Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 May 27.
Published in final edited form as: Obesity (Silver Spring). 2021 Nov 24;30(1):39–44. doi: 10.1002/oby.23312

Insulin resistance persists despite a metabolically healthy obesity phenotype

Kristin K Hoddy 1, Christopher L Axelrod 1, Jacob T Mey 1, Adithya Hari 2, Robbie A Beyl 1, Jourdan B Blair 1, Wagner S Dantas 1, John P Kirwan 1
PMCID: PMC9136885  NIHMSID: NIHMS1807725  PMID: 34816598

Abstract

Objective:

Metabolically healthy obesity (MHO) is often defined as the absence of metabolic syndrome in the presence of obesity. However, phenotypic features of MHO are unclear. Insulin sensitivity in MHO was cross-sectionally compared with metabolically unhealthy obesity (MUO) and a reference group of young healthy participants without obesity.

Methods:

Sedentary adults (n = 96) undergoing anthropometric, blood chemistries, maximal aerobic capacity, and euglycemic-hyperinsulinemic clamp measurements were classified by BMI (<25 and ≥30 kg/m2). MUO was defined as having obesity with metabolic syndrome (≥2 additional risk factors). Data were analyzed using a linear mixed models approach.

Results:

Body weight was similar between MHO and MUO. Body fat (percentage) and high-density lipoprotein cholesterol were higher (p < 0.001), and systolic blood pressure, triglycerides, glucose, and insulin were lower in MHO versus MUO (p < 0.03, all). The MHO group also had lower high-density lipoprotein cholesterol and higher low-density lipoprotein cholesterol, diastolic blood pressure, and insulin compared with the reference. Both the MHO and MUO groups displayed impaired insulin sensitivity compared with the reference control (p < 0.001).

Conclusions:

Participants with MHO had distinct clinical measures related to hyper-tension, lipid metabolism, and glycemic control compared with a healthy reference group. Peripheral insulin resistance in obesity independent of metabolic status portends increased risk for type 2 diabetes in the MHO patient population.

INTRODUCTION

The concept of metabolically healthy obesity (MHO) has become a controversial term within obesity medicine. The construct was introduced several decades ago (1) and was used to identify a subset of people with obesity who are at a lower risk for chronic disease compared with counterparts with metabolically unhealthy obesity (MUO) (2). Improved glycemic control (3), higher cardiorespiratory fitness (4), and less sedentary behavior (4) have been observed in MHO compared with MUO. Some longitudinal observations indicate a high conversion of MHO to MUO over time (5,6). Other reports using stricter criteria suggest that the MHO phenotype is rather stable (7). Nevertheless, the MHO phenotype appears to carry a heightened health risk compared with healthy participants with a normal body weight (7–9) and it may represent an intermediate phenotype within the spectrum of obesity.

Although there are several existing definitions of MHO (2), the MHO phenotype is most commonly characterized as obesity without metabolic syndrome (MetS). MetS encompasses several clinical facets of cardiometabolic health, including glycemic control, lipid metabolism, and endothelial function (10). Concern regarding the ability for cardiometabolic risk factor clusters, such as MetS, to identify insulin resistance was raised in the general population without diabetes (11). Considering the recent emphasis on cardiometabolic care for patients with obesity and type 2 diabetes, it is important to examine the utility of MetS criteria to differentiate MHO from MUO. Our objective was to first use MetS criteria to phenotype participants with obesity as having MHO or MUO and then compare indicators of cardiometabolic health and insulin sensitivity against a young healthy reference group without obesity.

METHODS

Human participants

This was a retrospective analysis of procedures conducted from 2000 to 2018 (NCT02697201) (12–14). Participants (n = 96) were weight stable (>6 months) and washed-out of antihypertensive medication prior to testing. We excluded those with a diagnosis or history of disease (i.e., history of heart, kidney, liver, thyroid, intestinal, and pulmonary diseases) and medication use known to alter insulin sensitivity, lipids, appetite, mood, and hormonal concentrations. Resting 12-lead electrocardiograms (ECG) and submaximal exercise stress tests excluded individuals with exercise contraindications. Menopausal women were excluded. All studies were subject to internal review, and participants provided written informed consent.

Inpatient control period

All study participants completed a standardized 3-day inpatient stay in a Clinical Research Unit and they were allowed to leave during the day (12–14). Maximal aerobic capacity (VO2MAX) was determined using a modified Bruce protocol (1800 hours) 2 days preceding all metabolic testing. On the evening prior to testing, all participants were provided a standardized meal (55% carbohydrate, 35% fat, and 10% protein) based on individual energy needs. Body composition was assessed after an overnight fast via dual-energy x-ray absorptiometry or underwater weighing. Participants were then discharged and instructed to abstain from structured exercise. After returning at 1800 hours, they were fed a similar standardized meal. A euglycemic-hyperinsulinemic clamp (90 mg/dL, 40 mU·m−2·min−1) was performed the following morning (0600 hours) with simultaneous infusion of insulin (constant) and 20% dextrose (variable) initiated at 0 minutes. Blood samples were drawn at 5-minute intervals. Glucose was measured immediately after each draw (YSI 2300, STAT Plus), and the glucose infusion rate was adjusted using the DeFronzo et al. correction algorithm (15). Steady-state insulin-stimulated whole-body glucose disposal (glucose disposal rate [GDR], in milligrams per kilogram per minute) and the quantity of glucose metabolized per unit of plasma insulin concentration (GDR/I; microunits per milliliter) were determined during the 90- to 120-minute time range of the clamp. Euglycemic-hyperinsulinemic clamp outcomes were expressed relative to total and fat free mass (FFM).

Statistical analysis

BMI-defined weight classification and metabolic status were determined using criteria established by the International Diabetes Foundation for MetS (10) (Figure 1). MHO participants presented with obesity and less than two additional MetS risk factors, whereas MUO participants had obesity and two or more additional MetS risk factors. Age and gender were identified as covariates and were adjusted using linear mixed models (SAS version 9.4). Post hoc t tests based on least squares means were analyzed, and Pearson correlations were performed to determine relationships between obesity metrics and glycemic outcomes. The significance level was set at p < 0.05.

FIGURE 1.

FIGURE 1

Open in a new tab

Participant classification. Those who previously completed metabolic inpatient stays were classified by BMI (<25 and ≥30 kg/ m2). Participants with a BMI ≥30 kg/m2 were identified as having MUO if they presented with two or more metabolic syndrome risk factor criteria, including hypertension (>130 mmHg systolic or > 85 mmHg diastolic), hyperglycemia (fasting glucose ≥100 mg/dL), triglycerides (≥150 mg/dL), and HDL cholesterol (<40 mg/dL [men] or <50 mg/dL [women]). There was no overall statistical difference observed for biological sex (p = 0.694). MHO, metabolically healthy obesity; MUO, metabolically unhealthy obesity

RESULTS

Participant characteristics and clinical measures are provided in Table 1. Participants with MHO and MUO were older than the young healthy reference group; thus, subsequent comparisons were age adjusted. Body weight and BMI were similar between MHO and MUO and elevated compared with the reference. Groups with obesity had increased body fat with MHO participants reporting the greatest proportion of body fat.

TABLE 1.

Clinical characteristics of participants

REF (N = 19) MHO (N = 26) vs. REF MUO (N = 51) vs. REF vs. MHO
Mean (SEM) Mean (SEM) p value Mean (SEM) p value p value
Age (y)a 33.1 (2.8) 59 (2.4) <0.001 60.3 (1.7) <0.001 NS
Body weight (kg) 61.5 (3.7) 98.4 (2.5) <0.001 104.0 (1.9) NS <0.001
BMI (kg/m2) 22.6 (1.1) 34.4 (0.8) <0.001 35.7 (0.6) NS <0.001
Body fat (%)b 28.9 (1.4) 46.2 (1.0) <0.001 41.3 (0.8) <0.001 <0.001
Fasting glucose (mg/dL) 86.8 (6.1) 91.2 (4.1) NS 106.8 (3.1) 0.003 0.009
Fasting insulin (μU/mL)c 5.8 (3.7) 15.4 (2.5) 0.042 22.6 (1.9) 0.020 <0.001
Systolic blood pressure (mmHg)d 119.4 (4.2) 125.6 (3.0) NS 134.3 (2.3) 0.012 0.005
Diastolic blood pressure (mmHg)d 67.3 (2.6) 78.5 (1.8) 0.001 83.4 (1.4) 0.026 <0.001
Total cholesterol (mg/dL)e 186.1 (8.9) 188.2 (6.1) NS 184.5 (4.7) NS NS
LDL cholesterol (mg/dL) 54.9 (9.5) 117.8 (6.5) <0.001 111.7 (5.0) NS <0.001
HDL cholesterol (mg/dL) 68.2 (3.3) 51.8 (2.3) <0.001 36.6 (1.7) <0.001 <0.001
Triglycerides (mg/dL)e 88.8 (17.7) 114.9 (12.2) NS 162.7 (9.1) 0.002 <0.001
Open in a new tab

Estimates derived from least squares means using linear models adjusted for age and sex.

Abbreviations: HDL, high-density lipoprotein; LDL, low-density lipoprotein; MHO, metabolically healthy obesity; MUO, metabolically unhealthy; REF, healthy reference group.

a

Variable not adjusted for age or gender.

b

n = 91.

c

n = 92.

d

n = 90.

e

n = 95.

Clinical measures highlight MHO as an intermediate phenotype between the reference and MUO. Systolic blood pressure was similar between MHO and the reference, whereas MUO values were clinically remarkable (>130 mmHg) and statistically higher compared with MHO and the reference. Diastolic blood pressure was lowest in the reference and progressively increased in MHO and MUO, respectively. VO2MAX was similar between MHO and MUO (p = 0.525), but both MHO (24.0 ± 0.9 mL/kg/min) and MUO (23.2 ± 0.8 mL/kg/min) groups were lower compared with the reference (40.4 ± 1.3 mL/kg/min; p < 0.001, all). All groups presented with similar total cholesterol. Low-density lipoprotein (LDL) cholesterol was similar in MHO and MUO and values were higher in both groups with obesity compared with the reference. MUO participants had lower high-density lipoprotein (HDL) cholesterol compared with MHO, and HDL cholesterol was also lower in both MHO and MUO compared with the reference. Triglyceride concentrations were similar between MHO and the reference and highest in MUO. MHO participants maintained normal fasting glucose values (<100 mg/dL) that were similar to the reference. However, normoglycemia appears to have been maintained at the expense of higher fasting insulin in MHO compared with the reference. Glucose and insulin values were highest in MUO versus both MHO and the reference. Across the entire cohort, body weight correlated with glucose (r = 0.22; p = 0.039) and insulin (r = 0.35; p < 0.001) concentrations. Additionally, percent body fat correlated with insulin concentrations (r = 0.26; p = 0.015).

The euglycemic-hyperinsulinemic clamp findings are presented in Figure 2. Both obesity phenotypes displayed lower GDR relative to total mass compared with the reference (5.7 ± 0.37 mg/kg/min; p < 0.001, all) with similar responses noted between MHO (2.00 ± 0.26 mg/kg/min) and MUO (1.52 ± 0.19 mg/kg/min; p = 0.124). GDR relative to total mass correlated significantly with body weight (r = −0.65; p < 0.001) and percent body fat (r = −0.57; p < 0.001). Aligned with this, GDR/I relative to total mass was markedly lower in MHO (0.0298 ± 0.0067 mg/kg/min/μU/mL) and MUO (0.0242 ± 0.0049 mg/kg/min/μU/mL) compared with the reference (0.0857 ± 0.010 mg/kg/min/μU/mL; p > 0.001). There was no difference between MHO and MUO for whole-body GDR/I (p = 0.490). Body weight (r = −0.47; p < 0.001) and percent body fat (r = −0.50; p < 0.001) associated with GDR/I relative to total mass across phenotypes. Steady-state insulin concentrations were similar between the reference (104.13 ± 15 μU/mL), MHO (93.5 ± 11 μU/mL), and MUO (80.6 ± 8 μU/mL; p ≥ 0.207, all).

FIGURE 2.

FIGURE 2

Open in a new tab

Comparison of whole body insulin sensitivity relative to total mass in each phenotype. Comparison of insulin sensitivity relative to total body mass in the healthy reference group (REF), MHO, and MUO. (A) Insulin-stimulated glucose disposal (GDR) (n = 94) and (B) insulin sensitivity relative to prevailing insulin concentrations (GDR/I) (n = 90). Data are displayed as a box (mean ± 95% CI) and whisker (minimum to maximum) plot. ****Denotes a significant (p < 0.001) difference relative to the healthy reference group. MHO, metabolically healthy obesity; MUO, metabolically unhealthy obesity

GDR and GDR/I relative to FFM values aligned with our findings for total mass. GDR relative to FFM was similar between MHO (4.21 ± 1.05 mg/kgFFM/min) and MUO (3.51 ± 0.81 mg/kgFFM/min; p = 0.585), and both GDR values were significantly diminished compared with the reference (22.21 ± 1.5 mg/kgFFM/min; p < 0.001, all). Likewise, GDR/I relative to FFM was lower in MHO (0.0691 ± 0.0365 mg/kgFFM/min/μU/mL) and MUO (0.0674 ± 0.0273 mg/kgFFM/min/μU/mL) compared with the reference (0.3642 ± 0.0517 mg/kgFFM/min/μU/mL; p < 0.001, all) but was similar to one another (p = 0.969).

DISCUSSION

The data presented herein demonstrate the presence of peripheral insulin resistance in obesity independent of clinically determined metabolic status. Participants with MHO in the present analysis were clinically healthier compared with MUO. However, our comparison of MHO with a young healthy reference group depicted clear clinical and subclinical cardiometabolic deficits. These abnormalities suggest cardiometabolic risk in the MHO phenotype and corroborates examples in the literature demonstrating higher chronic disease risk (7–9).

Similar analyses have demonstrated better lipid profiles (16–19) and glycemic outcomes (16,18) in participants with MHO compared with MUO. Several investigations have used percentile cutoffs derived from euglycemic-hyperinsulinemic clamps to define metabolic health status in obesity (3,16–19). Notably, many of the comparable studies with euglycemic-hyperinsulinemic clamp outcomes to evaluate MHO and MUO phenotypes have been in postmenopausal women (3,16–18). Marini et al. (3) defined MHO at the 75th percentile and MUO to be within the bottom 50th percentile for clamp-derived insulin sensitivity and included a lean healthy reference group in their analysis. Even though they noted similar insulin sensitivity between their lean healthy reference and MHO phenotypes, MHO participants still presented with intermediate clinical outcomes between MUO and their healthy reference (3). Messier et al. (18) alternatively considered clinical criteria to identify MHO and MUO phenotypes. Lipid concentrations and blood pressure values were healthier in MHO compared with MUO when MHO was defined as obesity with no more than one additional clinical cardiometabolic abnormality. Contrary to our findings, their MHO phenotype presented with higher insulin sensitivity than their MUO phenotype (18). It is possible that stricter criteria and the authors’ inclusion of high-sensitivity c-reactive protein as an additional clinical marker were able to identify a healthier sample of people with MHO. A lean healthy reference was not included in the report by Messier et al., thus making it difficult to determine whether MHO was an intermediate phenotype.

Our data depict the relationship between increasing body weight and percent body fat with decreasing insulin sensitivity. This aligns with the clinical practice of grading obesity severity and cardiometabolic risk, by proxy, according to BMI. Nevertheless, peripheral insulin resistance occurs early in the onset of obesity (20), and clinical tools are needed to detect early signs of chronic disease without obesity stigma. The Edmonton Obesity Staging System (EOSS) is one example of a clinical grading tool with aspects of cardiometabolic health and quality of life (21). Although the EOSS is considerably better than BMI alone, it still relies heavily on clinical disease presentation to inform therapeutic decisions rather than objective assessment of biomarkers present prior to chronic disease onset. As such, there is a missed opportunity for early intervention in patients with MHO who present with early indicators of cardiometabolic dysfunction. More granular staging of cardiometabolic status has been shown to be an attractive option to identify risk in this subset of patients with obesity (22,23).

We chose to use MetS to differentiate MHO from MUO because of its clinical utility and widespread use (2). Our findings, however, suggest that MetS criteria may lack sensitivity to detect subclinical presentation of chronic disease biomarkers. Our interpretation of the present findings aligns with seminal work depicting insulin resistance, assessed via euglycemic-hyperinsulinemic clamp, in 30% of people without MetS defined by criteria from the Third Report of the National Cholesterol Education Program Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (ATP III) (11). MHO presentation has been shown to vary by the identification approach. As an example, clamp-derived identification of MHO and MUO by Messier et al. (18) demonstrated similar HDL cholesterol and blood pressure values between the two phenotypes, with a higher aerobic capacity noted in those with MHO (18). In the same group of participants, clinical identification of MHO revealed no difference in aerobic capacity but indicated better HDL cholesterol and blood pressures in the MHO phenotype (18). Likewise, slightly different clinical identification criteria also altered cardiometabolic characteristics of the MHO phenotype (18). These examples collectively underscore the limitations of MetS criteria for MHO identification. Stricter criteria that excludes all clinical cardiometabolic abnormalities could be used to standardize MHO criteria. Yet this approach may not account for subclinical cardiometabolic deficits as those with MHO and free of clinical cardiometabolic abnormalities were shown to still be at a greater risk for chronic disease compared with those with a healthy body weight and clinical cardiometabolic status (7).

A major strength of this retrospective analysis is that the selected studies used gold-standard, robust measurement techniques in a carefully controlled, supervised inpatient setting. Confounders were minimized by excluding participants on medication and those with clinical disease while only including sedentary participants who were healthy enough for intense exercise. Although these efforts minimized influence from disease and behavioral factors, this approach could obstruct full representation of the obesity spectrum. Our reference group was significantly younger than the groups with obesity, and we adjusted for this statistically. Future comparisons should examine age-matched phenotypes. Despite this present limitation, our reference group represents optimal metabolic health and allows for important interpretations that would not be possible if only MHO and MUO phenotypes were considered. Lastly, glucose disposal rates were not adjusted for hepatic glucose production, which may have varied between phenotypes.

CONCLUSION

Our data underscore the presence of insulin resistance in both MHO and MUO determined by MetS. In the context of supporting literature, cardiometabolic status in patients with MHO as defined by absence of MetS does not parallel a healthy state without obesity. Scientifically valid, clinically relevant, and universally implemented tools to identify cardiometabolic status are necessary to advance physiological understanding of obesity progression and inform precise therapies.O

Study Importance.

What is already known?

  • Obesity is a known risk factor for cardiometabolic disease.

  • Obesity is a heterogenous condition, and a subset appears to be metabolically healthy.

What does this study add?

  • Clinical identification of metabolically healthy obesity (MHO) using metabolic syndrome criteria shows insulin resistance exists in this seemingly healthy phenotype.

  • Experimental and common clinical assessments indicate early signs of cardiometabolic risk in MHO and suggest that MHO may be an intermediate condition between a young healthy reference group without obesity and metabolically unhealthy obesity.

How might these results change the direction of research or the focus of clinical practice?

  • Standardization of MHO, informed by reliable and validated biomarkers that incorporate markers of insulin sensitivity, is necessary for integration in clinical practice.

  • Routine clinical assessments to detect early indicators of diminished insulin sensitivity stand to advance obesity treatment.

  • Participants with MHO may benefit from insulin-sensitizing interventions, but long-term clinical trials are needed to establish the potential benefits of precise obesity care.

ACKNOWLEDGMENTS

We thank the study volunteers and staff for their considerable time and effort on this project.

Funding information

This research was supported by NIH grants U54 GM104940 (JPK), R01 AG012834 (JPK), and R01 DK108089 (JPK). JTM was supported by T32 AT004094, and NIH National Center for Research Resources 1UL1 RR024989, Cleveland, Ohio.

Footnotes

CONFLICT OF INTEREST

At the time of publication, KKH will be employed by Abbott Laboratories. Abbott was not involved in design, analysis, interpretation, or funding of the present work. The other authors declared no conflict of interest.

REFERENCES

  • 1.Vague J The degree of masculine differentiation of obesities: a factor determining predisposition to diabetes, atherosclerosis, gout, and uric calculous disease. Am J Clin Nutr. 1956;4:20–34. [DOI] [PubMed] [Google Scholar]
  • 2.Smith GI, Mittendorfer B, Klein S. Metabolically healthy obesity: facts and fantasies. J Clin Invest. 2019;129:3978–3989. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Marini MA, Succurro E, Frontoni S, et al. Metabolically healthy but obese women have an intermediate cardiovascular risk profile between healthy nonobese women and obese insulin-resistant women. Diabetes Care. 2007;30:2145–2147. [DOI] [PubMed] [Google Scholar]
  • 4.Ortega FB, Cadenas-Sanchez C, Migueles JH, et al. Role of physical activity and fitness in the characterization and prognosis of the metabolically healthy obesity phenotype: a systematic review and meta-analysis. Prog Cardiovasc Dis. 2018;61:190–205. [DOI] [PubMed] [Google Scholar]
  • 5.Mongraw-Chaffin M, Foster MC, Anderson CAM, et al. Metabolically healthy obesity, transition to metabolic syndrome, and cardiovascular risk. J Am Coll Cardiol. 2018;71:1857–1865. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Kouvari M, Panagiotakos DB, Yannakoulia M, et al. Transition from metabolically benign to metabolically unhealthy obesity and 10-year cardiovascular disease incidence: the ATTICA cohort study. Metabolism. 2019;93:18–24. [DOI] [PubMed] [Google Scholar]
  • 7.Guo F, Garvey WT. Cardiometabolic disease risk in metabolically healthy and unhealthy obesity: stability of metabolic health status in adults. Obesity (Silver Spring). 2016;24:516–525. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Fan J, Song Y, Chen Y, Hui R, Zhang W. Combined effect of obesity and cardio-metabolic abnormality on the risk of cardiovascular disease: a meta-analysis of prospective cohort studies. Int J Cardiol. 2013;168:4761–4768. [DOI] [PubMed] [Google Scholar]
  • 9.Tajik S, Mirzababaei A, Ghaedi E, Kord-Varkaneh H, Mirzaei K. Risk of type 2 diabetes in metabolically healthy people in different categories of body mass index: an updated network meta-analysis of prospective cohort studies. J Cardiovasc Thorac Res. 2019;11:254–263. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Huang PL. A comprehensive definition for metabolic syndrome. Dis Mod Mech. 2009;2:231–237. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Liao Y, Kwon S, Shaughnessy S, et al. Critical evaluation of adult treatment panel III criteria in identifying insulin resistance with dyslipidemia. Diabetes Care. 2004;27:978–983. [DOI] [PubMed] [Google Scholar]
  • 12.Solomon TPJ, Sistrun SN, Krishnan RK, et al. Exercise and diet enhance fat oxidation and reduce insulin resistance in older obese adults. J Appl Physiol. 2008;104:1313–1319. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Solomon TPJ, Haus JM, Kelly KR, et al. A low–glycemic index diet combined with exercise reduces insulin resistance, postprandial hyperinsulinemia, and glucose-dependent insulinotropic polypeptide responses in obese, prediabetic humans. Am J Clin Nutr. 2010;92:1359–1368. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Fealy CE, Haus JM, Solomon TPJ, et al. Short-term exercise reduces markers of hepatocyte apoptosis in nonalcoholic fatty liver disease. J Appl Physiol. 2012;113:1–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.DeFronzo RA, Tobin JD, Andres R. Glucose clamp technique: a method for quantifying insulin secretion and resistance. Am J Physiol. 1979;237:E214–E223. [DOI] [PubMed] [Google Scholar]
  • 16.Brochu M, Tchernof A, Dionne IJ, et al. What are the physical characteristics associated with a normal metabolic profile despite a high level of obesity in postmenopausal women? J Clin Endocrinol Metab. 2001;86:1020–1025. [DOI] [PubMed] [Google Scholar]
  • 17.Karelis AD, Faraj M, Bastard J-P, et al. The metabolically healthy but obese individual presents a favorable inflammation profile. J Clin Endocrinol Metab. 2005;90:4145–4150. [DOI] [PubMed] [Google Scholar]
  • 18.Messier V, Karelis AD, Prud’homme D, Primeau V, Brochu M, Rabasa-Lhoret R. Identifying metabolically healthy but obese individuals in sedentary postmenopausal women. Obesity (Silver Spring). 2010;18:911–917. [DOI] [PubMed] [Google Scholar]
  • 19.Owei I, Umekwe N, Provo C, Wan J, Dagogo-Jack S. Insulin-sensitive and insulin-resistant obese and non-obese phenotypes: role in prediction of incident pre-diabetes in a longitudinal biracial cohort. BMJ Open Diabetes Res Care. 2017;5:e000415. 10.1136/bmjdrc-2017-000415 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Kahn SE, Hull RL, Utzschneider KM. Mechanisms linking obesity to insulin resistance and type 2 diabetes. Nature. 2006;444:840–846. [DOI] [PubMed] [Google Scholar]
  • 21.Padwal RS, Pajewski NM, Allison DB, Sharma AM. Using the Edmonton obesity staging system to predict mortality in a population-representative cohort of people with overweight and obesity. Can Med Assoc J. 2011;183:E1059–E1066. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Guo F, Moellering DR, Garvey WT. The progression of cardiometabolic disease: validation of a new cardiometabolic disease staging system applicable to obesity. Obesity (Silver Spring). 2014;22:110–118. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Mechanick JI, Garber AJ, Grunberger G, Handelsman Y, Garvey WT. Dysglycemia-based chronic disease: an American Association of Clinical Endocrinologist position statement. Endocr Pract. 2018;24:995–1011. [DOI] [PubMed] [Google Scholar]

RESOURCES