HbA1c May Be Less Effective to Diagnose Adolescent Diabetes

Laurie Barclay, MD

March 04, 2011

March 4, 2011 — Recommendations for diagnosis of diabetes using hemoglobin A1c (HbA1c) level in adults may not be accurately extrapolated to adolescents, according to the results of a study using data from the National Health and Nutrition Examination Surveys (NHANES 1999-2006), reported online in the February 23 issue of the Journal of Pediatrics.

The American Diabetes Association (ADA) past president of medicine and science, Larry C. Deeb, MD, told Medscape Medical News that this was "an interesting study making a very important point."

"We were all afraid of a massive epidemic of type 2 diabetes in children at the end of the century, but it hasn't happened," said Dr. Deeb, who is also pediatric endocrinologist and medical director at the Diabetes Center at Tallahassee Memorial Hospital and a clinical professor at Florida State University. "This study points out that the actual numbers are much smaller, with few cases uncovered in the screening. This makes predictions using HbA1c really difficult."

The goal of this study was to compare the diagnostic ability of HbA1c to detect diabetes mellitus or prediabetes in 1156 obese and overweight US adolescents aged 12 to 18 years vs adults, using NHANES data from 1999 to 2006.

Criteria for diabetes mellitus were a fasting plasma glucose (FPG) level of at least126 mg/dL and a 2-hour plasma glucose level (2-hr PG) of at least 200 mg/dL. For prediabetes, diagnostic criteria were an FPG of at least 100 mg/dL and less than 126 mg/dL and 2-hr PG of at least 140 mg/dL and less than 200 mg/dL. Receiver operator characteristic (ROC) analyses allowed assessment of HbA1c test performance.

Only 4 adolescents had undiagnosed diabetes mellitus. For detection of diabetes (based on FPG) in adolescents and adults, respectively, the sensitivity rates of an HbA1c level of 6.5% were 75.0% (95% confidence interval [CI], 30.1% - 95.4%) and 53.8% (95% CI, 47.4% - 60.0%). The specificity rates were 99.9% (95% CI, 99.5% - 100.0%) and 99.5% (95% CI, 99.3% - 99.6%), respectively.

With an HbA1c level of 5.7%, the sensitivity rates were 5.0% (95% CI, 2.6% - 9.2%) and 23.1% (95% CI, 21.3% - 25.0%) for adolescents and adults, respectively, and the specificity rates were 98.3% (95% CI, 97.2% - 98.9%) and 91.1% (95% CI, 90.3% - 91.9%). For adolescents vs adults, HbA1c level was a poorer predictor of diabetes mellitus (area under the ROC curve, 0.88 vs 0.93) and prediabetes (FPG area under the ROC curve, 0.61 vs 0.74). With use of either FPG or 2-hr PG measurements, test performance of HbA1c level was poor in adolescents.

HbA1c Not as Reliable

"Based on the study findings, a fasting blood glucose test should still be used for diagnosing diabetes in children," said lead author Joyce M. Lee, MD, MPH, a pediatric endocrinologist at C.S. Mott Children's Hospital at the University of Michigan, in a news release. "...We found that ...HbA1c is not as reliable a test for identifying children with diabetes and pre-diabetes compared with adults. Using this test in children may lead to missed cases."

Dr. Deeb noted that a major strength of this study was that it was performed in a large, population-based sample, and that a limitation was the low number of individuals with diabetes mellitus in the study sample.

"I wonder if the numbers of cases really doesn't make it harder to use the laboratory and epidemiological data to try to predict sensitivity, specificity, and ROCs," Dr. Deeb said.

The study authors agree that their findings highlight the dilemma of screening for diabetes mellitus in adolescents, because the prevalence of undiagnosed diabetes mellitus in the pediatric population is only 0.02%, resulting in a low positive predictive value for detecting diabetes mellitus using any test, and not just HbA1c level. Other limitations of this study acknowledged by the authors include study sample of only asymptomatic overweight and obese children, and limited ability to diagnose undiagnosed diabetes mellitus and prediabetes because of the absence of subsequent testing of FPG or 2-hr PG.

"My major fear isn't the adolescent getting diabetes, but the young adults in their 20's and 30's," Dr. Deeb said. "While the diabetes epidemic in children isn't happening as we feared, the obesity epidemic is in full swing, and we are having issues with predicting diabetes, but not obesity. Will it turn into diabetes? If the history of the decades past is true....yes."

Current ADA guidelines note that asymptomatic individuals would be classified as having diabetes if HbA1c values are at least 6.5% and having prediabetes if HbA1c values are 6% to 6.4% on 2 separate tests.

"It's important that we not take a cavalier attitude toward someone with an HbA1c of 6.5 and tell them they don't have diabetes — we have to take the whole person into account, including their BMI and family history," Dr. Deeb concluded.

The study authors were supported by the National Institute of Diabetes and Digestive and Kidney Diseases; the Clinical Sciences Scholars Program at the University of Michigan; National Institute of Child Health & Human Development, the National Institutes of Health; and the National Heart, Lung, and Blood Institute. Dr. Deeb has disclosed no relevant financial relationships.

J Pediatrics. Published online February 23, 2011. Abstract


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: