Are Obesity-Related Metabolic Risk Factors Modulated by the Degree of Insulin Resistance in Adolescents? Free

Fifty-four obese (mean BMI 34.9 ± 5.5 kg/m²) otherwise-healthy African-American and American white adolescents were studied. Subjects had exogenous obesity with no clinical evidence of endocrinopathy or syndromes. They were not involved in any regular physical activity or weight reduction programs and were not on medications that affect glucose metabolism. Female subjects were evaluated in the follicular phase of their menstrual cycle. All studies were approved by the institutional review board of the University of Pittsburgh. Study participants were recruited through newspaper advertisements in the community. Parental informed consent and child assent were obtained. Clinical characteristics of the study subjects are summarized in Table 1. All subjects were at Tanner stage II–V of puberty on examination, as confirmed by plasma testosterone in male and estradiol in female subjects.

Each participant underwent a hyperinsulinemic-euglycemic clamp and a hyperglycemic clamp study after 10–12 h of fasting, at 1- to 3-week intervals, in random order. Participants were admitted to the General Clinical Research Center the afternoon before the day of the testing.

A fasting blood sample was obtained for determination of cholesterol, LDL, HDL, VLDL, triglycerides, HbA_1c (A1C), proinsulin, C-peptide, and adiponectin. Fasting endogenous glucose production was measured with a primed constant rate infusion of [6,6-²H₂] glucose (0.306 ± 0.009 μmol · kg⁻¹ · min⁻¹) (Isotech, Miamisburg, OH) (7,8). Blood was sampled at the start of the 2-h isotope infusion and every 10 min from −30 to 0 times (basal period) for determination of plasma glucose, insulin, and isotopic enrichment of glucose. Fasting turnover calculations were made during the last 30 min of the basal period. Insulin-mediated glucose metabolism and insulin sensitivity were evaluated during a 3-h hyperinsulinemic-euglycemic clamp (7,8). Intravenous crystalline insulin (Humulin; Lilly, Indianapolis, IN) was infused at a rate of 80 mU/m² per min (7). Plasma glucose was clamped at 5.6 mmol/l with a variable rate infusion of 20% dextrose, based on arterialized plasma glucose determinations every 5 min. Continuous indirect calorimetry by a ventilated hood (Deltatrac metabolic monitor; Sensormedics, Anaheim, CA) was used to measure CO₂ production, O₂ consumption, and respiratory quotient. Measurements were made for 30 min at baseline and at the end of the euglycemic clamp (9).

First- and second-phase insulin secretion was evaluated during a 2-h hyperglycemic clamp (8). Plasma glucose was rapidly increased to 12.5 mmol/l and maintained by a variable rate infusion of 20% dextrose. The night before the clamp, blood pressure was measured between 10 and 11 p.m. and the next morning between 6 and 7 a.m. in the resting supine position (8).

Body composition was determined by dual-energy X-ray absorptiometry and subcutaneous abdominal adipose tissue and visceral adipose tissue (VAT) by a single-slice computed tomography scan at L₄-L₅ (8).

Maximal O₂ uptake was measured using the Bruce multistage treadmill protocol reported by us in children before (10). Vo_2max was indexed to total body mass (ml · kg⁻¹ · min⁻¹) and fat-free mass (FFM; ml · min⁻¹ · kg FFM⁻¹).

Plasma glucose was measured with a glucose analyzer (Yellow Springs Instrument, Yellow Springs, OH), insulin by radioimmunoassay (8), adiponectin by a radioimmunoassay kit (Linco Research) (7), A1C by high-performance liquid chromatography (Tosoh Medics), and lipids using the standards of the Centers for Disease Control and Prevention (7). Deuterium enrichment of glucose in the plasma was determined on a Hewlett-Packard 5971 mass spectrometer (Hewlett-Packard, Palo Alto, CA) coupled to a 5890 series II gas chromatograph as before (7).

Fasting hepatic glucose production was calculated during the last 30 min of the 2-h isotope infusion according to steady-state tracer dilution equations (7). Insulin-stimulated glucose disposal rate (R_d) was calculated during the last 30 min of the euglycemic clamp to be equal to the rate of exogenous glucose infusion. Peripheral insulin sensitivity was calculated by dividing the R_d by the steady-state (SS_EU) clamp insulin level (8). Metabolic clearance rate of insulin was calculated by dividing the insulin infusion rate by the Δ increase in circulating insulin during the SS_EU (8). Insulin-stimulated carbohydrate and lipid oxidation rates were calculated according to the formulas of Frayn (11) from the indirect calorimetry data. Nonoxidative glucose disposal was estimated by subtracting the rate of glucose oxidation from the total R_d.

During the hyperglycemic clamp, the first-phase insulin concentration was calculated as the mean of five determinations from 2.5 to 12.5 min after the dextrose bolus. The second phase was calculated as the mean of eight determinations from 15 to 120 min (8). Glucose disposition index was calculated as the product of insulin sensitivity times first-phase insulin.

Statistical analyses were performed using ANOVA followed by post hoc Bonferroni correction for three group comparisons. We used paired t test or Wilcoxon rank-sum analysis to compare the obese pair-matched groups, Spearman’s correlation and multiple regression analyses for bivariate and multivariate relationships, and curve estimation regression statistics to evaluate the relationship of insulin secretion to insulin sensitivity. Data are presented as means ± SE. Two-tailed P value ≤0.05 was considered statistically significant.

The obese adolescents were divided into two groups based on cut points for insulin sensitivity in normal-weight adolescents (NW group). The NW and obese adolescents were part of our ongoing investigations of determinants of insulin sensitivity and secretion during childhood. The insulin sensitivity of the MIR group was within 2 SDs below the mean of the NW subjects, and the insulin sensitivity of the SIR group was <2 SDs (<2.35 μmol · kg⁻¹ · min⁻¹ per pmol/l) of the mean insulin sensitivity of NW adolescents. Table 1 depicts the characteristics of the two obese groups (MIR versus SIR) versus NW adolescents. Of 54 obese adolescents, 23 (43%) were categorized as MIR and 31 (57%) as SIR subjects. The SIR group had higher BMI, waist-to-hip ratio, and visceral and subcutaneous fat compared with the MIR and NW groups. There were no differences in waist-to-hip ratio and lipid profile between the MIR and NW groups.

To control for differences in BMI, race, and sex between the SIR and MIR groups, we pair matched 17 MIR and 17 SIR subjects with respect to these variables. All subjects were pubertal (Tanner II–V). Despite similar BMI, percent body fat, fat mass, and subcutaneous abdominal fat, SIR adolescents had significantly greater VAT and higher waist-to-hip ratio, fasting insulin, C-peptide, proinsulin, and triglyceride-to-HDL ratio and lower HDL and a tendency for lower adiponectin levels than MIR subjects. The MIR and SIR adolescents had similar fasting glucose (5.4 ± 0.06 vs. 5.5 ± 0.09 mmol/l), A1C (5.3 ± 0.1% in both), and systolic and diastolic blood pressure (Table 2).

Because there were significant differences between the pair-matched MIR versus SIR subjects in HDL, VAT, and waist-to-hip ratio despite similar BMI, we proceeded to investigate if differences in insulin sensitivity among these obese children determine cardiovascular outcome measures (HDL, LDL, triglyceride-to-HDL ratio, and systolic and diastolic blood pressure) independent of BMI. We performed multiple regression analysis with each outcome as the dependent variable and BMI and insulin sensitivity as independent variables. Insulin sensitivity independent of BMI explained ∼8% of the variance in HDL (R² = 0.08, P = 0.03) but not in the other dependent variables. On the other hand, VAT independent of BMI and of insulin sensitivity explained ∼20% of the variance in triglyceride-to-HDL ratio (R² = 0.198, P < 0.001). Similarly, waist-to-hip ratio independent of insulin sensitivity explained ∼15% of the variance in the triglyceride-to-HDL ratio (R² = 0.15, P = 0.04). The variance in LDL (R² = 0.10, P = 0.02) and systolic (R² = 0.10, P = 0.03) and diastolic (R² = 0.12, P = 0.001) blood pressure was attributable to BMI independent of insulin sensitivity.

Per design, insulin sensitivity was lower in the SIR versus MIR pair-matched children (1.4 ± 0.1 vs. 3.6 ± 0.3 μmol · kg⁻¹ · min⁻¹ per pmol/l, P < 0.001). This remained true when insulin sensitivity was expressed per kilogram FFM (2.6 ± 0.2 vs. 6.7 ± 0.5 μmol · min⁻¹ · kg FFM⁻¹ per pmol/l, P < 0.001). Similarly, metabolic clearance rate of insulin was lower in the SIR group (6.5 ± 0.3 vs. 9.6 ± 0.6 ml · kg⁻¹ · min⁻¹, P = 0.001). Glucose disposition index was lower in the SIR group (1.91 ± 0.22 vs. 3.70 ± 0.37 mmol · min⁻¹ · kg⁻¹, P = 0.002), despite higher first-phase insulin (1,422.4 ± 194.3 vs. 1,066.2 ± 94.0 pmol/l, P = 0.022) (Fig. 1). Hepatic glucose production was not different between the two groups (13.4 ± 0.85 vs. 13.0 ± 0.70 μmol · kg⁻¹ · min⁻¹). Total, oxidative, and nonoxidative glucose disposal were lower in SIR subjects. During hyperinsulinemia, the suppression in fat oxidation was lower in SIR adolescents (1.29 ± 0.19 vs. 1.94 ± 0.26 μmol · kg⁻¹ · min⁻¹, P = 0.039) (Fig. 2).

Vo_2max, a measure of CRF, tended to be lower in SIR adolescents (Table 2). Resting energy expenditure, respiratory quotient, glucose, and fat oxidation were not different between the two groups (data not shown). However, during insulin-stimulated conditions of the hyperinsulinemic clamp, respiratory quotient (0.89 ± 0.01 vs. 0.93 ± 0.01, P = 0.017), energy expenditure (22.3 ± 0.67 vs. 23.9 ± 0.75 kcal · kg⁻¹ · min⁻¹, P = 0.05), and glucose oxidation (14.3 ± 0.9 vs. 17.6 ± 0.6 μmol · kg⁻¹ · min⁻¹, P = 0.005) were lower, while fat oxidation was higher (1.58 ± 0.19 vs. 0.90 ± 0.17 μmol · kg⁻¹ · min⁻¹, P = 0.022) in the SIR versus MIR group.

The present study demonstrates that despite similar BMI, there are obese adolescents who are only moderately insulin resistant and at lower risk for obesity-associated comorbidities compared with severely insulin-resistant ones. These MIR adolescents have higher physical fitness, lower VAT, higher adiponectin levels, and better substrate utilization and energy consumption compared with the SIR group. Moreover, they have a preserved balance of β-cell secretory compensation to insulin resistance, lessening their risk of progression to type 2 diabetes. Furthermore, their lower triglyceride-to-HDL ratio and higher HDL profile would suggest lower risk of CVD. This study adds to the limited existing literature by providing 1) a comparison between two groups of obese adolescents, severely versus moderately insulin resistant, strictly defined based on data in normal-weight youth and pair matched for BMI, ethnicity, sex, and puberty; 2) information on CRF; 3) information on energy and substrate utilization; and 4) in vivo evaluation of insulin sensitivity and secretion simultaneously.

In recent years, researchers became aware of the existence of “fat-fit” individuals (12). In population studies, based on measurement of insulin sensitivity with the hyperinsulinemic-euglycemic clamp in 1,146 healthy Caucasians, aged 18–85 years, it was shown that insulin resistance (defined as <10% of the M value of lean subjects) in “simple obesity” is not as prevalent as previously thought (1). Accordingly, only 26% of all obese subjects were insulin resistant. Bonora et al. (13) similarly reported that insulin resistance, estimated by the homeostasis model, was present in 42% of overweight subjects with no metabolic disorders.

In our current study, 57% of obese adolescents were severely insulin resistant, while 43% had insulin sensitivity within 2 SDs of that of the NW group. Adiposity and insulin sensitivity independently appear to modulate different CVD risk factors. In our multiple regression analysis, insulin sensitivity determined HDL independent of BMI, while LDL and blood pressure were determined by BMI. This is consistent with adult data showing a decrease in HDL with an increase in the tertile of insulin resistance, while LDL worsened with an increase in BMI tertile (14). Another observation is that despite similar BMI, MIR subjects had lower VAT, waist-to-hip ratio, and a tendency for higher adiponectin than SIR subjects. This is consistent with adult data of lower VAT in metabolically normal obese postmenopausal women (4) and lower adiponectin levels in obese and nonobese insulin-resistant versus insulin-sensitive men and women (3). Lower visceral and intramyocellular fat and higher adiponectin was reported (5) in obese insulin-sensitive youth. However, subjects were of three different ethnicities, potentially impacting the outcome measures, particularly race-related differences in visceral adiposity (15).

The difference in insulin sensitivity between the MIR and SIR groups was related to both oxidative and nonoxidative (storage) glucose disposal, which were lower in the SIR group. They also had less suppression of fat oxidation and lower respiratory quotient in response to hyperinsulinemia, indicative of insulin resistance in suppressing fat oxidation. This is consistent with the higher VAT in the SIR group. Our results differ from those of Weiss et al. (5), in which lipid and glucose oxidations were not significantly different between the insulin-sensitive and -resistant groups. This could again stem from including different racial groups in their study despite reports of race-related differences in lipolysis, substrate oxidation (16,17), and energy expenditure (18). Lower glucose oxidation and higher fat oxidation were observed in insulin-resistant subjects and were proposed as one of the mechanisms for limiting additional weight gain in insulin-resistant versus insulin-sensitive obese Pima Indians in a longitudinal study (19).

In addition, SIR youth have evidence of suboptimal insulin compensation resulting in a lower glucose disposition index compared with their MIR peers. This finding, along with hypoadiponectinemia, which was associated with the future development of type 2 diabetes (20), suggests a higher propensity to progress to an impaired glucose-tolerant state and diabetes (21).

Another observation is the tendency of MIR subjects to have higher CRF. CRF, which is influenced by physical activity, has been associated with lower rates of CVD and all-cause mortality in adult populations (22–24). In female adolescents with a wide range of BMI, Vo_2max was a more critical determinant of insulin sensitivity than percent body fat (25). In the study by Brochu et al. (4), the insulin-sensitive versus -resistant obese women had similar CRF. It remains to be determined whether the tendency for higher CRF in the MIR group will become more significant if more subjects are studied.

From the practical clinical perspective, the SIR children had significantly higher waist-to-hip ratio, higher fasting insulin, lower HDL, and a tendency for worse CRF and lower adiponectin levels. Similar findings were reported by the few studies addressing the metabolic risk of obesity in adults (3,4) and children (5). However, more research is needed in children to derive specific cut points for the different variables with careful assessment of sensitivity and specificity to enable the distinction between moderately versus severely insulin-resistant children.

In summary, we conclude that despite similar BMI in obese youth, there are differences in the degree of insulin sensitivity and metabolic consequences. The severely insulin-resistant youth are at greater risk for obesity-related comorbidities, including the risk of type 2 diabetes and dyslipidemia. Whether differences in insulin sensitivity stem from higher visceral adiposity or lead to abdominal obesity remains to be determined. Also, it remains to be determined whether this risk phenotype is genetically programmed yet environmentally modulated to allow for therapeutic interventions.

This work was supported by the U.S. Public Health Service Grants RO1 HD27503, K24 HD01357, and MO1-RR00084; the General Clinical Research Center (GCRC); and Eli Lilly.

We thank the GCRC nurses for expert nursing assistance and Pat Antonio for secretarial assistance. These studies would not have been possible without the recruitment efforts of Sandy Stange, the laboratory expertise of Resa Brna, and, most importantly, the commitment of the volunteer children and their parents.