HbA_1c Diagnostic Categories and β-Cell Function Relative to Insulin Sensitivity in Overweight/Obese Adolescents

Approval by the Institutional Review Board of the University of Pittsburgh, parental consent and child assent were obtained prior to any research procedure. A total of 204 overweight/obese youth (according to age- and sex-specific BMI percentiles [14]) (89 African Americans and 115 Caucasians; 88 males and 116 females; ages 9 to <20 years old; Tanner stage II–V) (15) who had complete hyperinsulinemic-euglycemic and hyperglycemic clamp data obtained while participating in our ongoing studies of National Institutes of Health grants “Childhood Metabolic Markers of Adult Morbidity in Blacks” and “Childhood Insulin Resistance” were included. Some of these participants’ data have been reported along with details of our recruitment and screening procedures (13,16–18). Tests were conducted at the Pediatric Clinical and Translational Research Center of the Children’s Hospital of Pittsburgh.

A 3-h hyperinsulinemic (80 μU/m²/min)-euglycemic (100 mg/dL) clamp was performed after a 10–12-h overnight fast (13,19). Fasting endogenous/hepatic glucose production (HGP) was measured in 164 participants using a primed (2.2 µmol/kg)–constant rate infusion of [6,6-²H₂] glucose (Isotech, Miamisburg, OH) at 0.22 µmol/kg/min for 2 h (−120 to 0 min) (20,21). Four baseline blood samples were collected (−30 to 0 min) for determination of glucose, insulin, and isotopic enrichment of glucose prior to the initiation (0 min) of the clamp. On a separate occasion, 1 to 3 weeks apart, and in random order, a 2-h hyperglycemic clamp (∼225 mg/dL) was performed (12,13,22). Either the day preceding one of the clamp procedures or on a separate visit within a 1- to 3-week period, a 2-h OGTT (1.75 g/kg glucola, maximum 75 g) was performed in 142 participants (23,24). Normal versus impaired glycemia was defined according to standards for fasting or 2-h OGTT glucose (2).

Percent body fat was measured with dual energy X-ray absorptiometry and abdominal visceral adipose tissue (VAT) with computed tomography in 154 subjects and magnetic resonance imaging in 50 subjects (17,21).

Plasma glucose was measured by the glucose oxidase method (Yellow Springs Instrument Co., Yellow Springs, OH), and insulin by a commercial radioimmunoassay (catalog number 1011; LINCO Research, St. Charles, MO) (13). HbA_1c was measured by high-performance liquid chromatography (Tosoh Medics), and “normal” was defined as HbA_1c <5.7% and “prediabetes” as HbA_1c ≥5.7% to <6.5% according to ADA criteria (2). Deuterium enrichment of glucose in the plasma was determined on a Hewlett-Packard 5971 mass spectrometer coupled to a 5890 series II gas chromatograph (Hewlett-Packard) (20,21).

Fasting HGP was measured during the last 30 min of the 2-h baseline isotope infusion and hepatic insulin sensitivity (HIS) was calculated as 1,000/(HGP × fasting insulin) (12,20). Peripheral insulin sensitivity (mg/kg/min per µU/mL) was calculated during the last 30 min (150–180 min) of the hyperinsulinemic-euglycemic clamp (12,13,19,22). First-phase insulin (µU/mL) was calculated as the mean insulin concentration at times 2.5, 5, 7.5, 10, and 12.5 min during the hyperglycemic clamp (21,25). β-Cell function relative to insulin sensitivity, the disposition index (DI; mg/kg/min), was calculated as the product of insulin sensitivity and first-phase insulin (13,16). In the subset of 142 participants with OGTT, area under the curve (AUC) for glucose and insulin was calculated by the trapezoidal rule and the insulinogenic index (ΔI₃₀/ΔG₃₀) and the oral DI (oDI) as before (24–26).

Differences in categorical variables were determined by χ² analysis, and differences in continuous variables were determined by two-tailed t test or by ANCOVA adjusting for race. Differences in insulin sensitivity, first-phase insulin, and DI across HbA_1c categories (normal versus prediabetes) were determined by two-tailed t test and also by ANCOVA adjusting for race and adiposity (BMI, percent body fat, or VAT). Overall differences across subgroups of NGT versus dysglycemia by OGTT and normal versus prediabetes by HbA_1c were determined using ANCOVA, adjusting for race and VAT in SPSS (PASW 18; SPSS Inc., Chicago, IL). Data are presented as mean ± SE. A P value of ≤0.05 was considered statistically significant, and P ≤ 0.10 was considered a trend.

Age, sex, and Tanner stage distributions were not different between participants divided according to HbA_1c category (normal versus prediabetes) (Table 1). There were more Caucasians in the normal HbA_1c category and more African Americans in the prediabetes category. Adiposity measures (BMI, BMI percentile, percent body fat, and VAT) were significantly higher in the prediabetes HbA_1c category than normal HbA_1c, before and after adjustment for race.

Fasting glucose, insulin, OGTT glucose AUC, and insulin AUC were higher, and HIS was lower, in the prediabetes versus normal HbA_1c category (Table 1). These differences persisted after adjusting for race and percent body fat. Peripheral insulin sensitivity was lower in the prediabetes versus normal HbA_1c category before and after adjustment for race (P = 0.004; data not shown), but was not different after additional adjustment for adiposity (Fig. 1A). First-phase insulin was not different between the two groups; however, DI, β-cell function relative to insulin sensitivity, was significantly lower in the prediabetes versus normal HbA_1c category after adjusting for race and adiposity (BMI, percent body fat, or VAT; Fig. 1A). In the subset of participants with OGTT-derived indices (Fig. 1B), results of insulin sensitivity (1/I_F) and oDI mirrored the results observed with the clamps.

In the normal HbA_1c category, 27% had dysglycemia, and in the prediabetes HbA_1c category, 41% had dysglycemia (P = 0.09). Within each HbA_1c category, DI was lower by ∼35% in dysglycemia versus NGT groups. After adjustment for race and VAT, there were significant differences in DI among the four groups (P < 0.001), with the highest value in NGT in the normal HbA_1c category (562 ± 33 mg/kg/min) and the lowest value in the dysglycemia group in the prediabetes HbA_1c category (308 ± 58 mg/kg/min) (Fig. 2).

The current study demonstrates that overweight/obese adolescents meeting the ADA HbA_1c diagnostic criteria (2) for prediabetes have evidence of impaired β-cell function relative to insulin sensitivity, a metabolic signal for heightened risk of type 2 diabetes (27). This finding persisted after adjustment for race and for the greater adiposity in the youth with elevated HbA_1c. These data support our hypothesis that alterations in the pathophysiologic mechanisms of type 2 diabetes are evident in the ADA-recommended at-risk/prediabetes category (HbA_1c 5.7 to <6.5%).

Since the release of the HbA_1c-based diagnostic criteria for prediabetes/diabetes by the ADA in 2010 (2), reports of adults and some in pediatrics have concluded that HbA_1c is inferior to glycemic measures for diagnostic purposes (4–6,11). These studies evaluated the diagnostic sensitivity and specificity of HbA_1c compared with glycemic measures, treating the latter as the gold standard. One such study determined that the inferior sensitivity of the recommended HbA_1c cutoff compared with OGTT was worse for adolescents than for adults (4). However, although these studies evaluated equivalence of the new HbA_1c criteria compared with glycemic criteria, they did not evaluate metabolic characteristics linked to future diabetes risk. Moreover, as noted in a Comment published in Diabetes Care, “it is a fallacy that the OGTT is a gold standard… if you define one test as a gold standard, all comparators will be inferior” (28). Indeed, in longitudinal studies, HbA_1c criteria predicted a similar rate of progression to diabetes as did IFG, and the combination of fasting glucose and HbA_1c provided the greatest predictive value of diabetes incidence (9,10).

The DI, which expresses β-cell function relative to insulin sensitivity, is an established metabolic predictor of progression to diabetes (12,13,29–31). Numerous longitudinal studies in adults have demonstrated the predictive value of DI (including oral DI) for the future development of diabetes (27,29,32,33). In addition, we (12,13) and others (30,31) have reported progressively declining DI across the spectrum of glycemia from normal to IFG and/or IGT to type 2 diabetes in youth. Therefore, even though the ADA HbA_1c diagnostic criteria have a lower sensitivity for diagnosis of prediabetes when compared with glycemic measures, we show in this study that these criteria are consistent with greater metabolic risk (lower clamp DI, oral DI, and insulinogenic index) using sensitive metabolic assessments.

Evaluating HbA_1c and glycemia together, youth with normal HbA_1c combined with normoglycemia had the highest DI compared with the other groups, and DI seemed to be lowest overall in the group with both elevated HbA_1c and dysglycemia (Fig. 2). Moreover, rates of dysglycemia were higher in the prediabetes (41%) versus normal HbA_1c category (27%) (Fig. 2). Overall, these data support the use of HbA_1c to identify youth with lower β-cell function for epidemiological observational and/or interventional studies of overweight/obese youth and progression to type 2 diabetes. However, the independent use of HbA_1c for clinical diagnostic purposes may be premature. Collectively, these observations indicate that the combination of HbA_1c and a glycemic measure may best pinpoint subjects at lowest or greatest risk of eventual diabetes (9,10). Longitudinal studies will be necessary to determine whether there are differences in rates of the development of diabetes between youth with normal HbA_1c and dysglycemia versus those with elevated HbA_1c and dysglycemia.

A potential challenge in the application of the HbA_1c is the higher HbA_1c in African Americans than Caucasians, present in our study and reported by others (5,34). Racial differences in HbA_1c result in more African Americans and fewer Caucasians being identified as having prediabetes using HbA_1c compared with glycemic criteria (6,7). Potentially, the ADA-recommended cut points may not be appropriate for all racial groups, and exploration of cut points customized for race/ethnicity in larger scale studies may be warranted. Regardless, in this study, we report impaired β-cell function as well as lower peripheral and hepatic insulin sensitivity in youth with elevated versus normal HbA_1c, independent of the effects of race.

Strengths of the current investigation include the simultaneous assessment of HbA_1c, clamp-measured in vivo insulin sensitivity and β-cell function, and an OGTT in a large group of adolescents. A limitation of this report is that we included only overweight/obese adolescents, and our population does not represent other minority groups (e.g., Hispanics) in which the utility of HbA_1c may vary (34). Lastly, the number of youth with elevated HbA_1c and dysglycemia is somewhat limited.

In conclusion, the recently recommended HbA_1c criteria coincide with lower peripheral and hepatic insulin sensitivity and lower β-cell function relative to insulin sensitivity in overweight/obese adolescents. Larger and longitudinal studies are required to more thoroughly investigate the interplay of race within categories of dysglycemia and HbA_1c in relation to metabolic risk factors for diabetes (e.g., DI). Although controversy is likely to continue surrounding the adoption of HbA_1c-based criteria for the diagnosis of prediabetes and diabetes in youth, these data support that the ADA HbA_1c criteria for prediabetes correspond to impaired β-cell function in overweight/obese youth, a metabolic marker of heightened type 2 diabetes risk.

This project was supported by U.S. Public Health Service grants R01 HD-27503 (to S.A.A.) and K24 HD-01357 (to S.A.A.), Richard L. Day Endowed Chair (to S.A.A.), the Department of Defense (to L.A.S., S.L., H.T., F.B., and S.A.A.), and the National Institutes of Health through grants M01-RR-00084, UL1-RR-024153, and UL1-TR-000005.

No potential conflicts of interest relevant to this article were reported.

L.A.S. analyzed the data and first-authored the manuscript. S.F.M. provided technical support for data analysis, editing, and review of the manuscript. S.L., H.T., and F.B. contributed research participants and data. L.F. managed the data entry and database maintenance. S.A.A. provided the study concept and design; acquired data; obtained funding; provided administrative, technical, and material support; supervised the study; and critically reviewed and edited the manuscript. S.A.A. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Parts of this study were presented orally at the 72nd Scientific Sessions of the American Diabetes Association, Philadelphia, Pennsylvania, 8–12 June 2012.

These studies would not have been possible without the nursing staff of the Pediatric Clinical and Translational Research Center, the devotion of the research team (Nancy Guerra, CRNP; Kristin Porter, RN; Sally Foster, RN, CDE; and Lori Bednarz, RN, CDE), the laboratory expertise of Resa Stauffer, and, most importantly, the commitment of the study participants and their parents.