Predictors of Nonalcoholic Steatohepatitis and Significant Fibrosis in Non-obese Nonalcoholic Fatty Liver Disease

Donghee Kim; Won Kim; Sae Kyung Joo; Jung Ho Kim; Stephen A. Harrison; Zobair M. Younossi; Aijaz Ahmed


Liver International. 2019;39(2):332-341. 

In This Article


Our findings demonstrated that non-obese NAFLD carries as severe liver histology as obese NAFLD and that visceral adiposity and diabetes are major drivers of NASH and fibrosis in non-obese NAFLD subjects. Taking into consideration the increased age of non-obese subjects with NAFLD, non-obese subjects with NASH were noted to have higher prevalence of significant fibrosis than obese subjects with NASH. These findings indicate that severity of hepatic histology in non-obese NAFLD is comparable to obese NAFLD with plausible influence of factor(s) other than obesity playing a role in the progression of NAFL to NASH and significant fibrosis. It is unclear which phenotype of non-obese NAFLD is associated with the propensity to develop advanced disease. We found that presence of diabetes, higher AST or ALT levels and higher VAT area were independent predictors associated with NASH or significant fibrosis in subjects with non-obese NAFLD. Additionally, HDL cholesterol in the non-obese population may have a protective role against NASH. Our results provided an explanation for this phenomenon, namely diabetes, visceral fat and hepatic injury are more likely to induce NASH or significant fibrosis in the non-obese NAFLD population. These clinical predictors can facilitate the timing of liver biopsy in non-obese subjects suspected of NAFLD.

Recent epidemiological studies have shown that approximately 10%-30% (7%-21% for Western and 3%-27% for Eastern studies) of non-obese individuals have non-obese NAFLD, when the ethnic differences in the BMI cut-off to determine obesity are taken into account.[6] A recent Asian study with 307 biopsy-proven NAFLD subjects showed that non-obese NAFLD had lower NAS compared to obese NAFLD, which was mainly driven by lower severity of steatosis and balloon degeneration.[25] In our study, non-obese subjects with NAFLD had lesser degrees of steatosis compared to obese subjects with NAFLD. Otherwise, morphological features of non-obese NAFLD and obesity-associated NAFLD are indistinguishable from each other, while non-obese subjects with NASH had a higher prevalence of significant fibrosis and advanced fibrosis compared with their obese counterparts. Recent genomewide association studies demonstrated that genetic polymorphisms in patatin-like phospholipase domain-containing protein 3 play a more important role in non-obese NAFLD compared to obese NAFLD,[26,27] and these findings may contribute to elucidation of predictors for non-obese NASH and significant hepatic fibrosis when integrated with recent emerging genetic data.[6]

A subset of normal-weight individuals displayed a metabolic profile similar to the profile that is often associated with being obese.[28] This subset of subjects is referred to as "MONW."[29,30] In the light of the well-known association between insulin resistance and NAFLD, non-obese NAFLD may represent a subset phenotype with NAFLD in MONW subjects.[6] The presence of non-obese NAFLD may be a harbinger of systemic inflammation that portends worse clinical outcomes in these subjects. We confirmed that increased VAT area had a higher influence on the pathogenesis of NASH and significant fibrosis in non-obese than obese subjects with NAFLD. A recent study suggested that non-obese NAFLD subjects with higher waist circumference may have higher risk of significant fibrosis than obese NAFLD subjects, which did not reach statistical significance due to small sample size and the limitation of waist circumference.[31] In addition, our multivariate analysis regarding non-obese NASH suggested that two major risk factors for NASH are clearly evident: diabetes [OR 3.65] and female sex [OR 2.49].

Asians generally have a higher percentage of VAT area compared with Europeans and Hispanics of the same age, gender and even BMI.[32–34] While waist circumference has been widely used to measure abdominal obesity,[20] it does not distinguish VAT from SAT.[20,35] SAD, which measures the anteroposterior diameter of the abdomen, reflects the VAT area because subcutaneous fat is displaced inferiorly by gravity.[20] Our previous study reported that SAD has a stronger correlation with abdominal VAT area than BMI or waist circumference, irrespective of age, gender and degree of obesity.[20] In particular, SAD has a stronger correlation with VAT area among non-obese subjects compared with obese subjects.[20] In the current study, we found that as a surrogate marker of VAT, SAD is independently associated with NASH or significant fibrosis among non-obese population. Regarding non-obese NAFLD, we suggest that the clinical use of SAD has advantages over other anthropometric measures in predicting NASH and significant fibrosis.

One strength of this study was that the histological diagnoses of NASH and other histological features were reviewed and established by an expert pathologist who specialized in liver pathology. The wealth of demographic, clinical and laboratory data allowed the comparison between non-obese and obese NAFLD with sufficient statistical power. The main limitation of this study is the causality of the observed relationships due to the cross-sectional nature of the study design regarding the determinants of NASH or significant fibrosis. Additionally, the number of lean subjects (BMI <23 kg/m2) with NAFLD remains relatively small. This precludes extensive analyses of risk factors associated with NASH or significant fibrosis in lean subjects with NAFLD. However, the cut-off point of BMI for observed risk varies from 22 to 25 kg/m2 in different Asian populations; for high risk, it varies from 26 to 31 kg/m2.[36,37] In the cohorts of 1 million East Asians, the lowest risk of death was seen among persons with a BMI in the range of 22.6-27.5 and the risk was elevated among persons with a BMI higher than 27.5.[38] In addition, the WHO Expert Consultation recommended that the WHO BMI cut-off points should be retained as international classifications in Asians.[36,37] Finally, because this study included only subjects of East Asian ethnicity, the conclusions might not be generalizable to other ethnic populations.

With these caveats in mind, we conclude that non-obese NAFLD subjects display a similar degree of histological severity compared to their obese counterparts. While the pathophysiology and natural history of non-obese NAFLD remains to be elucidated, these data suggest that a high index of suspicion for NASH or significant fibrosis is needed to promptly diagnose non-obese patients with NAFLD in the era of precision medicine with greater emphasis on prediction and prevention of disease initiation and progression.