Mass screening for significant liver injury in patients with NAFLD will be an important medical challenge in the years to come due to the epidemics of obesity and diabetes. The inability of liver biopsy to meet this challenge makes the development of non-invasive, readily available and easy to perform serum markers a high priority. In these patients the priority is to estimate the severity of fibrosis but also to identify patients with steato-hepatitis among those with steatosis.
Many studies on non-invasive panels with significant diagnostic value for fibrosis have been published but so far the most studied biomarker is the FT , having a specific validation in NAFLD . Many previous studies have highlighted the potential utility of FT for patients infected with HCV [26, 27, 28, 32, 33, 34], HBV [29, 30] and patients with ALD [31, 32, 34].
Few tests have yet been developed for the diagnosis of steatosis and steato hepatitis [17–21]. We recently highlighted the potential utility of ST for the prediction of steatosis in patients with NAFLD, as well as in patients infected with HCV, HBV and ALD . Since the validation of ST it is therefore possible to focus on the diagnosis of NASH in patient with NAFLD after exclusion of patients without steatosis. The algorithm of NT excluded patients with steatosis predicted by ST. When screened patients without histological steatosis were included in a sensitivity analysis the diagnostic value of NT was not different than in patients with steatosis only.
We also recently demonstrated the potential utility of AshTest for the prediction of alcoholic steato-hepatitis in heavy drinkers . AshTest was designed for the diagnosis of patients with severe alcoholic steato hepatitis needing specific treatment  and does not have significant diagnostic value for the diagnosis of NASH (data not shown). Therefore a specific test for NASH was necessary to complete the non invasive estimation of frequent histological features in patients with NAFLD.
The most significant components of NT were the metabolic factors (mostly weight, triglycerides and glucose), as previously observed [17–21], but also A2M and apoA1. These proteins have been proven to be associated with fibrosis  but also with steatosis , steato-hepatitis [24, 25] and insulin resistance pathways . A2M is a protease inhibitor but also has multiple functions as a binding, carrier and targeting protein . In patients with NAFLD we previously demonstrated a very significant association between A2M and insulin levels, a hallmark of insulin resistance . Other studies have observed an increase of A2M in diabetic patients . Insulin is covalently bound to A2M  in plasma and A2M is a binding protein of Insulin-like Growth Factor Binding Protein-1 (IGFBP-1) which modifies the IGFBP-1/IGF interaction . Therefore A2M can be directly involved both in the hepatic mechanisms of insulin resistance and fibrogenesis.
Comparisons with biomarkers of alcoholic steato hepatitis (ASH)  are important as ASH and NASH share many physio-pathological mechanisms and histological features. The same associations were observed for proteins in univariate analysis with a decrease in ApoA1, haptoglobin and an increase of A2M in NASH. However the decrease of apolipoprotein A1 in NASH was much lower than in patients with ASH. In our NASH population there was a much lower prevalence of cirrhosis, as well as severe steato-hepatitis in comparison with the population of ASH . In the present NAFLD population, only two patients had a polymorphonuclear infiltrate (1.2%). In the first case there was a dramatic decrease in ApoA1 (0.05 g/L), as observed in alcoholic steato-hepatitis. In the second case the absolute value of ApoA1 was not decreased (1.72 g/L) but was relatively low in comparison with HDL cholesterol (1.64 mmol/L).
The present study has several limitations. First, the variability of the end point, histological steato-hepatitis, is even greater than for the other features of chronic liver disease, fibrosis and steatosis. There is both a significant limitation of liver biopsy due to its sampling variability , as observed for HCV hepatitis , and a high intra- and inter- observer variability [13, 14]. In the present study only 25% (63/257) of biopsy samples reached the 25 mm minimum recommended by Bedossa et al for HCV . When we used sensitivity analyses to compare the AUROCs of NT according to biopsy quality, there was no significant difference, although there was a trend in favor of better AUROCs with non-fragmented biopsies (Table 4). To reduce the observer variability related to the NASH definition, we used the recent NAS scoring system recommended by Kleiner et al . To the extent that the NAS represents the severity of current liver injury, the proposed NT may separate those with more severe injury from those with little injury. This would be of great value in clinical trial situations where the investigator might want to enroll those with severe disease first or perhaps for identifying patients at greatest risk for progression. However the NAS was intended for use in monitoring changes in liver disease and other clinical situations, and was not intended to replace the pathologist's determination of whether NASH is present or not. Therefore we checked the utility of the NT for identifying patients with bona fide NASH using the pathologist determination. Indeed the value of NT for this diagnostic of NASH was fair (AUROC = 0.80 in the validation group).
Because of the biopsy variability, discordances between biomarker and biopsy results must be discussed case by case before attributing the cause of error to biomarkers or to biopsy. In the present study, 3.5 % of patients with discordance results were attributable to NT failure versus 0.5% to biopsy failure. Being a serum marker, NT has the advantage of giving a more global estimate of liver steato-hepatitis throughout the whole liver.
The first validation group included patients from a tertiary care center, which makes it liable to referral selection bias, but the second validation group was most representative of less specialized centers.
We have used less limited inclusion criteria concerning alcohol consumption with inclusion of patients consuming up to 49 g of alcohol per day, due to our national high consumption. There was no consensual limit. However when males consuming 30 g or women 20 g or more per day were excluded (only a total of 12 patients) according to recent guidelines for the diagnosis of NAFLD , the diagnostic value of NT was not significantly changed (Table 6 and Table 7). The prevalence of patients with metabolic risk factors and moderate alcohol consumption is important in many countries and should be also analyzed in diagnostic studies.
Another drawback of liver biopsy is that for most practitioners it seems almost unethical for it to be performed in patients with normal serum transaminases values. Unfortunately, many patients with NAFLD or NASH have normal ALT levels and some of them have advanced liver fibrosis [46–48]. In the present study 50% of patients with histological borderline NASH or NASH had ALT lower than 50 IU/L. NT AUROCs for the diagnosis of NASH or borderline NASH in NAFLD were unchanged in patients with ALT values lower than 50 IU/L (Table 6 and Table 7); therefore NT could be used to diagnose NASH even in patients that are not eligible for liver biopsy.
Although there is no specific treatment currently approved to treat liver injury in NAFLD, many are being developed. The diagnosis of advanced fibrosis or NASH could be very important for motivating patients to make diet or lifestyle modifications, for the intensive treatment of complications of the metabolic syndrome or for providing weight in favor of anti-obesity surgery. The early detection of advanced fibrosis or NASH is the first step reducing future cirrhosis-related deaths. Diagnosing silent cirrhosis has important consequences in terms of screening for portal hypertension and hepatocellular carcinoma, of preventing complications and of providing a timely indication for liver transplantation.