT-125.