The value of serum adipokine measurement in non-alcoholic fatty liver disease

Authors


Non-alcoholic fatty liver disease (NAFLD) is now considered as one of the most common liver disorders worldwide, affecting up to 30% of the general population (1, 2). NAFLD encompasses a clinicopathological spectrum of histological abnormalities, including various degrees of steatosis, lobular inflammation and fibrosis of the liver (3). Importantly, the histological picture seems to be related to clinical outcomes in the sense that, while pure hepatic steatosis has a benign clinical course, the presence of necroinflammatory changes [where the term non-alcoholic steatohepatitis (NASH) is used] and/or hepatic fibrosis is associated with an increased risk of developing advanced fibrosis, cirrhosis and hepatocellular carcinoma (4, 5). It has been estimated that up to one-third of subjects with NAFLD have NASH (5). Thus, identification of this subset of patients who develop the aggressive form of the disease is one of the most challenging aspects in the NAFLD field. At present, liver biopsy is still considered the gold standard to accurately differentiate between NASH and bland steatosis. However, the costly and invasive nature of the procedure as well as its potential risks (complications occur in 0.5–1% of cases) limit its wide use in daily clinical practice. Therefore, efforts have been made to develop non-invasive ways of detecting NAFLD patients at risk for progression. Several research groups have attempted to identify one or more clinical or biochemical variables that, individually or grouped in different scoring systems or predictive panels, allow to discriminate patients with NASH and/or fibrosis. Additionally, the use of new radiological techniques such as transient elastography for this purpose is also under intense scrutiny. Excellent reviews or commentaries on this topic have been published recently analysing the available data and future prospects (6, 7). In the current issue of Liver International, Lemoine et al. (8) add significant information to the matter by examining the predictive value of serum adipokine measurement in NAFLD.

The potential role of the fat secreted in the pathogenesis and progression of NAFLD has recently gained considerable attention (9, 10). Recently recognized as the largest endocrine organ of the body, adipose tissue secretes at least a dozen proteins (collectively called adipokines or adipocytokines) that are able to influence insulin sensitivity and have been shown to exert pro-inflammatory and anti-inflammatory effects in the liver (11). A significant body of information, mainly regarding the possible role of tumour necrosis factor-α (TNF-α), leptin and adiponectin in NAFLD, has been accumulated. Data on other potential players such as resistin, visfatin, apelin and other adipokines are still scarce, with few studies assessing the levels of these molecules in the setting of a fatty liver (11).

A number of human studies have examined the levels of adipokines in NAFLD aiming to explore associations with different stages of the disease. Table 1 summarizes published case–control studies performed in adults and involving mainly leptin, adiponectin and TNF-α. Of note, contradictory or conflicting results have been found. In the case of leptin for example, one study reported that its serum levels are increased in NASH patients independently of the degree of obesity, with higher levels in patients with more advanced disease (12) but three other studies show that leptin levels do not correlate with inflammation or fibrosis but only correlate with the severity of steatosis (13–15). Thus, at present, serum leptin alone cannot be used as a non-invasive marker for the diagnosis of NASH. In the case of adiponectin, decreased levels of this adipokine have been consistently found in patients with NAFLD compared with controls (Table 1) but its association with the presence and severity of inflammation and liver fibrosis is still controversial. Reduced serum adiponectin levels in patients with NASH compared with matched controls and patients with simple steatosis were first reported by Hui et al. (16). However, another study, with a larger number of subjects, found an association only with intrahepatic fat but not with inflammation or fibrosis (17). Other studies also yielded discrepant results (18–21) (Table 1). Differences among the published studies are likely related to methodological issues such as sample size, patient populations (morbidly vs. non-morbidly obese, inclusion of diabetic patients, gender proportions) as well as presentation and statistical treatment of data.

Table 1.  Summary of selected case–control studies analysing the association between serum levels of main adipokines with histological liver damage in non-alcoholic fatty liver disease
ReferencesPopulationMajor findings
  1. IL, interleukin; NAFLD, non-alcoholic fatty liver disease, NASH, non-alcoholic steatohepatitis; OR, odd ratio; TNF, tumour necrosis factor.

Uygun et al. (12)49 NASH patients, 32 patients with chronic viral hepatitis and 30 healthy adultsSerum leptin levels were significantly higher in patients with NASH, while they were not affected by chronic viral hepatitis
Chitturi et al. (15)47 NASH patients and matched controlsSerum leptin correlates with the severity of hepatic steatosis in NAFLD patients but was not an independent predictor of hepatic inflammation or fibrotic severity
Angulo et al. (14)88 NAFLD patients and matched controls.No relationship was demonstrated between leptin levels and fibrosis stage
Hui et al. (23)109 NAFLD patients and 109 matched controlsSubjects with NASH had reduced adiponectin levels and increased TNF-α but not leptin levels. Adiponectin, TNF-α and leptin were not associated with the extent of hepatic fibrosis
Bugianesi et al. (24)174 NAFLD patients and 42 controlsAdiponectin serum levels were inversely correlated to the percentage of hepatic fat content and no association was established with necroinflammation or fibrosis
Pagano et al. (18)17 NAFLD patients and 20 matched controlsPlasma adiponectin was significantly lower in NAFLD patients than controls but similar in simple steatosis and in steatohepatitis
Targher et al. (19)60 NAFLD patients and 60 controls.Low adiponectin levels independently predicted hepatic steatosis and necroinflammation but not fibrosis
Wong et al. (21)80 NAFLD patients and 41 healthy controlsLow adiponectin level and increased leptin level were associated with NAFLD. A high TNF-α level was independently associated with NASH. TNF-α level had a positive correlation with necroinflammatory grade and fibrosis stage
Jarrar et al. (20)26 NASH patients, 19 with simple steatosis and 50 controlsSerum TNF-α and IL-8 were higher in NAFLD patients. Visfatin in NASH patients was lower than simple steatosis and obese controls. In comparison with simple steatosis, IL-8 and adiponectin were independently associated with NASH. In this study, TNF-α was the only independent predictor of fibrosis in NASH
Lemoine et al. (8)57 NASH patients, 17 with simple steatosis and 10 controls.NASH was associated with Adiponectin/Leptin ratio <1.4 (OR=5.2). IL-6 levels were associated with liver fibrosis. The hepatic expression of adiponectin receptor 2 was significantly higher in patients with NASH compared with controls and was related to necroinflammatory injury

Lemoine et al. (8) conducted a retrospective study of patients who underwent a liver biopsy due to elevated liver enzymes. They assessed adipokine levels (leptin, adiponectin, soluble TNF receptor 1 and interleukin-6) and failed to find significant associations between histological severity of the disease with either leptin or adiponectin. However, they did find that the adiponectin to leptin (A/L) ratio was useful in predicting NASH in their patients with a sensitivity of 0.86 and a specificity of 0.54 for an A/L ratio <1.4 × 103. The rationale of the A/L ratio relies on the opposite effects and reciprocal changes of leptin and adiponectin in obesity and NAFLD. Moreover, the A/L ratio has been used in other scenarios such as type 2 diabetes, where it has potential utility as a marker of atherosclerosis (22) or as an indicator of insulin resistance (23). In their multivariate analysis, Lemoine and colleagues also found that a homeostasis model assessment score>3 was also independently associated with NASH and that the combination of this variable with an A/L ratio <1.4 × 103 was able to predict NASH with an area under the receiver operating characteristic curve of 0.82. Also, a predictive value of interleukin-6 levels for the presence of liver fibrosis in addition to the known association of this variable with body mass index and age was found. As in other studies published recently, the study of Lemoine and colleagues can be criticized for several methodological flaws such as a relatively small sample size, the inclusion of diabetic patients and the relatively small proportion of patients with severe fibrosis. That said, a common conclusion was reached: results must be validated in a larger group of patients. However, the use of the A/L ratio as a fibrosis predictor is novel to the field and provides a timely opportunity to review the data on the use of adipokine levels in the non-invasive prediction of NASH and fibrosis in NAFLD patients (Table 1). After performing this review, one can conclude that the information available on this particular topic is still controversial and that the role of determining the levels of one or two adipokines to reliably exclude the aggressive forms of the disease in NAFLD patients remains unclear. In contrast, the use of combined sets of clinical variables and serum determinations can augment the diagnostic performance across different stages of fibrosis and is promising (24). However, these panels must be validated in large and well-selected groups of NAFLD patients and carefully matched healthy controls before being accepted in daily clinical practice. Certainly, to correctly ascertain the prognosis and to have the ability to stratify NAFLD patients is still a challenging issue.

Acknowledgements

This work was partially supported by a grant from the National Commission for Scientific and Technological Research of the Chilean Government (CONICYT); FONDECYT No. 1080170.

Ancillary