Degree of Obesity and Glucose Allostasis Are Major Effectors of Glucose Tolerance Dynamics in Obese Youth

Participants in this cohort were recruited to perform an OGTT from the Yale Pediatric Obesity Clinic as part of a study of the pathophysiology of type 2 diabetes in youth. Some of these subjects have been described in previous publications (7,8). All participants were between the ages of 4 and 18 years. All underwent a physical examination that included anthropometric determinations and pubertal status (9) and were classified as prepubertal or pubertal. Subjects with overt type 2 diabetes at baseline and those with medical conditions or using medications that may affect glucose metabolism were excluded from the study. All subjects had a BMI >95th percentile for age and sex and were thus classified as obese. To standardize the BMI levels, conversion to BMI Z scores was performed based on Centers for Disease Control growth charts (10). Participants were followed biannually as outpatients by the clinical staff and received nutritional guidance as well as recommendations for physical activity, as previously described (11). All participants of this cohort are encouraged to return to repeat their OGTT within ∼18 months. Data from the 784 obese children and adolescents who had a baseline OGTT were analyzed for the cross-sectional analysis, and data from the subset of 181 participants who performed a second OGTT was used for the longitudinal analysis. The mean follow-up interval was 22 ± 12 months. Anthropometric and metabolic characteristics of the baseline cross-sectional cohort and the subset that had a second OGTT were comparable (Table 1). The protocol was approved by the institutional review board of the Yale University School of Medicine. Written informed consent was obtained from the parents and assent from the children and adolescents.

Subjects were studied at the Yale Children's Clinical Research Center at 8:00 a.m. after a 12-h overnight fast as previously described (7,8). Two baseline samples were then obtained for measurements of plasma glucose, insulin, C-peptide, and lipids. Thereafter, flavored glucose (Orangedex; Custom Laboratories, Baltimore, MD) in a dose of 1.75 g/kg body wt (up to a maximum of 75 g) was given orally, and blood samples were obtained every 30 min for 180 min for the measurement of plasma glucose, insulin, and C-peptide.

Plasma glucose level was determined with a YSI 2700 STAT Analyzer (Yellow Springs Instruments). Plasma insulin was measured with a radioimmunoassay made by Linco (St. Charles, MO), which has <1% cross-reactivity with C-peptide and proinsulin. Plasma C-peptide levels were determined with an assay made by Diagnostic Products (Los Angeles, CA). The intraassay variation was 11% for insulin and 13% for C-peptide, and the interassay variation was 12% for insulin and 12% for C-peptide

The insulinogenic index (IGI), calculated as the ratio of the increments in insulin and glucose levels during the first 30 min after the ingestion of glucose, was used to assess early β-cell response (12). Insulin sensitivity was determined by the whole-body insulin sensitivity index (WBISI, the Matsuda index) (13). The DI was calculated as the product of IGI and the square root of WBISI, based on the curvilinear relation of these OGTT derived indexes, as previously described (6). To evaluate a component along the DI hyperbola, we followed the approach of Stumvoll et al. (3) and calculated the ratio of IGI and WBISI to derive the β-cell demand index (BCDI). This index represents the additional burden placed on the β-cell in the face of a decrease in insulin sensitivity while maintaining a constant DI. As shown in Fig. 1, for a given DI, the BCDI rises as insulin sensitivity decreases and may be used to estimate the metabolic burden placed on the β-cell in order to provide adequate compensation.

Data are presented as means ± SD. Parameters that did not have a normal distribution were log transformed for the analysis. Adjustment for comparisons for potential confounders was performed using ANCOVA using the general linear model procedure. Adjustment for multiple comparisons was performed using the Bonferonni procedure. Division of the cohort for the cross-sectional analysis to quartiles of IGI, WBISI, DI, and BCDI was based on the 25th, 50th, and 75th percentile of these parameters for presentation purposes. Adjustments were made for age, sex, ethnicity, and BMI Z score. For the longitudinal analysis, we similarly divided the cohort into four categories of changes in BCDI by initially dividing the cohort to those who had a positive or negative change and then dividing these two groups using the median of each one. Thus, the BCDI change categories reflect the direction (positive or negative) and the magnitude (large or small) of BCDI change. Baseline BMI Z score was divided into equal quartiles based on corresponding percentiles in order to evaluate the impact of baseline degree of obesity on later glycemic levels, independent of dynamics of insulin sensitivity, secretion, and their interrelations. An α < 0.05 was considered statistically significant. All analyses were performed using SPSS 12.0 for Windows.

The relations of the 2-h glucose, fasting glucose, and quartiles of IGI and WBISI are shown in the online appendix (available at http://dx.doi.org/10.2337/dc07-0325). The effect of IGI and WBISI quartiles on 2-h glucose was significant (P < 0.001), while the interaction between them was not (implying that they are independent and additive but not synergistic). These effects remained statistically significant after adjustment for age, sex, ethnicity, pubertal status, and BMI Z score. Similarly, IGI and WBISI quartiles had significant effects on fasting glucose (P < 0.001), while the interaction between them did not. These effects remained statistically significant after similar adjustments, although, interestingly, sex had a significant effect in this model (P < 0.001), with males having a greater fasting glucose than females (92 vs. 89 mg/dl, P < 0.001).

The relations of the DI and the BCDI and 2-h and fasting glucose levels are shown in Fig. 2. The effects of DI and BCDI on 2-h glucose levels were significant individually (P < 0.001), as was the interaction between them (P < 0.001), reflecting the concept of allostasis per se (i.e., to maintain a constant DI in the face of worsening insulin sensitivity, the signal for the β-cell to compensate by adequate secretion of insulin is rising ambient glycemia). The significant interaction between the two implies a synergistic effect, i.e., per given insulin sensitivity, a lower DI implies a greater BCDI, leading to an even greater glycemic signal in order to provide adequate compensation. These effects remained significant after adjustment for age, sex, ethnicity, pubertal status, and BMI Z score. The effects of DI and BCDI on fasting glucose were significant (P < 0.001), yet the interaction between them was not (P = 0.10). After similar adjustment, the effects of DI and BCDI on fasting glucose remained highly significant. Sex again had a significant effect on fasting glucose (92 mg/dl for males vs. 89 mg/dl for females, P < 0.001).

To evaluate the impact of changes in the BCDI on glucose levels, we modeled the changes in fasting and 2-h glucose between baseline and follow-up OGTTs while adjusting for baseline DI, baseline BCDI, change in DI, baseline fasting glucose, age, sex, ethnicity, pubertal status, time between studies, and baseline BMI Z score. As shown in Fig. 3B, the upper quartile of BCDI change, which had the greatest increase of β-cell demand, had significantly greater 2-h glucose level change on follow-up compared with the 1st and 2nd BCDI change categories, which had a reduction of β-cell demand (P = 0.006 and P = 0.003, respectively). Similarly, those in the upper category of BCDI change had a greater change in fasting glucose levels on follow-up compared with the first category (P = 0.02). Baseline BMI Z, baseline BCDI, and change in DI had a significant effect on 2-h glucose in this model (P = 0.007, P = 0.007, and P < 0.001, respectively). The most obese participants at baseline had significantly greater increases in 2-h glucose on follow-up compared with the other three quartiles (Fig. 4D). This difference remained significant after adjustment for age, sex, ethnicity, pubertal status, time between studies, baseline fasting glucose, BCDI and DI, and changes in BCDI and DI. Interestingly, adding the change in obesity over time to the model, expressed as BMI Z change or as absolute weight change, had no significant effect on the model and did not modify the significance of baseline BMI or BCDI change category. Baseline degree of obesity had no effect on follow-up fasting glucose (Fig. 4B).

In the present analysis we used the OGTT to demonstrate the phenomenon of glucose allostasis, manifested as an increase in ambient glycemia in order to maintain a constant DI while insulin sensitivity is decreasing. The main findings of this analysis show that OGTT-derived indexes can indeed demonstrate the allostatic effect of rising ambient glycemia in the face of increased β-cell demand. At least one of the signals for the β-cell to increase insulin secretion is glucose, which is mildly yet significantly elevated in order to provide a continuous stimulation to the pancreas. Our longitudinal analysis demonstrated that increasing β-cell demand is independently related to elevations in ambient glycemia and that the degree of obesity at baseline has an adverse effect on follow-up glucose levels, independent of dynamics of insulin sensitivity and β-cell demand. Moreover, baseline BMI Z score was a significant contributor to elevated glucose levels on the second OGTT, while the change in degree of obesity during the follow-up was not.

The emergence of type 2 diabetes in the pediatric age-group has raised questions regarding the underlying pathophysiology of this condition in comparison with adults. The rapid tempo of the development of β-cell failure in youth suggests that parameters less relevant in adults, such as hormonal changes of puberty, might accelerate this process. Our data clearly demonstrate that the phenomenon of glucose allostasis (14), previously described in adults, is similarly relevant in obese children and contributes to the development of hyperglycemia over time. Indeed, changes in the β-cell demand index in this cohort were independently related to changes in fasting and 2-h glucose levels. This phenomenon can be observed in puberty-related hormonal changes in nonobese adolescents, which are related to a ∼33% decrease in insulin sensitivity, and with a transient increase of 3.5 mg/dl in fasting glucose, that probably returns to prepubertal levels upon reaching Tanner stage V (15). Obesity is the major cause of peripheral insulin resistance in childhood and is tightly related to the development of altered glucose metabolism (16). Indeed, the mean value of the insulin sensitivity index (WBISI) in this large cohort was ∼2.0, reflecting baseline insulin resistance. At this level of sensitivity, further small decreases of insulin sensitivity necessitate large changes of insulin secretion in order to maintain a constant DI, thus causing major increases in β-cell demand due to the increasing slope of the DI hyperbola in that numerical range of WBISI (Fig. 1). In this scenario, adding a second element that further reduces insulin sensitivity, such as pubertal hormonal changes (17) or exogenous glucocorticoids (18), might tip the balance toward extreme β-cell demand that can no longer be compensated by the pancreas, leading to β-cell failure and increased glucose levels.

The longitudinal assessment demonstrated that the degree of obesity at baseline had an effect on increased levels of fasting and 2-h glucose levels during follow-up. This effect was independent of changes in insulin sensitivity, baseline DI and BCDI, and changes in DI and BCDI and, importantly, more significant than changes in relative adiposity (BMI Z score change) or absolute weight over time. This finding suggests that fat-derived factors may have an independent role in the development of deteriorating glucose tolerance that is not mediated by effects on insulin sensitivity and that those with severe obesity are a specifically high-risk group for deteriorating glucose tolerance. Potential mediators of the adverse effects of severe obesity on β-cells include, among others, increased levels of leptin (19), free fatty acids (20), and tumor necrosis factor-α and reduced adiponectin (21). Elevated leptin (22), free fatty acids (23), and tumor necrosis factor-α (24) have all been shown to hamper insulin secretion, while adiponectin seems to have protective effects on β-cells (25).

The major impact of the changes of β-cell demand and of baseline degree of obesity over time on dynamics of fasting and postprandial glucose levels raises several therapeutic implications. Obviously, prevention of severe obesity should be in the frontline of all therapeutic interventions. In the severely obese child, reduction of β-cell demand through use of insulin sensitizers or increasing the early insulin response using secretogogues may be of benefit. Such therapies, as well as pharmacologic interventions directed at weight loss per se may also have beneficial effects on circulating concentrations of adipocytokines that will further reduce the burden on the β-cell. Indeed, targeting both β-cell enhancement and weight reduction in combination may be superior to treating only one. Treatment of obesity-related subclinical inflammation, regardless of weight loss, may provide another therapeutic target in this context (26).

In summary, increasing β-cell demand related to worsening insulin sensitivity and degree of obesity per se have independent roles in the development of elevated glucose levels over time. Elevation of glucose levels may be a normal physiological adaptation to create a signal for the β-cell to face the increased demand (allostasis) and may be a consequence of the unique proinflammatory milieu characteristic of the severely obese child. This implicates that peripheral insulin sensitization and/or β-cell enhancement alongside a significant reduction in obesity may be needed to prevent the development of altered glucose metabolism in obese youth.

This study was supported by the National Institutes of Health (RO1-HD40787, RO1-HD28016, and K24-HD01464 to S.C.; K12-DK063709 to W.V.T.; and MO1-RR00125 to Yale General Clinical Research Center) and the Hadassah-Yale Pediatric Diabetes Research Fund (to R.W.).