Article Text

Download PDFPDF

Non-alcoholic steatohepatitis (NASH): why biopsy?
  1. A McNair
  1. 1Queen Elizabeth Hospital, Stadium Road, London SE18 4QH, UK; a.mcnair{at}
  1. C P Day2
  1. 2Centre for Liver Research, Newcastle University, Newcastle-upon Tyne NE2 4HH, UK; C.P.Day{at}

Statistics from

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.

The leading article by Day (

) provides a valuable summary of the current understanding of the pathogenesis and clinical relevance of non-alcoholic steatohepatitis (NASH). The article also makes two points clear: (a) we have little ability to provide accurate prognostic information in an individual patient, even when liver histology is available, and (b) although there is the promise of new treatments, the only known effective therapy at present, for the obese patient, is weight loss. Why then should these patients be subjected to liver biopsy?

Day proposes that a subgroup of patients with suspected fatty liver should undergo biopsy, including those with alanine aminotransferase (ALT) more than twice the upper limit of normal, aspartate aminotransferase >ALT values, “moderate” central obesity, non-insulin dependent diabetes, hypertension, and hypertriglyceridaemia. Gastroenterologists are commonly referred patients fulfilling these criteria but is liver biopsy likely to affect their management? The only therapeutic option at present is weight loss and all obese patients should of course lose weight, whether they have simple steatosis, NASH, or even normal liver biochemistry.

A number of arguments may be used to justify liver biopsy in these patients; histopathology may reveal unexpected findings and the results may allow more accurate prognostic information to be given to the patient.

Sherwood and colleagues1 identified 342 patients found on screening by their general practitioner to have liver enzymes raised above twice the normal upper limit who had not been referred to a specialist for further assessment. Of these, less than half were thought to require further investigation, approximately one third of whom had normal results on repeat testing. Following investigation of the remainder in a gastroenterology clinic, alcoholic liver disease and non-alcoholic fatty liver disease (NAFLD) accounted for 43% and in nearly all of the others a diagnosis could have been reached with the aid of careful history taking (alcohol and drugs), serological testing (viral hepatitis, markers of iron overload, primary biliary cirrhosis, α1 antitrypsin deficiency), and ultrasound examination (common bile duct stones). These data would suggest that biopsy for those with raised liver enzymes rarely yields an unexpected diagnosis and can be reserved for a selected subgroup of patients following non-invasive testing. None of the patients with NAFLD were cirrhotic on biopsy, although 11 (42%) had fibrosis. The results of liver biopsy would not have affected the clinical management of the small number of patients with unexpected histological findings; six patients had “autoimmune” or “cryptogenic” hepatitis, but assuming transaminase levels were only approximately twice the normal upper limit there would be no justification for immunosuppressive therapy.

It is difficult to justify liver biopsy simply to provide the patient with better prognostic information. Patients with marked fibrosis are more likely to develop cirrhosis and die from liver disease2 but it is not yet clear that patients with simple steatosis and mild fibrosis can be reassured they will not, in time, develop more severe liver disease.

Most gastroenterologists, particularly those who work in district general hospitals, adopt a pragmatic approach to the management of patients referred with abnormal liver biochemistry. This is especially important for investigations that are costly and/or are associated with significant morbidity, as with liver biopsy.3 The presence of cirrhosis will affect patient management but, as Day points out in his article, published cirrhosis rates in NAFLD are likely to be overestimates because most studies have not been done on unselected patient groups. Improved criteria for selecting patients likely to have marked fibrosis or cirrhosis are required to target biopsies more appropriately.

When effective medical treatments are available, clear criteria for liver biopsy should follow. Until then, liver biopsy in the majority of these patients outside a research setting cannot be justified.


Author’s reply

In his letter, Dr McNair states that liver biopsy is not justified in patients with suspected non-alcoholic steatohepatitis (NASH) outside a research setting. His principal arguments are that liver biopsy is unlikely to: (a) improve diagnostic accuracy, (b) provide more accurate prognostic information, and (c) effect management.

To support his first argument, Dr McNair cites a study which examined whether abnormal liver function test results are investigated appropriately in primary care and reported on the eventual diagnosis after full investigation.1 I am sure that the authors of this study would be surprised at this use of their data by Dr McNair both in view of the inaccuracies present in the citation and the conclusions drawn. For example, contrary to McNair’s letter, just over 20% (36/157) of patients thought to require further investigation had normal results on repeat testing and in the remainder, 56% (68/121) rather than 43%, had alcoholic liver disease or non-alcoholic fatty liver disease. Contrary to McNair’s statement that: “These data would suggest that biopsy for those with raised liver enzymes rarely yields an unexpected diagnosis and can be reserved for a selected group of patients following non-invasive testing”, the cited report states that: “Liver biopsy provided a diagnosis in 81 (patients) when serological tests gave normal results”. Furthermore, contrary to McNair’s statement that liver biopsy did not effect clinical management, the cited report states that: “Of these 157 patients, 97 (62%) had an identifiable diagnosis requiring hospital intervention or follow-up or both”. Therefore, the cited report suggests strongly that liver biopsy is useful for diagnosis in patients with negative serology and that the eventual diagnosis has implications for management. Specifically referring to NASH, further evidence of the utility of liver biopsy for diagnostic reasons comes from a study in which a group of eminent hepatologists were given clinical and laboratory data on a variety of different “liver” patients and asked to predict the correct diagnosis without knowledge of the findings on liver biopsy.2 Compared with other liver diseases the clinical diagnosis of NASH was the most inaccurate.

Even if I accepted Dr McNair’s argument that the diagnosis of non-alcoholic fatty liver disease (NAFLD) can be made on clinical grounds, it would be difficult to agree with his second argument that liver biopsy does not provide the clinician and patient with more accurate prognostic information. The study by Matteoni and colleagues3 along with our own study4 showed that, over at least a 10 year follow up period, patients with simple steatosis are highly unlikely to progress to more advanced disease whereas approximately 25% of patients with NASH with or without fibrosis will progress to cirrhosis within eight years. I agree with Dr McNair that further natural history studies are required but I consider that the information available at present allows clinicians to give patients with simple fatty liver a good prognosis and, importantly, in view of the large numbers of these patients referred to liver clinics, enables them to be discharged from regular follow up. The smaller number of patients with NASH with or without fibrosis however should be monitored for the development of advanced liver disease during regular follow up. Until alternative methods of distinguishing between fatty liver and more advanced disease are developed, without liver biopsy, clinicians will be committed to the indefinite follow up of the increasing number of patients under their care with suspected NAFLD.

Dr McNair’s final argument against liver biopsy in these patients is that it does not influence treatment, which he states is limited to advising weight loss. There have been no randomised controlled trials of any treatment (including weight loss) providing evidence of histological benefit in patients with NASH. However, there have been a number of encouraging pilot studies reported thus far and randomised controlled trials based on these reports are now underway. It seems inconceivable therefore that within two or three years we will not have proven treatment of benefit for NASH and unless we are happy to treat all patients with NAFLD regardless of severity, then biopsy will be required to determine which patients require treatment. Dr McNair’s argument that weight loss should be advised for all obese patients regardless of the severity of their liver disease is reasonable but it ignores the fact that, similar to the situation with heavy drinkers, obese individuals seem more likely to adhere to dietary advice if they know their obesity has seriously damaged their liver. It also assumes that the efforts made by clinicians in achieving weight loss in their patients will not be influenced by knowledge of the severity of any obesity associated end organ damage.