Article Text

Download PDFPDF

Are non-invasive tests accurate enough to predict hepatic fibrosis in non-alcoholic fatty liver disease (NAFLD)?
  1. Norberto C Chavez-Tapia1,
  2. Claudio Tiribelli1,2
  1. 1
    Centro Studi Fegato, Liver Research Center, AREA Science Park, Trieste, Italy
  2. 2
    Department of Life Sciences, University of Trieste, Italy
  1. Dr C Tiribelli, Centro Studi Fegato, AREA Science Park, Bldg. Q., SS 14 km 163.5, 34012 Trieste, Italy; ctliver{at}

Statistics from

In recent decades several diseases primarily associated with obesity have reached epidemic proportions and many of them such as high blood pressure, hyperlipidaemia and type 2 diabetes mellitus, are already an important social problem. It has been calculated that in 2030 type 2 diabetes mellitus per se will be responsible of 3% of deaths in the world together with ischaemic heart disease and cerebrovascular disease, which will account for 13 and 11% of all cause of deaths.1 Non-invasive, low-cost, easy to perform and reproducible diagnostic tools have been used in epidemiological studies to allow the screening of large populations. While the assessment of fasting glycaemia, oral glucose tolerance test blood pressure, and serum cholesterol and triglycerides is routine in clinical practice, the use of a hyperinsulinaemic–euglycaemic clamp in every patient with type 2 diabetes mellitus is clearly not feasible. Since the diagnostic accuracy, assessed by the value of the area under the receiver operator characteristic curve (ROC), of all 19 candidate indexes of insulin resistance to predict incident diabetes varies from 0.75 to 0.8,2 these variables are extensively used both in clinical practice and trials.

A very similar scenario occurs in clinical hepatology. Fatty liver (FL) or hepatic steatosis is a very common finding for the general practitioner and is the most frequent cause of referral to liver or gastroenterology centres.3 After exclusion of viral infections, drugs, metabolic diseases and autoimmune disorders, FL is classified as non-alcoholic (NAFLD) or alcoholic fatty liver disease on the basis of daily ethanol intake. NAFLD is characterised fatty infiltration of hepatocytes associated with histological abnormalities ranging from simple hepatic fat accumulation with mild non-specific inflammation to non-alcoholic steatohepatitis (NASH) characterised by the presence of moderate to severe necro-inflammation, ballooning degeneration, Mallory hyaline and/or fibrosis.4 Discrimination between NAFLD from NASH, or in other words the assessment of the degree of fibrosis and inflammation, is important as the two conditions have a very different clinical outcome, the former being rather benign while the second often progressing to cirrhosis.5 Studies performed in the general population showed that almost 20% of population suffer from NAFLD6 and 3% from NASH,7 and almost 12% of patients with fibrosis will develop cirrhosis.8 In a recent study correlating the natural history of NAFLD with the histological picture9 steatosis increased in only a small percentage (6–18%) of patients while fibrosis showed a larger percentage (23–53%). Although intrinsically limited by sampling variability, this study indicated that the progression, if any, of NAFLD-induced fibrosis is slow but the survival is related to the presence and the degree of fibrosis at the time of the diagnosis.10 11

Determination of the extent of liver fibrosis in NAFLD subject is therefore of great clinical importance. Although liver biopsy is the “gold standard” to assess the extent of inflammation and fibrosis, it is technically and ethically impossible to perform this invasive procedure in all subjects with FL. Therefore a non-invasive test will provide very useful information to discriminate NAFLD from NASH, quantifying the degree of liver fibrosis and allowing to design effective preventive programs. Several approaches may be used as clinical and/or biological markers, imaging or hybrid image techniques such as elastography. Although generally acceptable in terms of diagnostic accuracy, each of these approaches has intrinsic limitations which have to be considered. For example, the diagnostic performance of the clinical markers is modified by the distribution of the variables included in the model12 while in the case of the biological markers, the lack of widely availability and standardisation may limit their utility. In spite of substantial advances in imaging techniques, most of them are primarily focused in measuring the degree of fatty infiltration rather than fibrosis of the liver.13

In this issue Harrison et al14 (see page 1441) report the development of a new score to identify and assess the degree of hepatic fibrosis in NAFLD patients using four easy, non-invasive, low-cost, and reproducible variables (BMI, AST/ALT Ratio, presence of diabetes, BARD). Based on a large convenience sample of obese subjects, they demonstrate that a BARD score ranging from 2 to 4 was associated with an odds ratio of 17 to predict fibrosis, with a negative predicted value close to 100%. BARD is a new, additional score which adds to a number of previous attempts to non-invasively assess hepatic fibrosis. As in table 1, other scoring systems have been reported showing similar efficacy with comparable ROC values ranging from 0.79 to 0.98. With the exception of transient elastography (limited by the presence of central adiposity in NAFLD patients) and breath test, all “fibrosis scores” consider a series of blood parameters to be computed to end in a number. Unfortunately, none of them have been validated by prospectively following changes, thus questioning the applicability in a large cohort of patients. In the case of the FibroTest,15 the access to the calculating algorithm is limited by a fee, rendering this approach not widely available. In any case the possibility to assess, non-invasively, hepatic fibrosis appears to be particularly useful in clinical practice, as well as in pharmacological research. Quantitative parameters have shown to be clinically important as in hepatitis C where the availability of the viral load and genotype led to substantial improvement in the treatment of this disease, and the MELD score is now the key parameter to define the need for liver transplantation.

Table 1 Comparisons among non-invasive markers of fibrosis in non-alcoholic fatty liver disease

More than 20 years after the classical description of NAFLD,4 we start having reliable elements to define this epidemic disease by easy, non-invasive, low-cost, and reproducible diagnostic tools. Although a critical and prospective comparison of the different scores is still missing and would be highly desirable, there are few doubts that they will be useful to predict the final outcome of liver disease and the potential benefits of the drugs recently suggested beneficial in treating NAFLD and/or NASH. The future will tell which of the scores will win the race.


View Abstract


  • Funding: NCC-T was supported by a Fellowship in Translational Hepatology by the Centro Studi Fegato.

  • Competing interests: None declared.

  • The permanent address of NCC-T is Department of Gastroenterology, Instituto Nacional de Ciencias Médicas y Nutrición Salvador Zubirán, Mexico.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles