Hemoglobin A1c and Diagnosis of Diabetes in Adolescents
The diagnosis of diabetes has major ramifications for the family and child/adolescent and therefore must not be made without certainty. There are four ways to diagnose diabetes (all types including: type 1, type 2, type 1.5, gestational, permanent neonatal diabetes, MODY, steroid or medication induced diabetes, Cystic Fibrosis related diabetes, etc.). They are as follows:
- Fasting blood sugar on two occasions greater than 126 mg/dl.
- Random blood sugar greater than 200 mg/dl along with symptoms related to new onset diabetes including increased drinking, increased urination, increased appetite, and weight loss.
- Two hour oral glucose tolerance test (75 grams of carbohydrates): blood sugar greater than or equal to 200 mg/dl.
- Hb A1c greater than or equal to 6.5 percent (reflecting a 3-month estimated blood sugar average of 139.85 mg/dl.
Pre-diabetes (or impaired glucose tolerance) is diagnosed by the following criteria:
- Fasting blood sugar greater than 100 mg/dl and less than 126 mg/dl on two separate occasions.
- Two hour postprandial blood sugar (oral glucose tolerance test) greater than 140 mg/dl and less than 200 mg/dl.
- Hb A1c between 6.0 and 6.4 percent on two separate occasions.
There has been considerable discussion among the different endocrine organizations in regard to the hb A1c as a diagnostic tool; however, the American Diabetes Association has adapted it as the fourth way in which to diagnose diabetes. As I have mentioned previously, diagnosis of type 1 diabetes in children and adolescents is typically based on criteria number two above, but not always. In situations of uncertainty, especially in the possibility of type 2 or other forms of diabetes, I use several criteria to confirm the diagnosis.
A recent article in the Journal of Pediatrics posted online ahead of print,
"Diagnosis of Diabetes using Hemoglobin A1c: Should Recommendations in Adults Be Extrapolated to Adolescents?" (doi: 10.1016/j.jpeds.2010.11.026), is an epidemiologic study looking at the ability of the hb A1c to diagnose diabetes in the adolescent population. The authors Lee, Wu, Tarini, Herman and Yoon's objectives were to compare efficacy of hemoglobin A1c for detecting diabetes/pre-diabetes for adolescents versus adults in the United States. They choose individuals with criteria for the diagnosis of diabetes mentioned above as well as those with pre-diabetes. Fasting blood sugars were obtained in 1,156 obese and overweight adolescents from 12 to 18 years of age. The blood sugars were then compared to fasting blood sugars in 6,751 adults from 19 to 79. The authors also compared results of the two-hour oral glucose tolerance test in 267 adolescents and 1,476 adults.
The hb A1c test was evaluated statistically applying receiver operator characteristic (ROC) analyses. Sensitivity versus specificity in the diagnosis of diabetes also was evaluated. (Sensitivity refers to the proportion of people** with** the disease who have a** POSITIVE** test result.** Specificity** refers to the proportion of people** without** the disease who have a** NEGATIVE** test result).
What were the results?
- Few adolescents had undiagnosed diabetes.
- When assessing sensitivity and specificity analyses to evaluate fasting blood sugars, a hb A1c of 6.5 percent had sensitivity rates of 75.0 and 53.8 percent and specificity rates of 98.3 and 91.1 percent for adolescents and adults respectively.
- When assessing a fasting blood glucose to detect DM, a hb A1c of 5.7 percent (estimated average blood sugar of 116.89%) had sensitivity rates of 5.0 versus 23.1 percent and specificity rates of 98.3 and 91.1 percent for adolescents and adults, respectively.
- ROC analyses also suggested that hb A1c was a poorer predictor of diabetes for adolescents versus adults.
- Hb A1c ROC analysis also was less accurate for fasting blood sugar or two hour post-prandial readings in adolescents as compared to adults.
What do these statistical analyses conclude?
- The fasting blood sugar was significantly more accurate in detecting diabetes and pre-diabetes in adolescents than the hb A1c.
- The hb A1c was more sensitive in detecting diabetes in adults than adolescents.
- The two-hour glucose tolerance test also was more sensitive than the hb A1c in detecting diabetes in adolescents as compared to adults.
Keep in mind that the predominant populations used in the study were those with type 2 diabetes and that previous studies using hb A1c to diagnose diabetes were conducted in adults. Clearly, more studies will be required in the pediatric population to determine the efficacy of hb A1c in the diagnosis of diabetes. In conclusion, the diagnosis of diabetes should be made based on the history of symptoms, blood sugar measurements (fasting versus random blood sugar or oral glucose tolerance test) in addition to the hb A1c. In many cases, I have used a stat hb A1c to rule out new onset type 1 diabetes in patients with stress-induced hyperglycemia or steroid/medication induced hyperglycemia (specificity analysis). Finally, the hb A1c is only one tool in an armamentarium of choices: it is important to use it wisely and cautiously until more evidence-based studies become available.