Metabolic Inflexibility in Youth with Obesity: Is it a Feature of Obesity or Distinctive of Youth Who are Metabolically Unhealthy?

Abstract

Background:

Individuals with obesity have metabolic inflexibility with diminished fasting fat oxidation and blunted increase in respiratory quotient (RQ) in insulin-stimulated states. However, it is unclear if metabolic inflexibility is a characteristic of obesity per se or is unique to youth who have metabolically unhealthy obesity (MUO) compared with metabolically healthy obesity (MHO).

Objective:

We investigated metabolic flexibility in youth with MUO, MHO and normal weight (NW) and compared their metabolic characteristics.

Patients/Design:

Youth (n=188) were divided, based on cut-off points for in vivo insulin sensitivity (IS) of adolescents with NW, into 137 with MUO and 51 with MHO. Fasting hepatic IS (HIS) from hepatic glucose production by [6,6-²H₂]glucose, adipose tissue IS (ATIS) from whole-body lipolysis by [²H₅]glycerol, RQ (indirect calorimetry) during fasting and a hyperinsulinemic (80mu/m²/min)-euglycemic clamp were measured.

Results:

Youth with MUO vs. MHO had blunted ΔRQ (p=0.035) and lower HIS and ATIS (both p<0.0001), while ΔRQ, HIS and ATIS were not different between youth with MHO and NW. In a pair-matched sub-analyses of 30 MUO and 30 MHO the results were similar to the total cohort.

Conclusions:

Metabolic inflexibility, does not appear to be a feature of obesity per se rather distinctive of youth with MUO, who also have worse HIS and ATIS compared with youth with MHO.

1. Introduction

Childhood obesity is described by the World Health Organization as “one of the most serious public health challenges of the 21^st Century” as it continues to increase worldwide (1). According to the 2015-2016 National Health and Nutrition Examination Survey, 18.5% of U.S youth aged 2-19 years have obesity, including 5.6% with severe obesity and another 16.6% who are overwieght (1). Furthermore, between 1988-1994 and 2013-2014 the overall prevalence of obesity and extreme obesity increased particularly among adolescents aged 12 to 19 years old (2).

Pediatric obesity is associated with significant comorbidities including insulin resistance, prediabetes/type 2 diabetes, prehypertension/hypertension, dyslipidemia, nonalcoholic fatty liver disease (NAFLD), hyperandrogenemia/polycystic ovary syndrome (PCOS), early subclinical atherosclerosis and premature mortality in adulthood (3). Nevertheless, accumulating evidence suggests that not all individuals with obesity have these comorbidities (4). Consequently, obesity is now classified as either metabolically healthy obesity (MHO) or metabolically unhealthy obesity (MUO) with the former being protected against cardiometabolic risks (4-8). Because there is no uniform definition of pediatric MHO, most available literature has defined metabolic health based on the absence of obesity-associated metabolic abnormalities including insulin resistance (9-11). Studies in pediatrics, similar to adults, show that adolescents with MHO (defined based on various methods of assessment for insulin resistance/insulin sensitivity) have favorable physical, metabolic, hormonal, and cardiovascular disease characteristics (9-15).

Over the past few decades great progress has been made to understand metabolic flexibility and its underpinnings in health and disease (16-20). Metabolic flexibility reflects the ability to adjust fuel oxidation to fuel availability and to switch from lipid oxidation to carbohydrate oxidation during insulin stimulation. Metabolic flexibility is impaired during long-term caloric excess, and metabolic inflexibility has been demonstrated in the context of obesity, insulin resistance, metabolic syndrome and type 2 diabetes (17,19,21). Individuals with obesity have metabolic inflexibility evidenced by lower fasting fat oxidation, impaired suppression of lipid oxidation, blunted stimulation of glucose oxidation and diminished increase in respiratory quotient (RQ) from fasting to insulin-stimulated states (16,22-24). Currently however, it is unknown if metabolic inflexibility is a feature of obesity per se, or a characteristic of individuals with MUO. Since youth with MUO, unlike their peers with MHO, have adverse metabolic characteristics, we hypothesized that metabolic inflexibility is distinctive of youth with MUO. Therefore, we aimed to investigate metabolic flexibility in youth with obesity characterized as having MHO and MUO, and examine their physical and metabolic characteristics.

2. Materials and Methods

2.1. Participants

Data from 188 youth that are overweight/affected by obesity without diabetes or PCOS, and 72 adolescents with normal weight (NW); ages 10 to <20 years; Tanner stage II-V, who participated in our National Institutes of Health-funded Childhood Metabolic Markers of Adult Morbidity in Blacks, and Childhood Insulin Resistance grants were used in the present analysis. Data resulting from these grants but unrelated to this manuscript have been reported (25-30). Twenty-one of these participants had overweight with body mass index (BMI) ≥85^th percentile for age and sex and 167 had obesity with BMI ≥95^th percentile. From here on we will refer to overweight and youth affected by obesity as “youth with obesity” for simplicity reasons. Of the recruited youth with obesity, there were 26 with impaired glucose tolerance (IGT) and 162 with normal glucose tolerance (NGT). Participants were recruited through newspaper advertisements, flyers posted in the outpatient clinics of the Weight Management and Wellness Center and the Division of Pediatric Endocrinology, city bus routes and the medical campus. The study was approved by the Institutional Review Board of the University of Pittsburgh and written informed parental consent and child assent were obtained from all participants before any research procedures were performed.

2.2. Procedures

All procedures were performed at the Pediatric Clinical and Translational Research Center (PCTRC) of the UPMC Children's Hospital of Pittsburgh. All participants provided a medical history and underwent a physical examination, and hematologic and biochemical tests. Height and weight were assessed to the nearest 0.1 cm and 0.1 kg, respectively, and used to calculate BMI. Pubertal development was assessed using Tanner criteria (31). Body composition was evaluated with dual energy X-ray absorptiometry (DEXA) for measurement of % body fat. Abdominal visceral adipose tissue (VAT) was assessed by either magnetic resonance imaging (MRI) or computed tomography (CT) at L4-5 intervertebral space (29,32). The switch from CT to MRI was imposed by the study section during the competitive grant renewal process. However, Klopfenstein et al. demonstrated strong correlation (r=0.89-0.95) and a good agreement (indicated by a minimal average difference) between CT and MRI for the measurement of abdominal adipose tissue (33).

2.3. Metabolic Studies

All participants were admitted to the PCTCR and underwent a hyperinsulinemic-euglycemic clamp combined with stable isotope tracers and continuous indirect calorimetry after a 10- to 12-hour overnight fast. Fasting blood samples were obtained for determination of lipid profile, adiponectin and HbA1c. Before the start of the hyperinsulinemic-euglycemic clamp, fasting hepatic glucose production was measured with a primed (2.2 μmol/kg) constant infusion of [6,6-²H₂]glucose at 0.22 μmol/kg/min for 2 hours as published (34). Whole-body lipolysis was measured at baseline and during the hyperinsulinemic-euglycemic clamp with a primed (1.2 umol/kg) constant-rate infusion of [²H₅]glycerol (MSD Isotopes, St. Louis, MO), which was started 2 hours before the clamp (35,36). Following the 2-hour baseline isotope infusion period, in vivo insulin sensitivity was assessed during a 3-hour hyperinsulinemic (80 mu/m²/min)-euglycemic (100 mg/dL) clamp. Continuous indirect calorimetry (Deltatrac Metabolic Monitor; SensorMedics, Anaheim, CA) was used to measure CO₂ production, O₂ consumption, and RQ for 30 min at baseline and at the end of the euglycemic clamp (13).

2.4. Biochemical Measurements

Plasma glucose was measured by the glucose oxidase method using a glucose analyzer (Yellow Springs Instrument Co., Yellow Springs, Ohio), insulin by commercially available radioimmunoassay (Millipore), HbA1c by high-performance liquid chromatography (Tosoh Medics) and adiponectin (ug/mL) by a radioimmunoassay kit (Linco Research) as previously reported (29,37). The intra- and inter- assay coefficients for adiponectin were 3.6 and 9.3% for low, and 1.8 and 9.3% respectively for high serum concentrations (37). Plasma lipid concentrations were determined using the standards of the Centers for Disease Control and Prevention (13). Deuterium enrichment of glucose in the plasma was determined on a Hewlett-Packard Co. 5971 mass spectrometer (Palo Alto, CA) coupled to a 5890 gas chromatograph (13). Deuterium enrichment of glycerol in the plasma was determined according to previously described methods (24).

2.5. Classification of MHO and MUO

Since insulin resistance is universally accepted to be the linchpin of the metabolic syndrome and its components (38), we chose to define metabolic health based on the cut points for in vivo insulin sensitivity of 72 healthy adolescents with NW (age 13.8 ± 1.8 yrs [SD]; BMI 19.6 ± 2.1 kg/m²). The in vivo insulin sensitivity was 10.49 ± 4.70 mg/kg/min per uU/mL in these participants with NW, and 2.89 ± 1.99 mg/kg/min per uU/mL in the total cohort with obesity. The insulin sensitivity of the group with MHO was chosen to be within 1.5 SDs of the mean of the participants with NW, and that of the group with MUO group to be <1.5 SDs of the mean insulin sensitivity of the subjects with NW (<3.44 mg/kg/min per uU/mL). Accordingly, of the 188 participants with obesity, 51 (27.1%) were classified as having MHO and 137 (72.9%) as MUO. These cut-off values for in vivo insulin sensitivity were chosen because the prevalence of the MHO obtained using these cut points matched the 21.5-31.5% range of MHO prevalence reported in 8-17 years old adolescents with obesity and a BMI above the 85^th percentile (5,9).

2.6. Calculations

Fasting endogenous glucose production was calculated during the last 30 min of the 2-hr isotope infusion according to steady-state tracer dilution equations (37). Hepatic insulin sensitivity was calculated as the inverse of the product of hepatic glucose production and fasting plasma insulin concentration (34). Adipose tissue insulin sensitivity was calculated as 1/(whole-body lipolysis rate x fasting insulin) as before (39). Peripheral insulin sensitivity was calculated by dividing the insulin-stimulated glucose disposal by the steady-state clamp insulin concentrations as reported by us (40). Metabolic flexibility was calculated as ΔRQ from fasting (baseline) to clamp steady-state.

2.7. Statistical Analysis

Univariate ANOVA using Bonferroni post-hoc test and chi-square analyses were used to compare physical and metabolic characteristics among the three groups (NW, MHO and MUO). ANCOVA was used to compare characteristics after adjusting for potential confounding effects of race, Tanner stage, glycemic status and % body fat differences among the three groups. Spearman’s correlation analysis was used to analyze bivariate relationships between metabolic flexibility (ΔRQ) and physical and metabolic parameters. Data that did not meet the assumptions for normality were log₁₀ transformed; untransformed data are presented for ease of interpretation. An additional analysis was performed by pair-matching 30 adolescents with MHO and 30 with MUO for age (± 2 yrs), sex (male/female), race (African American [AA]/American White [AW]), and BMI (± 2 kg/m²). Paired t-test was used to compare the pair-matched adolescents with MHO and MUO. Data were analyzed using SPSS 23.0 statistical software package with significance set at p-value <0.05. Data are mean ± SEM unless indicated otherwise.

Since there are no data on metabolic flexibility in youth with MHO vs. MUO, we used our two previous publications (15,24) to estimate the total sample size needed for the current study. Based on our data on metabolic flexibility (ΔRQ) in adolescent girls affected by obesity with vs. without polycystic ovary syndrome (pair-matched comparison of two metabolically different groups) (15), the calculated sample size for 80% and 90% power for detecting significant difference in ΔRQ was 16 and 21, respectively in each group. Additionally, another study with the unpair-matched design in 19 adolescents with MHO vs. 51 adolescents with MUO comparing leptin-adiponectin ratio (24), the calculated sample size was 18 (for MHO) and 48 (for MUO) to reach 80% and 24 (for MHO) and 64 (for MUO) to reach 90% power. Collectively, we elected to increase our sample size to avoid any chance of type 1 error in detecting a significant difference with respect to metabolic flexibility between the two groups in the present study.

3. Results

3.1. Total Cohort: Physical and Metabolic Characteristics

Physical and metabolic characteristics of the three groups are presented in Table 1. The three groups were similar in age and sex but the group with MUO had more AA youth, advanced Tanner stage, higher BMI, and more IGT than the other two groups. Among the three groups, % body fat was highest in youth with MUO and significantly higher than MHO (Table 1). While BMI percentile was significantly different among the three groups, it was not different between youth with MHO and MUO after adjusting for race, Tanner stage and glycemic status (Table 1). Similarly, VAT was significantly different among the three groups, but not different between adolescents with MHO and MUO after adjusting for race, Tanner stage, glycemic status and % body fat (Table 1). HbA1c was highest in youth with MUO but not different from MHO. Fasting glucose concentrations were not different among the three groups, nor between MHO and MUO, but fasting insulin concentrations and triglycerides were highest in youth with MUO and higher than MHO (Table 1). HDL was lowest while LDL was highest in adolescents with MUO, but not different from MHO. Adiponectin was lowest in MUO group and significantly lower than MHO (Table 1). ANCOVA analyses with adjustment for race, Tanner stage, glycemic status and % body fat did not change the three group significant differences except for VAT, fasting glucose and LDL (Table 1).

Table 1:

Physical and Metabolic Characteristics of Youth with NW, MHO and MUO.

	NW (1) n=72	MHO (2) n=51	MUO (3) n=137	P-ANOVA	P-ANCOVA	P, post hoc
						1 vs. 2	1 vs. 3	2 vs. 3
Age (yrs.)	13.8 ± 0.2	13.9 ± 0.3	14.1 ± 0.2	NS	NS	NS	NS	NS
Race (%, AA/AW)	46/54	47/53	62/38	0.039	-	-	-	-
Sex (%, M/F)	49/51	59/41	45/55	NS	-	-	-	-
Tanner stage (%, II/III/IV/V)	14/25/26/35	20/10/25/45	7/21/17/55	0.009	-	-	-	-
Glycemic status (%, NGT/IGT)	100/0	92/8	84/16	0.001	-	-	-	-
Physical Characteristicsa
BMI (kg/m2)	19.6 ± 0.2	29.0 ± 0.7	35.4 ± 0.5	<0.0001	<0.0001	<0.0001	<0.0001	<0.0001
BMI %	53.7 ± 2.6	95.4 ± 0.6	98.5 ± 0.1	<0.0001	<0.0001	<0.0001	0.001	NS
% body fat	20.6 ± 1.0	35.5 ± 1.3	43.5 ± 0.5	<0.0001	<0.0001	<0.0001	<0.0001	<0.0001
VAT (cm2)b	23.0 ± 1.7	45.4 ± 3.4	69.1 ± 3.0	<0.0001	NS	NS	NS	NS
Metabolic Characteristicsb
HbA1c (%)	5.27 ± 0.05	5.15 ± 0.06	5.41 ± 0.04	0.004	0.012	0.019	NS	NS
Fasting glucose (mg/dL)	95.3 ± 0.6	95.3 ± 0.8	96.5 ± 0.5	NS	0.023	NS	0.029	NS
Fasting insulin (uU/mL)	17.8 ± 0.9	20.2 ± 0.9	39.7 ± 1.6	<0.0001	<0.0001	NS	<0.0001	<0.0001
Total cholesterol (mg/dL)	151.7 ± 3.2	149.3 ± 4.6	157.6 ± 2.8	NS	NS	NS	NS	NS
Triglycerides (mg/dL)	70.7 ± 3.7	80.9 ± 3.6	102.3 ± 4.8	<0.0001	<0.0001	0.005	<0.0001	<0.0001
HDL (mg/dL)	51.4 ± 1.2	44.5 ± 1.2	40.7 ± 0.7	<0.0001	<0.0001	0.001	<0.0001	NS
LDL (mg/dL)	86.1 ± 2.7	88.7 ± 4.1	96.6 ± 2.4	0.018	NS	NS	NS	NS
VLDL (mg/dL)	14.1 ± 0.7	16.2 ± 0.7	20.0 ± 0.9	<0.0001	<0.0001	0.006	<0.0001	0.001
Adiponectin (ug/mL)	12.6 ± 0.7	10.1 ± 0.8	6.8 ± 0.3	<0.0001	<0.0001	0.008	<0.0001	<0.0001

Data are mean ± SEM. NW: normal weight; MHO: metabolically healthy obesity; MUO: metabolically unhealthy obesity; M: Male; F: Female; AA: African American; AW: American White; NGT: normal glucose tolerance; IGT: impaired glucose tolerance; BMI: body mass index; VAT: visceral adipose tissue; HDL: high density lipoprotein; LDL: low density lipoprotein; VLDL: very low density lipoprotein; NS: not significant. Note that Chi-square test is not subjective to provide post hoc results.

^aAdjusted for race, Tanner stage and glycemic status

^bAdjusted for race, Tanner stage, glycemic status and % body fat

Among the three groups, metabolic flexibility or ΔRQ, was lowest in MUO (Supplemental Table 1; see supplementary material), and significantly lower in MUO vs. MHO before and after adjustment for race, Tanner stage, glycemic status and % body fat (Figure 1, upper panel). However, metabolic flexibility was not different between youth with NW and MHO (Supplemental Table 1). Among the three groups, hepatic insulin sensitivity and adipose tissue insulin sensitivity were lowest in youth with MUO (Supplemental Table 1), and significantly lower in MUO vs. MHO before and after adjustment for confounders (Figures 2A and B, upper panel). By study design, peripheral insulin sensitivity was lowest in adolescents with MUO (Supplemental Table 1) and significantly lower than peers with MHO (Figure 2C, upper panel). Metabolic flexibility, hepatic insulin sensitivity and adipose tissue insulin sensitivity remained significantly different between groups with MHO and MUO after adjusting for race, Tanner stage, glycemic status, and % body fat. Additional adjustment for VAT did not change the significant differences between the two groups.

Figure 1:

Metabolic flexibility (ΔRQ) from fasting to clamp steady-state hyperinsulinemia of 51 youth with MHO and 137 with MUO (Upper panel), and of the pair-matched youth with MHO and MUO (Lower panel). Data are mean ± SEM. Upper panel *p is adjusted for race, Tanner stage, glycemic status and % body fat.

Figure 2:

Hepatic Insulin Sensitivity (A), Adipose tissue Insulin Sensitivity (B) and Peripheral Insulin Sensitivity (C) of 51 youth with MHO and 137 with MUO (Upper panel), and of the Pair-matched youth with MHO and MUO (Lower Panel). Data are mean ± SEM. Upper panel *p is adjusted for race, Tanner stage, glycemic status and % body fat.

3.2. Relationship of Metabolic Flexibility to Physical and Metabolic Parameters

In the total cohort with obesity, ΔRQ correlated positively with hepatic insulin sensitivity, adipose tissue insulin sensitivity, peripheral insulin sensitivity, adiponectin, and Tanner stage (p≤0.001 for all), and negatively with baseline RQ, triglycerides, total cholesterol, LDL, and VAT (p<0.0001 for all), but not with HDL (Supplemental Table 2). In youth with MUO, the correlations were similar to the total cohort with the added positive correlation of ΔRQ to HDL, while in youth with MHO the significant correlations were limited to baseline RQ, hepatic insulin sensitivity, triglycerides, and total cholesterol (Supplemental Table 2).

3.3. Pair-Matched Cohort

To avoid the potential confounding effects of differences in race, sex, Tanner stage and BMI (or BMI percentile) between youth with MHO and MUO, we pair-matched 30 adolescents with MHO and 30 with MUO with respect to age (± 2 yrs), sex (male/female), race (AA/AW), and BMI (± 2 kg/m²) as detailed under statistics above. Physical characteristics of the pair-matched cohort including age, race, sex, Tanner stage, BMI, BMI percentile, % body fat and VAT were not different between the pair-matched group with MHO and MUO (Table 2). Both fasting glucose and insulin concentrations were significantly greater in youth with MUO compared with their pair-matched peers with MHO (Table 2). Similar to our findings in the total cohort, youth with MUO had blunted ΔRQ in comparison with their peers with MHO (Figure 1, lower panel), and lower hepatic insulin sensitivity, adipose tissue insulin sensitivity and peripheral tissue insulin sensitivity (Figure 2, lower panel).

Table 2:

Physical and Metabolic Characteristics of the Pair-Matched Youth with MHO and MUO.

	MHO (n=30)	MUO (n=30)	P-value
Age (yrs.)	14.3 ± 0.4	14.1 ± 0.3	NS
Race (%, AA/AW)	57/43	57/43	NS
Sex (%, M/F)	63/37	63/37	NS
Tanner stage (%, II/III/IV/V)	13/10/23/53	7/30/10/53	NS
Glycemic status (%, NGT/IGT)	87/13	87/13	NS
Physical Characteristics
BMI (kg/m2)	31.2 ± 0.8	31.6 ± 0.8	NS
BMI %	97.6 ± 0.3	97.7 ± 0.4	NS
% body fat	38.0 ± 1.6	40.0 ± 1.4	NS
VAT (cm2)	51.7 ± 4.7	56.7 ± 4.9	NS
Metabolic Characteristics
HbA1c (%)	5.1 ± 0.1	5.3 ± 0.1	NS
Fasting glucose (mg/dL)	93.3 ± 0.7	96.4 ± 1.1	0.004
Fasting insulin (uU/mL)	20.8 ± 1.2	32.8 ± 2.5	<0.0001
Total cholesterol (mg/dL)	146.5 ± 5.8	152.7 ± 5.8	NS
Triglycerides (mg/dL)	80.1 ± 4.4	102.2 ± 9.1	0.055
HDL (mg/dL)	43.9 ± 1.6	41.8 ± 1.2	NS
LDL (mg/dL)	86.6 ± 5.1	90.5 ± 4.4	NS
VLDL (mg/dL)	16.0 ± 0.9	20.4 ± 1.8	0.055
Adiponectin (ug/mL)	9.2 ± 1.0	6.7 ± 0.5	NS

M: Male; F: Female; AA: African American; AW: American White; NGT: normal glucose tolerance; IGT: impaired glucose tolerance; BMI: body mass index; VAT: visceral adipose tissue; HDL: high density lipoprotein; LDL: low density lipoprotein; VLDL: very low density lipoprotein; NS: not significant.

4. Discussion

The present investigation reveals that (1) youth with MUO compared with MHO have metabolic inflexibility, i.e., blunted ΔRQ from fasting to insulin-stimulated states, with no difference in metabolic flexibility between youth with MHO and NW; (2) in the total and the pair-matched cohorts, youth with MUO compared with MHO have increased risk for type 2 diabetes manifested by lower hepatic and adipose tissue insulin sensitivity, and higher fasting glucose and insulin concentrations despite similar BMI percentile and VAT; and (3) there is a divergence in the relationship of metabolic flexibility to physical and metabolic characteristics between MUO and MHO. In youth with MUO the diminished ΔRQ correlates with global insulin resistance (hepatic, adipose and peripheral tissue) and with an atherogenic lipid profile, while the significant correlations in youth with MHO are far fewer.

In both adults and youth, it is widely recognized that obesity is strongly associated with reduced capacity to switch fuel oxidation to adjust for fuel availability, i.e., impaired metabolic flexibility or metabolic inflexibility (16,23,24,41). In line with this, metabolic inflexibility manifested by a blunted ΔRQ in the insulin-stimulated state is further associated with obesity-related comorbidities including insulin resistance, type 2 diabetes and cardiovascular disease risk (17-19,21,23,39,42). However, given the heterogeneity of obesity and the presence of cardiometabolic risk factors in some but not all individuals with obesity, it is questioned if metabolic inflexibility is a characteristic of obesity per se or it might be unique to a subgroup of individuals with unhealthy obesity. In order to probe this, we divided youth with obesity into those with MHO vs. MUO based on their clamp-measured peripheral insulin sensitivity cut-off derived from NW youth. Our results reveal that metabolic flexibility is significantly lower, ~20%, in youth with MUO compared with MHO youth, while metabolic flexibility did not differ in youth with MHO from that with NW. These observations suggest that metabolic inflexibility is a distinctive feature of MUO, and not obesity per se, because it is not globally present in all youth with obesity. Whether or not youth with MHO will develop metabolic inflexibility over time remains unknown. Furthermore, it needs to be investigated if the current observations in youth hold true for adults with obesity. Ultimately, it is critical to determine if youth with MUO and MHO require different obesity management strategies, such as a more concerted effort to enhance physical activity and consequently improve fat oxidation (43) besides nutritional counseling in MUO, in an effort to improve metabolic flexibility.

Our comparison of physical and metabolic characteristics between youth with MHO and MUO in the total cohort and in the pair-matched cohort confirms previous findings of worse cardiometabolic profile in MUO vs. MHO (12,13,44,45). Youth with MUO had worse lipid profile (higher triglycerides and VLDL) and higher fasting glucose and insulin concentrations compared with their peers with MHO, indicative of heightened cardiometabolic risk in youth with MUO consistent with previous findings in adults and youth (12,43,44). However, our findings further advance the field by demonstrating the presence of global insulin resistance in youth with MUO vs. MHO and their enhanced risk for youth type 2 diabetes. Both hepatic insulin sensitivity and adipose tissue insulin sensitivity, key pathophysiological components of type 2 diabetes (34,39,45) were significantly lower in MUO compared with MHO. Hepatic insulin sensitivity was ~40% lower and adipose tissue insulin sensitivity was ~60% lower. In the total cohort as well as the pair-matched cohort, despite similar BMI percentile and VAT, youth with MUO compared with MHO demonstrate resistance to the antilipolytic effect of insulin, while adipose tissue insulin sensitivity of youth with MHO is similar to that of peers with NW. Collectively, it could be postulated that altered adipose tissue function plays a significant role in differentiating youth with MUO from MHO, and could contribute not only to metabolic inflexibility but also be a biomarker of enhanced risk for type 2 diabetes.

Another uniqueness of the present study is the investigation of the relationship of metabolic flexibility with metabolic parameters separately in each of the groups with MUO and MHO. In fact, our correlation analysis in the total cohort showing that ΔRQ correlates negatively with baseline RQ and fasting lipids, and positively with insulin sensitivity and adiponectin, is concordant with previous findings in girls affected by obesity with PCOS (24). However, our data suggest that most of those relationships are driven by the characteristics of MUO. The separate analyses in the groups with MHO and MUO show that while both groups share common correlations between metabolic flexibility and Tanner staging, baseline RQ, Triglycerides, total cholesterol and hepatic insulin sensitivity, the associations appear to be much broader in youth with MUO including adipose tissue and peripheral insulin sensitivity, adiponectin, HDL, LDL and VAT (Supplemental Table 2). The observation that Tanner staging correlates positively with metabolic flexibility suggests that metabolic flexibility might change with pubertal maturation. Thus, it may be critical to match for pubertal development while studying adolescents with obesity with respect to metabolic flexibility. The wider associations observed in MUO with respect to metabolic flexibility could be reflective of their amplified risks for atherogenesis and diabetogenesis closely linked to metabolic inflexibility, which might not be the case yet in MHO but might evolve over time. On the other hand, one could theorize that any intervention aiming to improve metabolic flexibility might be more effective in a less severely impacted group such as MHO than MUO. Given that exercise increases both peripheral and hepatic insulin sensitivity in youth with obesity (47-50), it would be germane to examine the effects of different exercise modalities in improving metabolic flexibility in youth with MUO separate from MHO and assess the potential conversion from MUO to MHO phenotype.

The strengths of the present investigation include: (1) a first-time comparison of metabolic flexibility or ΔRQ between youth with NW vs. MUO vs. MHO in order to determine if metabolic inflexibility is a feature of obesity per se or is unique to MUO; (2) the comprehensive assessments from physical to state-of-the-art metabolic tests for a thorough characterization of MUO vs. MHO; (3) a rigorous matching between youth with MUO and MHO with respect to age, sex, race and BMI to confirm our total cohort analyses; and (4) the additional analyses of metabolic flexibility and its relationships separately in each MUO and MHO groups to gain insight into uniqueness of each group. Potential limitations would be that a cut-off point for defining MHO vs. MUO is not a universal criterion. However, we used a robust measure of in vivo insulin sensitivity in peers with NW, and assured that the (27%) prevalence of MHO in our total cohort was consistent with the range of MHO prevalence (21.5-31.5%) reported in 8-17 year old adolescents with obesity (5,9). Another limitation would be that our findings may not be representative of other ethnic groups as we focused on a balanced representation of African-American and American-White youth. Further, we could not pair-match exactly by Tanner staging; however, we did not include Tanner stage I and only included adolescents with Tanner staging II-V, and each group had 53 adolescents who were fully pubertal at Tanner V, and ultimately there was no statistically significant difference in Tanner stages between the two groups (p=0.153, Table 2). With respect to race-specific analysis, all of our analyses were statistically adjusted for race. Lastly, the nature of the current cross-sectional design hinders any conclusions about the development of MUO from MHO, and the chronology of metabolic inflexibility. It remains to be determined if over time youth with MHO will develop deteriorating physical and metabolic characteristics similar to MUO, paralleled with evolving metabolic inflexibility.

5. Conclusions

Our data suggest that metabolic inflexibility is unique to a subgroup of youth with obesity, i.e., MUO, as metabolic flexibility (ΔRQ) is lowest in youth who have MUO but is similar between peers with MHO and NW. Until further advancement in our understanding of the heterogeneity of obesity and associated cardiometabolic risks, caution should be practiced against uniformly characterizing all youth with obesity as having metabolic inflexibility without sub-grouping them. Furthermore, youth with MUO compared with MHO have evidence of global insulin resistance and worse cardiometabolic profile enhancing their risk for both diabetes and atherogenesis. Our observations of broader associations between metabolic flexibility and insulin sensitivity and cardiometabolic risks in MUO suggest a more advanced stage of obesity severity than in those with MHO. Lastly, longitudinal studies are warranted to examine if over time youth with MHO will develop metabolic inflexibility alongside worsening of cardiometabolic phenotypes akin to youth with MUO.

Abbreviations:

AA

African American

AW

American White

BMI

body mass index

CT

computed tomography

DEXA

dual energy X-ray absorptiometry

IGT

impaired glucose tolerance

MRI

magnetic resonance imaging

MHO

metabolically healthy obesity

MUO

metabolically unhealthy obesity

NGT

normal glucose tolerance

NAFLD

nonalcoholic fatty liver disease

NW

normal weight

PCOS

polycystic ovary syndrome

PCTRC

Pediatric Clinical and Translational Research Center

RQ

respiratory quotient

VAT

visceral adipose tissue