Percutaneous liver biopsy, a technique first accredited to Paul Ehrlich over 100 years ago, is still considered an essential component in the management of most liver diseases. However, the revolution in imaging techniques and development of serological investigations means that, at least for certain conditions, biopsy may no longer be necessary.

The usual technique for obtaining liver tissue for histological evaluation of diffuse parenchymal liver diseases is percutaneous needle biopsy, which is performed either blind or guided by ultrasound or computed tomography. The attendant mortality ranges from 0.01% to 0.1%,1,2 the major cause of death being intraperitoneal haemorrhage. Controversy still exists over whether guided biopsies can reduce the complication rate and whether an increased diagnostic yield renders them more cost-effective.3,4 To maximise diagnostic yield, it is usual practice to perform liver biopsy with a 14 G or 16 G needle. Increased complications are observed when clinicians perform liver biopsy less frequently5 and when more than three passes are made.6

It is only relatively recently that guidelines have been established for performing liver biopsies on a day-case basis7,8 and these recommend subsequent observation of patients for 6–8 hours, although the majority of complications are apparent within the first three hours.

In this issue of the Journal, Pokorny and Waterland (page 67)9 have examined the safety and possible cost benefits of performing liver biopsy in an out-of-hospital radiology clinic. Of 251 patients who underwent liver biopsy with an 18 G needle, 91.2% were discharged at 60 minutes and none were kept for longer than 2 hours 45 minutes. Moderate to severe pain was reported in 3.6% of patients, but no serious complications were observed. While a histological diagnosis was possible for all patients, it is not usual practice to use an 18 G needle for assessment of diffuse parenchymal liver disease, and there are no data regarding the adequacy of the biopsy (as regards number of portal tracts and number of sections). Could the low morbidity of the study simply reflect the use of a small biopsy needle rather than a positive benefit of radiological control? While a cost benefit was suggested when compared with in-hospital, day-case liver biopsies, the method of cost comparison relied upon two discrepant funding models — the Australian Medicare Benefits Schedule (for out-of-hospital biopsy) and DRG-based funding (for in-hospital biopsy). Ideally, actual resource costs should have been derived for each biopsy setting in order to determine an economic advantage.

Does this mean that we should be performing all our routine liver biopsies out of hospital? Although the mortality rate after liver biopsy in Pokorny and Waterland's study was low, the study does not have the power to demonstrate a clearly comparable safety profile between out-of-hospital and in-hospital liver biopsy. However, it is unlikely that a formal study with sufficient power will ever be conducted to prove this point, and perhaps ongoing audit of in-hospital and out-of-hospital biopsy and documentation of similar outcomes is all that is required to prove the safety of shorter recovery times and out-of-hospital biopsy. Nonetheless, if a major complication such as haemorrhage did occur, it would require urgent intervention that would be easier to provide in a hospital environment. The American Gastroenterological Association recommends that biopsies be performed in a unit with blood-banking facilities and an approved laboratory.7

Apart from the logistics of where liver biopsy is carried out, there is a major issue of why biopsy should be performed. There are differing opinions on the indications for liver biopsy. Current practice has recently been reviewed2 and the British Society of Gastroenterology has published guidelines for the use of liver biopsy in the United Kingdom.8

A key factor in the decision-making process should be whether knowledge of liver histology is likely to affect patient management. Histological assessment can help either to reach a diagnosis or to grade severity of disease in patients with a known hepatic disorder. Therefore, deciding whom to biopsy can be approached in one of two ways. In patients with a known or suspected disorder, biopsy enables staging of inflammation and fibrosis, providing the clinician and patient with a well-informed and accurate prognosis. This can guide or determine eligibility for treatment regimens and, in patients with cirrhosis, determine whether they should be enrolled in screening programs for hepatocellular carcinoma. The other major role of liver biopsy is in investigating patients with abnormal liver function tests (LFTs) for whom serology and imaging have been unhelpful in reaching a diagnosis.

Chronic hepatitis C is probably the most rapidly growing indication for liver biopsy in Australia. Current S100 Pharmaceutical Benefits Scheme regulations require liver biopsy before consideration of treatment of chronic hepatitis C, except in patients for whom liver biopsy is contraindicated. Should liver biopsy be performed routinely in all patients before antiviral treatment, and how does it really benefit the patient? There is growing debate on this issue within Australia10 and overseas,11 indicating that we need to re-evaluate the role of liver biopsy in hepatitis C.

The commonest cause of persistently abnormal LFTs (in the absence of markers for infectious, metabolic, autoimmune or hereditary liver disease) is non-alcoholic fatty liver disease. In a study by Daniel et al of 81 marker-negative patients with abnormal LFTs who had liver biopsies, eight patients had normal liver histology, while the remaining 73 patients all had some degree of steatosis.12 Non-alcoholic steatohepatitis was found in 26 of these patients and two had cirrhosis. Although there is a risk of cirrhosis in patients with non-alcoholic steatohepatitis, this is minimal in the absence of diabetes and obesity and in patients under 45 years.13 The result of a biopsy in patients at low risk of cirrhosis is unlikely to influence management and has not been shown to improve the benefit–risk ratio.

A final issue relates to who should perform liver biopsy. Clearly, appropriately trained clinicians should do so.5 If out-of-hospital biopsy becomes routine, there will be less opportunity to provide supervised training for registrars in this procedure. Current Gastroenterological Society of Australia guidelines for advanced training recommend that around 50 successful, supervised biopsies be performed to ensure adequate training. So, who should be trained to perform liver biopsy and how do we accredit such individuals?

Liver biopsy provides invaluable information and a histological diagnosis remains the gold standard in many liver disorders. However, the benefits for diagnosis and management need to be clearly defined before subjecting patients to an invasive procedure, albeit one with low risk. While the study by Pokorny and Waterland9 tempts us to move to an out-of-hospital, short-stay approach to liver biopsy, the issues of safety and diagnostic adequacy of relatively small core biopsies need to be well proven before this approach can be more widely adopted.