Liver

Subject Category: Liver

Am J Gastroenterol 2016; 111:1133–1140; doi:10.1038/ajg.2016.178; published online 17 May 2016

Metabolically Healthy Obesity and the Development of Nonalcoholic Fatty Liver Disease

Yoosoo Chang MD, PhD1,2,3,12, Hyun-Suk Jung MD1,12, Juhee Cho PhD1,3,4, Yiyi Zhang MHS, PhD4,5,6, Kyung Eun Yun MD, PhD1, Mariana Lazo MD, PhD4,5,6, Roberto Pastor-Barriuso PhD7, Jiin Ahn MSPH1, Chan-Won Kim MD1, Sanjay Rampal MD, PhD4,6,8, Miguel Cainzos-Achirica MD4,6,9, Di Zhao4,5,6, Eun Cheol Chung MD, PhD10, Hocheol Shin MD, PhD11, Eliseo Guallar MD, DrPH4,5,6 and Seungho Ryu MD, PhD1,2,3

  1. 1Center for Cohort Studies, Total Healthcare Center, Kangbuk Samsung Hospital, Sungkyunkwan University School of Medicine, Seoul, Republic of Korea
  2. 2Department of Occupational and Environmental Medicine, Kangbuk Samsung Hospital, Sungkyunkwan University School of Medicine, Seoul, Republic of Korea
  3. 3Department of Health Sciences and Technology, Samsung Advanced Institute for Health Sciences and Technology, Sungkyunkwan University, Seoul, Republic of Korea
  4. 4Department of Epidemiology, Johns Hopkins University Bloomberg School of Public Health, Baltimore, Maryland, USA
  5. 5Department of Medicine, Johns Hopkins School of Public Health, Baltimore, Maryland, USA
  6. 6Welch Center for Prevention, Epidemiology, and Clinical Research, Johns Hopkins University Bloomberg School of Public Health, Baltimore, Maryland, USA
  7. 7National Center for Epidemiology, Carlos III Institute of Health and Consortium for Biomedical Research in Epidemiology and Public Health (CIBERESP), Madrid, Spain
  8. 8Department of Social and Preventive Medicine, Julius Centre University of Malaya, Faculty of Medicine, University of Malaya, Kuala Lumpur, Malaysia
  9. 9Ciccarone Center for the Prevention of Heart Disease, Department of Cardiology, Johns Hopkins Medical Institutions, Baltimore, Maryland, USA
  10. 10Department of Radiology, Kangbuk Samsung Hospital, Sungkyunkwan University School of Medicine, Seoul, South Korea
  11. 11Department of Family Medicine, Kangbuk Samsung Hospital, Sungkyunkwan University School of Medicine, Seoul, South Korea

Correspondence: Eliseo Guallar, MD, DrPH, Welch Center for Prevention, Epidemiology, and Clinical Research, Johns Hopkins Bloomberg School of Public Health, 2024 E. Monument Street, Room 2-645, Baltimore, Maryland 21205, USA. E-mail: eguallar@jhu.edu; Seungho Ryu, MD, PhD, Kangbuk Samsung Hospital, Sungkyunkwan University School of Medicine, Samsung Main Building B2, 250, Taepyung-ro 2ga, Jung-gu, Seoul 100-742, Korea. E-mail: sh703.yoo@gmail.com

12The first two authors contributed equally to this work as co-first authors of this paper.

Received 3 February 2016; Accepted 13 April 2016
Advance online publication 17 May 2016

Top

Abstract

OBJECTIVES:

 

The risk of nonalcoholic fatty liver disease (NAFLD) among obese individuals without obesity-related metabolic abnormalities, a condition referred to as metabolically healthy obese (MHO), is largely unexplored. Therefore, we examined the association between body mass index (BMI) categories and the development of NAFLD in a large cohort of metabolically healthy men and women.

METHODS:

 

A cohort study was conducted in 77,425 men and women free of NAFLD and metabolic abnormalities at baseline, who were followed-up annually or biennially for an average of 4.5 years. Being metabolically healthy was defined as not having any metabolic syndrome component and having a homeostasis model assessment of insulin resistance <2.5. The presence of fatty liver was determined using ultrasound.

RESULTS:

 

During 348,193.5 person-years of follow-up, 10,340 participants developed NAFLD (incidence rate, 29.7 per 1,000 person-years). The multivariable adjusted hazard ratios (95% confidence intervals) for incident NAFLD comparing overweight and obese with normal-weight participants were 2.15 (2.06–2.26) and 3.55 (3.37–3.74), respectively. In detailed dose–response analyses, increasing baseline BMI showed a strong and approximately linear relationship with the incidence of NAFLD, with no threshold at no risk. This association was present in both men and women, although it was stronger in women (P for interaction <0.001), and it was evident in all clinically relevant subgroups evaluated, including participants with low inflammation status.

CONCLUSIONS:

 

In a large cohort of strictly defined metabolically healthy men and women, overweight and obesity were strongly and progressively associated with an increased incidence of NAFLD, suggesting that the obese phenotype per se, regardless of metabolic abnormalities, can increase the risk of NAFLD.

Top

INTRODUCTION

Nonalcoholic fatty liver disease (NAFLD) is the most common cause of chronic liver disease worldwide (1) and an emerging risk factor for type 2 diabetes mellitus, cardiovascular disease, and all-cause mortality (2, 3, 4). NAFLD may have a key role in the metabolic abnormalities associated with excess adiposity and is even considered a precursor of metabolic syndrome (5, 6).

While excess adiposity is a major determinant of the high prevalence of NAFLD worldwide (6, 7, 8, 9, 10), the connection between overweight/obesity and NAFLD is still incompletely understood. Specifically, it is unclear whether obesity is a risk factor for NALFD in subjects who do not have any of the metabolic abnormalities associated with excess adiposity, a group often described as metabolically healthy obese (MHO) (11, 12). The health implications of MHO are controversial (13, 14, 15). Only one study has evaluated the association of MHO with NAFLD (16). In this study, MHO subjects had an increased prevalence of NAFLD compared with metabolically healthy non-obese subjects, but the cross-sectional design of this study limited its ability to establish a temporal relation between obesity and NAFLD.

No cohort study has evaluated the role of MHO as a determinant of incident NAFLD among subjects free of NAFLD at baseline. Therefore, we examined the association between body mass index (BMI) categories and the development of NAFLD in a large cohort of metabolically healthy men and women who participated in a health screening examination program.

Top

METHODS

Study population

The Kangbuk Samsung Health Study is a cohort study of men and women 18 years of age or older who underwent a comprehensive annual or biennial health examination at the clinics of the Kangbuk Samsung Hospital Total Healthcare Screening Center in Seoul and Suwon, South Korea, between 1 January 2002 and 31 December 2013. More than 80% of participants were employees of various companies and local governmental organizations and their spouses. In South Korea, the Industrial Safety and Health Law requires annual or biennial health screening exams of all employees, offered free of charge. The remaining participants voluntarily purchased screening exams at the health exam center.

ADVERTISEMENT

The present analysis included all study participants with at least one follow-up visit through 31 December 2013 (n=233,676). We selected metabolically healthy participants by excluding any participant with at least one of the following metabolic abnormalities at baseline (11, 17): (i) fasting blood glucose ≥100mg/dl or current use of blood glucose-lowering agents (n=49,634) (18); (ii) blood pressure ≥130/85mmHg or current use of blood pressure-lowering agents (n=48,596) (18); (iii) triglyceride levels ≥150mg/dl or current use of lipid-lowering agents (n=58,629) (18); (iv) high-density lipoprotein-cholesterol (HDL-C) <40mg/dl in men or <50mg/dl in women (n=32,880) (18); or (v) insulin resistance defined as homeostasis model assessment of insulin resistance (HOMA-IR) ≥2.5 (n=40,559) (19).

We further excluded participants who had any of the following conditions at baseline: fatty liver on ultrasound (n=60,522); history of malignancy (n=2,484); history of known liver disease (n=30,380); alcohol intake ≥30g/day for men and ≥20g/day for women (n=20,506); (1) positive serologic markers for hepatitis B or C virus (n=9,585); or use within the past 1 year of medications that could induce fatty liver such as amiodarone, tamoxifen, methotrexate, or corticosteroids (n=720) (1). Finally, we excluded participants with missing data for anthropometric measures, metabolic parameters, or liver ultrasonography (n=6,331). As some individuals met more than one criterion for exclusion, the total number of metabolically healthy subjects without NAFLD at baseline included in the study was 77,425 (30,502 men and 46,923 women; Figure 1).

Figure 1.
Figure 1 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Flow chart of study participants.

Full figure and legend (133K)

The study was approved by the Institutional Review Board of the Kangbuk Samsung Hospital, which waived the requirement for informed consent as we used only de-identified data obtained as part of routine health-screening exams.

Data collection

All examinations were conducted at Kangbuk Samsung Hospital Total Healthcare Screening Center clinics in Seoul and Suwon. At each clinic visit, demographic characteristics, smoking status, alcohol consumption, regular exercise, medical history, and medication use were collected by standardized, self-administered questionnaires. Smoking status was categorized into never, former, and current. Alcohol consumption was categorized into none, ≤10g/day, and >10g/day. We also assessed the weekly frequency of moderate- or vigorous-intensity physical activity.

Height and weight were measured by trained nurses with the participants wearing a lightweight hospital gown and no shoes. Height was measured to the nearest 1mm using a stadiometer with the participant standing barefoot. Weight was measured to the nearest 0.1kg on a bioimpedance analyzer (InBody 3.0 and Inbody 720, Biospace Co., Seoul, Korea) that was validated with regards to reproducibility and accuracy for body composition (20) and calibrated every morning before the testing started. BMI was calculated as weight in kilograms divided by height in meters squared. Blood pressure was measured by trained nurses while subjects were in a sitting position with the arm supported at heart level.

Laboratory analyses

Blood specimens were sampled from the antecubital vein after at least a 10-h fast. Methods for measuring serum levels of glucose, uric acid, total cholesterol, low-density lipoprotein cholesterol (LDL-C), triglycerides, HDL-C, aspartate aminotransferase, alanine aminotransferase, gamma-glutamyltransferase, insulin, and high-sensitivity C-reactive protein (hsCRP) have been reported elsewhere (21, 22). The Laboratory Medicine Department of the Kangbuk Samsung Hospital is accredited by the Korean Society of Laboratory Medicine (KSLM) and the Korean Association of Quality Assurance for Clinical Laboratories (KAQACL), and participates in the College of American Pathologists (CAP) Survey Proficiency Testing. Insulin resistance was assessed with the HOMA-IR equation: fasting blood insulin (μU/ml) × fasting blood glucose (mmol/l)/22.5.

Abdominal ultrasound for NAFLD assessment

Abdominal ultrasounds were performed using a Logic Q700 MR 3.5-MHz transducer (GE, Milwaukee, WI) by 11 experienced radiologists unaware of the study aims. Images were captured in a standard manner with the patient in the supine position with the right arm raised above the head. An ultrasonographic diagnosis of fatty liver was defined as the presence of a diffuse increase of fine echoes in the liver parenchyma compared with the kidney or spleen parenchyma (23). The inter- and intra-observer reliability for the diagnosis of fatty liver was very high (kappa statistics of 0.74 and 0.94, respectively) (24). Because we had already excluded participants with excessive alcohol use (≥20g/day for women and ≥30g/day for men) (1) as well as other identifiable causes of fatty liver at baseline as described in the exclusion criteria, incident cases of fatty liver were considered NAFLD.

Statistical analysis

BMI was categorized on the basis of criteria established for Asian populations (25, 26): underweight, BMI <18.5kg/m2; normal-weight, BMI 18.5–23kg/m2; overweight, BMI 23–25kg/m2; and obesity, BMI ≥25kg/m2. The primary endpoint was the development of incident NAFLD. Follow-up for each participant extended from the baseline exam until the development of NAFLD or the last health exam conducted prior to 31 December 2013, whichever came first. Incidence rates were calculated as the number of incident cases divided by person-years of follow-up.

Because we started follow-up with all participants free of NAFLD at baseline, we could establish the first visit at which a participant showed NAFLD, but we could not determine the precise time of outcome development (which occurred at some point between the first visit with NAFLD and the previous visit). As a consequence, we used a parametric proportional hazards model to take into account this type of interval censoring (stpm command in Stata) (27). In these models, the baseline hazard function was parameterized with restricted cubic splines in log time with four degrees of freedom.

The primary analysis was based on clinical BMI categories, and we estimated adjusted hazard ratios (aHRs) with 95% confidence intervals (CIs) for incident NAFLD comparing BMI categories at baseline with the normal-weight category. To determine the linear trends of risk, we used a continuous variable with the category number and tested its statistical significance in the regression models. Models were initially adjusted initially for age, sex, center, and year of examination, and then further adjusted for potential confounding factors including education level, smoking status, alcohol intake and exercise. We assessed the proportional hazards assumption by examining graphs of estimated log (−log) survival.

We conducted two additional types of dose–response analyses. First, we estimated the aHRs with 95% CIs associated with a BMI increase of 1kg/m2 at baseline using BMI as a continuous variable in the regression models. Second, we modeled BMI using restricted quadratic splines with knots at the 5th, 50th, and 95th percentiles of the baseline distribution to provide a smooth yet flexible description of the relationship between BMI and incident NAFLD.

We also conducted two additional types of analyses. First, we conducted subgroup analyses to identify interactions between BMI categories and clinically relevant groups defined by age (<40 vs. ≥40 years), sex (women vs. men), smoking (current vs. noncurrent smokers), alcohol intake (<10 vs. ≥10g/day), regular exercise (<3 vs. ≥3 times/week), and hsCRP levels (<1.0 vs. ≥1.0mg/l). Second, we evaluated potential mediation by intermediate physiological variables by including baseline systolic blood pressure, fasting serum glucose, triglycerides, HDL-C, LDL-C, HOMA-IR, and hsCRP in the regression models. All analyses were performed using STATA version 13.0 (StataCorp LP, College Station, TX).

Top

RESULTS

The mean (s.d.) age and BMI of study participants at baseline was 35.7 (6.4) years and 21.5 (2.4) kg/m2 (BMI range 12.6–36.1kg/m2), respectively. Participants in higher BMI categories were more likely to be older, male, and current smokers, were more likely to exercise and to drink alcohol, and have higher levels of fasting glucose, systolic and diastolic blood pressure, total cholesterol, triglycerides, LDL-C, uric acid, hepatic enzymes, insulin, HOMA-IR, and hsCRP and lower levels of HDL-C than normal-weight or underweight participants (Table 1).


We identified 10,340 incident cases of NAFLD during 348,193.5 person-years of follow-up (incident rate 29.7 per 1,000 person-years). The average follow-up period for participants was 4.5 years. In this group of metabolically healthy men and women, increasing baseline BMI showed a strong and graded dose–response relationship with the incidence of NAFLD (Table 2). In models adjusted for age, sex, center, and year of visit, the aHRs (95% CIs) for incident NAFLD comparing underweight, overweight, and obese participants with normal-weight participants were 0.28 (0.24–0.33), 2.12 (2.03–2.22), and 3.47 (3.30–3.66), respectively. In spline regression models, the dose–response relationship between baseline BMI and incident NAFLD was approximately linear throughout the range of BMI levels, with no threshold for risk (Figure 2). The aHR associated with a 1kg/m2 increase when BMI was introduced as a continuous variable in regression models was 1.29 (1.28–1.30). The association was virtually unchanged after adjustment for smoking, alcohol intake, exercise, and education.

Figure 2.
Figure 2 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Multivariable-adjusted hazard ratios for incident nonalcoholic fatty liver disease by level of body mass index. Curves represent adjusted hazard ratios for incident nonalcoholic fatty liver disease based on restricted quadratic splines with knots at the 5th, 50th, and 95th percentiles of the body mass index distribution. The reference value (diamond dot) was set at 18.5kg/m2 (corresponding to the 10th percentile in the overall population, 14th percentile in women, and 4th percentile in men). Models were adjusted for age, sex (overall analysis only), study center, year of visit, education, smoking, alcohol, and physical activity.

Full figure and legend (50K)


The association between BMI and the incidence of NAFLD was observed in both men and women, but it was significantly stronger in women (P for interaction <0.001; Table 3) although the absolute incidence rate of NAFLD was much lower in women than in men (15.5 vs. 50.6 per 1,000 person-years, respectively). None of the other interactions tested was statistically significant (not shown). The association between BMI and incident NAFLD was evident even among the 66,642 participants with low hsCRP levels (<1.0mg/l) at baseline. In this subgroup, the aHRs (95% CIs) for NAFLD comparing underweight, overweight, and obese participants to normal-weight participants were 0.28 (0.24–0.33), 2.16 (2.05–2.28), and 3.41 (3.20–3.62), respectively.


In order to explore whether the increased risk of NAFLD associated with increased BMI levels was mediated by residual metabolic components in MHO participants, we performed additional analyses adjusted for metabolic risk factors. Adjusting for fasting blood glucose, systolic blood pressure, triglyceride levels, HDL-C, HOMA-IR, hsCRP, and LDL-C slightly reduced the associations but they remained strong and statistically significant (Appendix Table A1).

Top

DISCUSSION

In this large cohort study of metabolically healthy men and women without fatty liver at baseline, overweight and obesity were associated with an increased risk of incident NAFLD compared with normal weight. The association between BMI and incident NAFLD was strong and progressive, and was observed across the whole range of BMI values and in all subgroups evaluated. Furthermore, this association could not be explained by residual levels of metabolic risk factors below the standard cutoffs used to define metabolic abnormalities. Our findings indicate that excess adiposity is not a harmless condition and can induce the development of NAFLD even in subjects with no metabolic abnormalities.

A major strength of our study is that we used a strict definition of metabolic health to focus on a large cohort of men and women with no metabolic abnormalities and with no insulin resistance (11). Although there is no general agreement on the criteria to define MHO (13, 15), other studies have typically focused on subjects with two or fewer metabolic components, but the risk of cardiovascular disease increases progressively with the number of metabolic syndrome components, beginning with one component (28, 29). Indeed, each metabolic syndrome criterion is an established cardiovascular risk factor (30) and has been associated with increased risk of diabetes (31). Even though we used very strict criteria to define metabolically healthy participants using both metabolic syndrome criteria and insulin resistance, we still found a strong association between BMI and incident NAFLD that persisted after adjusting for metabolic components and after excluding participants with increased levels of inflammatory markers. As a consequence, our findings support that obesity per se, irrespective of insulin resistance or metabolic abnormalities, is a crucial element in the development of NAFLD.

To our knowledge, this is the first cohort study to evaluate the association between the MHO phenotype and the risk of developing NAFLD. Although obesity, insulin resistance, and metabolic abnormalities are well-known risk factors for NAFLD (32, 33), the effect of excess adiposity in the absence of metabolic abnormalities on the risk of NAFLD has rarely been reported in previous studies. A previous cross-sectional study showed a higher prevalence of fatty liver among MHO individuals, but this study did not consider other risk factors for fatty liver such as excessive alcohol consumption, or compare obese with non-obese participants (including overweight and underweight participants in the reference group) (16). Furthermore, the cross-sectional design resulted in temporal ambiguity of exposure and outcome.

The mechanisms whereby obesity contributes to NAFLD remain incompletely elucidated. Obesity is characterized by expanded adipose tissue that can increase free fatty acid delivery to the liver, possibly contributing to lipid accumulation in the liver (7). Furthermore, adipose tissue is an active endocrine organ that produces and releases adipokines with a number of pro-inflammatory and other potentially harmful effects (34). Adipocyte-derived hormones and cytokines including leptin, adiponectin, interleukin-6, and tumor necrosis factor-alpha may be involved in the pathogenesis of NAFLD (7). Indeed, the expression of inflammatory genes was similarly altered in both MHO and metabolically unhealthy obese subjects, highlighting that MHO individuals may exhibit an adverse profile comparable with that of metabolically unhealthy obese patients and that the MHO concept should be applied with caution (35). Finally, recent studies suggest that the MHO phenotype is transient and progresses to overt metabolic abnormalities in a significant proportion of individuals (36, 37), although the association between BMI and incident NAFLD in our study was evident across all BMI levels and was not restricted to MHO participants. Further mechanistic studies, however, are needed to further understand why the MHO phenotype is associated with an increased incidence of NAFLD.

In the current study, the association between BMI and incident NAFLD was observed in both men and women but was stronger in women, even though the absolute incidence of NAFLD was over three times higher in men. Previous population-based studies also reported a significantly higher prevalence of NAFLD in men than in women (38), possibly related to more favorable lifestyle factors and the protective effect of sex hormones in women. The reasons for a stronger association of BMI with incident NAFLD in women are unclear, but excess adiposity may be a more important contributor to fatty liver in women than in men. Further research is needed to understand this sex-related difference.

Several limitations of our study need to be considered. First, our study used ultrasonography to detect incident fatty liver as the study endpoint: we did not perform histological confirmation or use more accurate measures such as MR spectroscopy to detect incident fatty liver, as these methods are not feasible in large population-based epidemiological studies (1, 39). Abdominal ultrasound, however, has a high diagnostic accuracy for steatosis (40) and many population-based studies have relied on ultrasound assessment of NAFLD (39, 40, 41). Second, the definition of insulin resistance used in this study was based on HOMA-IR levels, not on invasive and time consuming euglycemic insulin clamp analyses. HOMA-IR and euglycemic insulin clamp data are strongly correlated (42), but we cannot discard some misclassification of insulin resistance status in our study. Third, we used BMI as a measure of obesity, but BMI does not distinguish fat tissue from lean tissue. If the MHO group had a higher proportion of lean mass compared with normal-weight participants, the association between MHO and incident NAFLD in our study could be attenuated. In addition, waist circumference measurements were available for only a small fraction of study participants, limiting our ability to examine the role of fat distribution on the development of NAFLD. Fourth, we were unable to include dietary information, which could confound the association between BMI and NAFLD. Fifth, duration of obesity has been shown to be an independent risk factor for adverse health outcomes (43, 44), but this information was not available in this study. Finally, our study was conducted in asymptomatic relatively young Korean men and women and our findings may not apply to other populations.

In addition to the large sample size that allowed us to select a group of participants with no metabolic abnormalities or insulin resistance at baseline, other strengths of our study included the prospective design, the use of carefully standardized clinical, imaging, and laboratory procedures, and the availability of detailed phenotyping of study participants. The relatively young age of our study cohort is also an advantage of our study, as findings in this group are less likely to be affected by survivor bias, biases induced by comorbidities, or use of multiple medications than findings in older cohorts.

In conclusion, overweight and obesity were strongly and progressively associated with an increased incidence of NAFLD in a strictly selected group of metabolically healthy men and women. These findings provide compelling evidence to support to the hypothesis that MHO is not a harmless condition and that the obese phenotype per se, regardless of metabolic abnormalities, can increase development of NAFLD. As a consequence, physicians should adequately address the increased risk of NAFLD and of other cardiovascular and metabolic abnormalities in metabolically healthy obese individuals in addition to counseling them about healthy weight and lifestyle.

Top

Study Highlights

Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Top

Conflict of interest

Guarantors of the article: Seungho Ryu, MD, PhD and Eliseo Guallar, MD, DrPH.

Specific author contributions: Y. Chang, H.S. Jung, S. Ryu, and E. Guallar planned and designed the study, and directed its implementation, including quality assurance and control. Y. Chang and S. Ryu analyzed the data and designed the study’s analytic strategy. S. Ryu, J. Cho, E.C. Chung, and H. Shin helped supervise the field activities. Y. Chang, H.S. Jung, S. Ryu, E. Guallar, Y. Zhang, K.E. Yun, M. Lazo, R. Pastor-Barriuso, J. Ahn, C.W. Kim, M. Cainzos-Achirica, D. Zhao, B.S. Suh, S. Rampal, and J. Cho helped conduct the literature review and prepare the Materials and Methods and the Discussion sections of the text. Y. Chang, H.S. Jung, S. Ryu, and E. Guallar drafted the manuscript. Y. Chang, H. Jung, J. Cho, Y. Zhang, K.E. Yun, M. Lazo, R. Pastor-Barriuso, J. Ahn, C.W. Kim, S. Rampal, M. Cainzos-Achirica, D. Zhao, B.S. Suh, E.C. Chung, B.S. Suh, H. Shin, E. Guallar, and S. Ryu interpreted the results. All authors contributed to critical revision of the manuscript.

Financial support: None.

Potential competing interests: None.

Top

References

  1. Chalasani N, Younossi Z, Lavine JE et al. The diagnosis and management of non-alcoholic fatty liver disease: practice guideline by the American Gastroenterological Association, American Association for the Study of Liver Diseases, and American College of Gastroenterology. Gastroenterology 2012;142:1592–609. | Article | PubMed | ISI |
  2. Marchesini G, Marzocchi R, Agostini F et al. Nonalcoholic fatty liver disease and the metabolic syndrome. Curr Opin Lipidol 2005;16:421–7. | Article | PubMed | CAS |
  3. Musso G, Gambino R, Cassader M et al. Meta-analysis: natural history of non-alcoholic fatty liver disease (NAFLD) and diagnostic accuracy of non-invasive tests for liver disease severity. Ann Med 2011;43:617–49. | Article | PubMed | ISI |
  4. Anstee QM, Targher G, Day CP. Progression of NAFLD to diabetes mellitus, cardiovascular disease or cirrhosis. Nat Rev Gastroenterol Hepatol 2013;10:330–44. | Article | PubMed | ISI | CAS |
  5. Lonardo A, Ballestri S, Marchesini G et al. Nonalcoholic fatty liver disease: a precursor of the metabolic syndrome. Dig Liver Dis 2015;47:181–90. | Article | PubMed |
  6. Yki-Jarvinen H. Non-alcoholic fatty liver disease as a cause and a consequence of metabolic syndrome. Lancet Diabetes Endocrinol 2014;2:901–10. | Article | PubMed | CAS |
  7. Fabbrini E, Sullivan S, Klein S. Obesity and nonalcoholic fatty liver disease: biochemical, metabolic, and clinical implications. Hepatology 2010;51:679–89. | Article | PubMed | ISI | CAS |
  8. Chang Y, Ryu S, Sung E et al. Weight gain within the normal weight range predicts ultrasonographically detected fatty liver in healthy Korean men. Gut 2009;58:1419–25. | Article | PubMed | ISI | CAS |
  9. Marchesini G, Moscatiello S, Di Domizio S et al. Obesity-associated liver disease. J Clin Endocrinol Metab 2008;93:S74–80. | Article | PubMed |
  10. Vernon G, Baranova A, Younossi ZM. Systematic review: the epidemiology and natural history of non-alcoholic fatty liver disease and non-alcoholic steatohepatitis in adults. Aliment Pharmacol Ther 2011;34:274–85. | Article | PubMed | ISI | CAS |
  11. Primeau V, Coderre L, Karelis AD et al. Characterizing the profile of obese patients who are metabolically healthy. Int J Obes (Lond) 2011;35:971–81. | Article | PubMed | CAS |
  12. Stefan N, Kantartzis K, Machann J et al. Identification and characterization of metabolically benign obesity in humans. Arch Intern Med 2008;168:1609–16. | Article | PubMed | ISI |
  13. Fruhbeck G, Toplak H, Woodward E et al. Need for a paradigm shift in adult overweight and obesity management—an EASO position statement on a pressing public health, clinical and scientific challenge in Europe. Obes Facts 2014;7:408–16. | Article | PubMed |
  14. Puri R. Is it finally time to dispel the concept of metabolically-healthy obesity? J Am Coll Cardiol 2014;63:2687–88. | Article | PubMed |
  15. Kramer CK, Zinman B, Retnakaran R. Are metabolically healthy overweight and obesity benign conditions?: A systematic review and meta-analysis. Ann Intern Med 2013;159:758–69. | Article | PubMed | ISI |
  16. Sung KC, Cha SC, Sung JW et al. Metabolically healthy obese subjects are at risk of fatty liver but not of pre-clinical atherosclerosis. Nutr Metab Cardiovasc Dis 2014;24:256–62. | Article | PubMed |
  17. Kuk JL, Ardern CI. Are metabolically normal but obese individuals at lower risk for all-cause mortality? Diabetes Care 2009;32:2297–99. | Article | PubMed | ISI |
  18. Alberti KG, Eckel RH, Grundy SM et al. Harmonizing the metabolic syndrome: a joint interim statement of the International Diabetes Federation Task Force on Epidemiology and Prevention; National Heart, Lung, and Blood Institute; American Heart Association; World Heart Federation; International Atherosclerosis Society; and International Association for the Study of Obesity. Circulation 2009;120:1640–45. | Article | PubMed | ISI | CAS |
  19. Matthews DR, Hosker JP, Rudenski AS et al. Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia 1985;28:412–9. | Article | PubMed | ISI | CAS |
  20. Malavolti M, Mussi C, Poli M et al. Cross-calibration of eight-polar bioelectrical impedance analysis versus dual-energy X-ray absorptiometry for the assessment of total and appendicular body composition in healthy subjects aged 21-82 years. Ann Hum Biol 2003;30:380–91. | Article | PubMed | ISI | CAS |
  21. Zhang Y, Chang Y, Ryu S et al. Thyroid hormone levels and incident chronic kidney disease in euthyroid individuals: the Kangbuk Samsung Health Study. Int J Epidemiol 2014;43:1624–32. | Article | PubMed |
  22. Chang Y, Kim BK, Yun KE et al. Metabolically-healthy obesity and coronary artery calcification. J Am Coll Cardiol 2014;63:2679–86. | Article | PubMed |
  23. Mathiesen UL, Franzen LE, Aselius H et al. Increased liver echogenicity at ultrasound examination reflects degree of steatosis but not of fibrosis in asymptomatic patients with mild/moderate abnormalities of liver transaminases. Dig Liver Dis 2002;34:516–22. | Article | PubMed | ISI |
  24. Ryu S, Chang Y, Choi Y et al. Age at menarche and non-alcoholic fatty liver disease. J Hepatol 2015;62:1164–70. | Article | PubMed |
  25. WHO Western Pacific Region IaI. The Asia-Pacific Perspective: Redefining Obesity and Its Treatment. Health Communications Australia Pty. Ltd. Sydney: Australia. 2000.
  26. Wen CP, David Cheng TY, Tsai SP et al. Are Asians at greater mortality risks for being overweight than Caucasians? Redefining obesity for Asians. Public Health Nutr 2009;12:497–506. | Article | PubMed | ISI |
  27. Royston P, Parmar MK. Flexible parametric proportional-hazards and proportional-odds models for censored survival data, with application to prognostic modelling and estimation of treatment effects. Stat Med 2002;21:2175–97. | Article | PubMed | ISI |
  28. Ho JS, Cannaday JJ, Barlow CE et al. Relation of the number of metabolic syndrome risk factors with all-cause and cardiovascular mortality. Am J Cardiol 2008;102:689–92. | Article | PubMed | ISI |
  29. Tsai SP, Wen CP, Chan HT et al. The effects of pre-disease risk factors within metabolic syndrome on all-cause and cardiovascular disease mortality. Diabetes Res Clin Pract 2008;82:148–56. | Article | PubMed | ISI |
  30. Kahn R, Buse J, Ferrannini E et al. The metabolic syndrome: time for a critical appraisal: joint statement from the American Diabetes Association and the European Association for the Study of Diabetes. Diabetes Care 2005;28:2289–304. | Article | PubMed | ISI |
  31. Wilson PW, D'Agostino RB, Parise H et al. Metabolic syndrome as a precursor of cardiovascular disease and type 2 diabetes mellitus. Circulation 2005;112:3066–72. | Article | PubMed | ISI | CAS |
  32. Bellentani S, Saccoccio G, Masutti F et al. Prevalence of and risk factors for hepatic steatosis in Northern Italy. Ann Intern Med 2000;132:112–7. | Article | PubMed | ISI | CAS |
  33. Marchesini G, Bugianesi E, Forlani G et al. Nonalcoholic fatty liver, steatohepatitis, and the metabolic syndrome. Hepatology 2003;37:917–23. | Article | PubMed | ISI |
  34. Ronti T, Lupattelli G, Mannarino E. The endocrine function of adipose tissue: an update. Clin Endocrinol (Oxf) 2006;64:355–65. | PubMed | CAS |
  35. Gomez-Ambrosi J, Catalan V, Rodriguez A et al. Increased cardiometabolic risk factors and inflammation in adipose tissue in obese subjects classified as metabolically healthy. Diabetes Care 2014;37:2813–21. | Article | PubMed | CAS |
  36. Chang Y, Ryu S, Suh BS et al. Impact of BMI on the incidence of metabolic abnormalities in metabolically healthy men. Int J Obes (Lond) 2012;36:1187–94. | Article | PubMed |
  37. Soriguer F, Gutierrez-Repiso C, Rubio-Martin E et al. Metabolically healthy but obese, a matter of time? Findings from the prospective Pizarra study. J Clin Endocrinol Metab 2013;98:2318–25. | Article | PubMed | ISI |
  38. Pan JJ, Fallon MB. Gender and racial differences in nonalcoholic fatty liver disease. World J Hepatol 2014;6:274–83. | Article | PubMed |
  39. Clark JM, Diehl AM. Defining nonalcoholic fatty liver disease: implications for epidemiologic studies. Gastroenterology 2003;124:248–50. | Article | PubMed |
  40. Hernaez R, Lazo M, Bonekamp S et al. Diagnostic accuracy and reliability of ultrasonography for the detection of fatty liver: a meta-analysis. Hepatology 2011;54:1082–90. | Article | PubMed | ISI |
  41. Lazo M, Hernaez R, Eberhardt MS et al. Prevalence of nonalcoholic fatty liver disease in the United States: the Third National Health and Nutrition Examination Survey, 1988-1994. Am J Epidemiol 2013;178:38–45. | Article | PubMed | ISI |
  42. Katz A, Nambi SS, Mather K et al. Quantitative insulin sensitivity check index: a simple, accurate method for assessing insulin sensitivity in humans. J Clin Endocrinol Metab 2000;85:2402–10. | Article | PubMed | ISI | CAS |
  43. Abdullah A, Wolfe R, Stoelwinder JU et al. The number of years lived with obesity and the risk of all-cause and cause-specific mortality. Int J Epidemiol 2011;40:985–96. | Article | PubMed | ISI |
  44. Abdullah A, Stoelwinder J, Shortreed S et al. The duration of obesity and the risk of type 2 diabetes. Public Health Nutr 2011;14:119–26. | Article | PubMed | ISI |
Top

Appendices

Appendix

Mediation analysis of the association between body mass index and the development of nonalcoholic fatty liver disease

Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author