Metabolically Healthy Versus Unhealthy Obesity and Risk of Fibrosis Progression in Non-alcoholic Fatty Liver Disease

Yejin Kim; Yoosoo Chang; Yong Kyun Cho; Jiin Ahn; Hocheol Shin; Seungho Ryu


Liver International. 2019;39(10):1884-1894. 

In This Article


In this large cohort study of 59 957 young- and middle-aged Korean men and women with NAFLD, the risk of fibrosis progression based on noninvasive fibrosis markers increased across BMI categories regardless of metabolic health status. This dose-dependent positive association was similarly observed when the associations of body fat percentage and waist circumference with fibrosis progression were additionally analysed. To the best of our knowledge, this is by far the largest cohort study evaluating the association of BMI, body fat percentage and waist circumference with fibrosis progression according to metabolic health status. We found that the associations persisted after further adjustments for hsCRP and metabolic parameters, and even in individuals who remained metabolically healthy over follow-up, indicating that overweight and obesity per se can contribute to fibrosis progression in NAFLD independent of metabolic health status.

While several previous studies showed that individuals with MHO are at increased risk of NAFLD, the impact of MHO on fibrosis progression in NAFLD is little known.[9,22] A growing body of evidence suggests that MHO is a transient state in which individuals initially defined as metabolically healthy may progress to unhealthy status.[23] In fact, studies with longer follow-up (6-10 years) reported that approximately 50%-75% of the subjects with MHO became metabolically unhealthy and the progression rate steadily increased with the length of follow-up.[23–25] In our study, approximately 70% of metabolically healthy individuals at baseline became metabolically unhealthy during median follow-up of 7.7 years. When we considered the changes in components of metabolic health over follow-up as time-varying covariates, the increased risk of fibrosis marker worsening was consistently observed in higher than normal BMI, waist circumference and body fat percentage categories, suggesting that metabolic health status is indeed not static, and healthy metabolic profile does not guarantee the favourable prognosis in obese individuals with NAFLD.

In our sensitivity analysis, after excluding metabolically healthy individuals at baseline who transitioned to metabolically unhealthy status over follow-up, the risk of fibrosis marker worsening was similarly elevated with increasing BMI category as observed in our primary analysis, further underscoring the possibility that being overweight and obese per se may predict worse outcomes in NAFLD patients. In light of this, all overweight and obese individuals with NAFLD, regardless of metabolic status, should be carefully monitored and given appropriate management for reducing fibrosis progression.

The mechanism by which excess adiposity or obesity per se in the absence of metabolic abnormality affects fibrosis progression remains unclear. Obesity is characterized by enlarged adipose tissue, which promotes the release of free fatty acid into the circulation.[26] The toxic by-products of the surplus free fatty acid metabolism as a result of excess adipose tissues are thought to contribute to the development of fibrosis.[27] Also, adipose tissue is recognized as an active endocrine organ, which produces various secretory products called adipokines.[28] The several adipocytokines, including adiponectin and leptin, appear to be implicated in the pathogenesis of NAFLD such as accumulation of liver fat, insulin resistance, inflammatory processes and development of liver fibrosis.[29,30] Leptin increases transforming growth factor β1 which is an important profibrogenic cytokine in liver fibrosis.[31] Further study is needed to elucidate the exact mechanism as well as the relationship between MHO with fibrosis progression in NAFLD.

In the sex-stratified analysis, the association between adiposity parameters and fibrosis progression was stronger in men compared to women. Although the sex difference in NAFLD progression is not well understood due to relatively scarce research findings on this issue, studies suggest that the protective role of oestrogen in premenopausal women account for this sex-related difference.[32,33] Women typically have more fat tissue than men, but have adipose tissue in the gluteofemoral area of the body which is associated with a favourable metabolic profile.[34] Due to the small number of postmenopausal women in our study, the association between BMI and fibrosis progression in women in our study primarily reflects the findings in premenopausal women, and we were not able to perform stratified analysis by menopausal status. Future studies should further investigate the differential effect of sex on the prognosis of NAFLD and elucidate the underpinning mechanism.

Several limitations of our study should be considered. First, we used ultrasonography instead of liver biopsy, which is the reference standard for diagnosis of NAFLD but was not feasible in this large-scale cohort study. Second, we used NFS and APRI in our analyses to assess fibrosis progression. We acknowledge that there is no currently available longitudinal data to examine the correlation between worsening noninvasive fibrosis markers and histological progression of fibrosis stage over time. NFS and FIB-4 are the most widely used noninvasive fibrosis markers in the setting of NAFLD.[35,36] However, these non-invasive fibrosis markers include variables in the formula that may affect the association between obesity and fibrosis progression. For instance, the FIB-4 formula has an age factor, which would potentially lead to an increase in fibrosis score during longitudinal follow-up even if the other parameters remain unchanged. The BMI variable included in the NFS formula could also affect the main finding based on BMI; however, because only a small fraction of age and BMI are taken into account in the NFS formula, we chose NFS for our main analysis. We also demonstrated a consistent finding in the sensitivity analysis using APRI, which includes neither age nor BMI in the formula. Furthermore, the positive relationship associated with NFS worsening or APRI worsening was observed when body fat percentage and waist circumference were evaluated in addition to BMI. NFS and APRI are non-invasive liver fibrosis markers that have been validated as offering good diagnostic performance in identifying advanced liver fibrosis, confirmed by liver biopsy.[20,37–39] Additionally, previous studies have demonstrated that higher liver fibrosis scores (intermediate and high fibrosis score) are associated with increased liver disease-related mortality in adults both in the United States and Korea.[40,41] Transient elastography is considered to be a non-invasive and reliable method for the assessment of liver fibrosis and can be reasonably used to quantify liver fibrosis in epidemiological research settings.[42–45] Unfortunately, in our study, data on transient elastography or liver biopsy were not available because our study was based on de-identified, retrospective cohort data from individuals who underwent a routine health check-up programme in which transient elastography was not included. Given these limitations, however, future studies should validate these findings with more accurate diagnostic modalities. Fourth, we could not assess genetic polymorphisms associated with fibrosis progression of NAFLD (eg PNPLA3 and TM6SF2 genes). Fifth, the reliability tests including inter- and intra-observer reliability for fatty liver diagnosis were conducted in the year 2013.[18] Throughout the long period of follow-up between 2002 and 2017, different radiologists were involved over time, but intra-observer and inter-observer reliability tests among the all radiologists who have worked in our centre were not performed. However, all of them were unaware of the study aims. If the degree of misclassification does not differ in relation to adiposity parameters, it is likely that this type of error would be non-differential, resulting in an underestimation of the association between adiposity parameters and incident NAFLD with/without fibrosis. Lastly, the cohort of relatively young and middle-aged Koreans used in our study limits the generalizability of the findings to other age groups, populations with a higher prevalence of comorbidities or other race/ethnicity groups.

Our study has several notable strengths. First, a longitudinal, prospective design enabled us to observe temporal associations of metabolically healthy and unhealthy obesity with the risk of fibrosis progression based on noninvasive fibrosis marker. Second, the large sample size, the use of carefully standardized clinical, imaging and laboratory procedures to ensure accuracy, the inclusion of lifestyle factors and repeated measurements allowed us to account for BMI, metabolic parameters and possible confounders as time-varying covariates. Moreover, to the best of our knowledge, our study is the first cohort study to incorporate waist circumference and body fat percentage in the analysis investigating their relationship with fibrosis progression in NAFLD in the context of metabolic health status. By adding these measures that may be better indicators of obesity and metabolic health, we were able to overcome the intrinsic limitation of BMI in distinguishing fat tissue from lean mass and thereby strengthen the validity of the study findings. Lastly, the incorporation of asymptomatic individuals with NAFLD at relatively younger age in our study reduced the possibility of introducing survivor bias caused by selecting subjects with severe disease or bias related to comorbidities and the use of multiple medications, which have been limitations in previous studies with patients at advanced stage of biopsy-proven NAFLD or NASH.

In summary, our study findings suggest that excessive adiposity per se, even without accompanying metabolic abnormality, can contribute to fibrosis progression in NAFLD. Individuals with NAFLD, even those with a normal metabolic profile, should be encouraged to maintain normal weight to prevent fibrosis progression and its associated consequences.