Motion – All patients with NASH need to have a liver biopsy : Arguments against the motion

This article was originally presented at a symposium entitled, “Controversies in Gastroenterology”, sponsored by Axcan Pharma, Toronto, Ontario, April 8 to 10, 2002 Hepatology Section, University of Maryland, Baltimore, Maryland, USA Correspondence: Dr J Laurin, Hepatology Section, University of Maryland, 22 South Greene Street, N2W50, Baltimore, Maryland 21201-1595, USA. Telephone 410-328-1358, fax 410-328-1897, e-mail jlaurin@medicine.umaryland.edu J Laurin. Motion – All patients with NASH need to have a liver biopsy: Arguments against the motion. Can J Gastroenterol 2002;16(10):722-726.

P hysicians are often faced with asymptomatic patients who have elevated levels of alanine aminotransferase (ALT), aspartate aminotransferase (AST), alkaline phosphatase or gamma-glutamyl transferase.These enzymes are present in many organs, but significant elevations usually imply a hepatic origin.History, physical examination, blood tests and radiological imaging studies are helpful in determining the cause of abnormal liver enzymes.
Minor elevations of liver enzymes are usually of no clinical consequence if the disorders listed in Table 1 have been excluded (1).Laboratory tests that may identify the cause of asymptomatic aminotransferase elevation are listed in Table 2.A myriad of medications, herbs and substances of abuse are reported to cause elevated levels of liver enzymes.Liver biopsy is considered the most accurate method of staging liver disease.It is not clear, however, that the results would lead to a change in the management of a patient who is not taking hepatotoxic medication or significant amounts of alcohol and in whom all markers of chronic liver disease are absent.Such patients would most likely have nonalcoholic fatty liver disease (NAFLD), nonspecific histological abnormalities or normal findings.
Pratt and Kaplan (1) recommended that liver biopsy be undertaken if aminotransferase levels are persistently more than twice the normal value.They noted that, even though the procedure is unlikely to lead to a diagnosis or changes in management, the results often reassure the patient and physician that no serious disease is present.There are, however, inherent risks and significant costs associated with performing liver biopsies.Considering that millions of persons in North America may have NAFLD, it is difficult to recommend liver biopsies to confirm the diagnosis when there is no proven therapy and when the results would not change the management of the patient.

NONALCOHOLIC STEATOHEPATITIS
Nonalcoholic steatohepatitis (NASH) is a liver disease characterized by the presence of fat (steatosis) with necroinflammation in patients with no history of alcohol consumption.NAFLD encompasses a broader spectrum of liver disease that also includes steatosis alone, or steatosis plus inflammation without hepatocyte ballooning degeneration, necrosis or fibrosis.It may occur in a variety of clinical settings, but is most frequently associated with obesity and type II diabetes mellitus.The diagnosis of NASH is, in part, based on the exclusion of other liver diseases, which can be done by taking a thorough history of alcohol intake and ingestion of potentially hepatotoxic medications and by performing laboratory and serologic testing.The finding of steatosis and necroinflammation on liver biopsy is the only method of making a definitive diagnosis of NASH.
NASH has been described in many populations throughout the world, including North and South America, Japan, Australia, Europe and the Middle East (2).Its prevalence has been estimated using several methodologies, including liver biopsy, ultrasonography and postmortem analysis.In persons undergoing liver biopsy, the prevalence of NAFLD and NASH range from 15% to 39% and 1.2% to 4.8%, respectively (3)(4)(5).Autopsy examinations following sudden deaths from automobile and airplane crashes revealed that 15.6% to 24% had NAFLD and 2.1% to 2.4% had NASH (6,7).Ultrasound screening studies of the general population in Japan and Italy estimated that the prevalence of fatty infiltration of the liver is 16.4% to 23% (8)(9)(10).It thus appears that the prevalences of NAFLD and NASH are 20% and 2% to 3%, respectively.
Recent studies have confirmed these estimates (11,12).Jeanne Clark et al (11), using the large sample population (n=13,500) from the Third National Health and Nutrition Examination Survey (1988 to 1994), found that the prevalence of NAFLD was 23.5%.This condition was the most common cause of abnormal liver enzymes, and was more prevalent among men than women, and among black people than white people.On the other hand, Luisa Santos and colleagues (12) undertook a retrospective analysis of 51 patients with NASH among 861 patients in a hepatology clinic, and found that the condition was rare among black people (1.4%), compared with white people (7%) or Hispanic people (8%).

SHOULD LIVER BIOPSY BE PERFORMED
TO DIAGNOSE NASH?NASH and NAFLD are diagnoses of exclusion and require liver biopsy for definitive diagnosis.Most persons with NASH present with asymptomatic elevations of aminotransferases.The key question thus becomes, 'should all persons with asymptomatic elevation of liver enzymes undergo liver biopsy?'Many of the studies that examined the usefulness of liver biopsy in the evaluation of asymptomatic patients with elevated liver enzymes occurred before the development of serologic markers for hepatitis C, and featured lenient inclusion and exclusion criteria.
A 1987 study of 100 asymptomatic persons with abnormal ALT levels included 36% with intermittent elevations of ALT, 28% with persistent elevations, and 33% with abnormalities that occurred only once (13).Most of the subjects were obese or drank alcohol.No apparent reason for the abnormalities could be found in 22 patients, but serologic testing for hepatitis C was not available at that time.In a follow-up study of the same subjects, 17 tested positive for hepatitis C antibody, as determined by both radioimmunoassay and second generation, four-antigen recombinant immunoblot assay (RIBA-2) (14).
Hay et al (15) undertook liver biopsies in 47 asymptomatic persons with three-to eight-fold elevations of AST that persisted for at least six months, and who had already been screened for hepatitis B infection, alcohol abuse and drug-related disease (15).They did not, however, undergo biochemical or radiological evaluations for autoimmune disease, cholestasis or hepatitis C. Liver biopsy revealed 'chronic active hepatitis' in 72% of the cases, including 34% with cirrhosis, while 21% had steatohepatitis and 6% had a variety of other disorders.Neither clinical nor labora-tory features could distinguish between patients with or without cirrhosis, and there was overlap in the presence of antinuclear antibodies between patients with chronic active autoimmune hepatitis and steatohepatitis.The authors suggested that liver biopsy might be justified in these patients.
Van Ness and Diehl ( 16) evaluated 90 patients with elevated liver enzymes and compared the diagnosis made before liver biopsy, based on the history, physical examination, laboratory tests and imaging studies, with the final histological diagnosis.The positive predictive value for the prebiopsy diagnosis ranged from 88% for alcoholic liver disease to 56% for fatty liver.Liver diseases that required specific therapy, other than alcohol abstinence, were diagnosed only after review of the histology in five cases (four with granulomatous hepatitis and one with druginduced liver diseases).On the other hand, specifically treatable liver disease was excluded in four cases in whom the prebiopsy evaluation suggested its presence (including two cases each of suspected chronic autoimmune liver disease and granulomatous hepatitis).The authors concluded that liver biopsies should be used in conjunction with noninvasive evaluation.This study was published in 1989, however, before the advent of hepatitis C testing.Moreover, strict criteria for the diagnosis of various chronic liver diseases were not employed.
Hultcrantz et al (3) studied 149 asymptomatic patients who had moderately elevated serum aminotransferase levels, excluding those with alkaline phosphatase levels of more than twice the upper limit of normal, or symptoms or physical signs of chronic liver disease.The patients were tested for viral (hepatitis B), autoimmune and genetic causes of liver disease.Liver biopsy revealed fatty liver in 94 patients (63%), 'chronic active hepatitis' in 15 (10%), 'chronic persistent hepatitis' in 15 (10%), and cirrhosis in nine (18%).Two patients had granulomas in the liver, one with sarcoidosis and one with antibodies to smooth muscle and antinuclear antibodies (presumably due to autoimmune hepatitis).Five patients had genetic hemochromatosis, as suggested by elevated ferritin levels, three had alpha-1antitrypsin deficiency, and three were heterozygous for alpha-1-antitrypsin deficiency.One patient had schistosomiasis and two had either nonspecific findings or normal histology.Unfortunately, hepatitis C testing was not yet available at the time of the study, but the authors noted that 73% (22 of 30) of the patients with chronic hepatitis and 12% (1 of 9) of those with cirrhosis had risk factors for bloodborne diseases.Ultrasonography and liver scintigraphy could not differentiate cirrhosis from fatty infiltration of the liver.The authors concluded that histological assessment was essential to the diagnosis of the cause of elevated aminotransferases.
A more recent prospective study of 1124 patients with persistently elevated liver enzymes, from the Henry Ford Hospital in Michigan (17), included 81 patients in whom there were no specific biochemical (for infectious, metabolic, autoimmune or hereditary) or radiological abnormalities, alcohol abuse or use of hepatotoxic drugs.Liver histology revealed steatosis in 41 patients (51%), steatohepatitis in 26 patients (32%), fibrosis in four patients (5%) and cirrhosis in two patients (2%).All 73 abnormal liver biopsies showed some degree of steatosis.There was no significant association between histological findings and the presence of obesity, hyperlipidemia or diabetes mellitus.Nor was there a significant association between histological findings and sex or symptoms.This study showed that, in marker-negative patients with chronically elevated aminotransferases, the most likely histological diagnosis is fatty metamorphosis of the liver with occasional associated fibrosis.
In 2000, Sorbi et al (18) prospectively evaluated the utility of liver biopsy in 36 asymptomatic patients with chronic elevation (greater than 50% above normal) of AST, ALT or alkaline phosphatase levels.The patients were screened for diabetes mellitus and hypertriglyceridemia, consumed less than 40 g of alcohol per day, took no potentially hepatotoxic medications, and did not exhibit signs of chronic liver disease or strong evidence for a particular liver disease.A diagnosis and a management plan (including lifestyle modification, weight loss, avoidance of excessive alcohol use or hepatotoxic medications, monitoring and specific therapies) were formulated before liver biopsy was undertaken.Histological assessment led to a change in diagnosis in five patients and a change in treatment in two.In three patients with a prebiopsy diagnosis of presumed NASH, the liver biopsy revealed normal histology, and, in a fourth patient, abnormalities consistent with the diagnosis of primary sclerosing cholangitis (PSC) were found and later confirmed with endoscopic retrograde cholangiopancreatography.This patient had liver enzyme abnormalities suggestive of hepatocellular injury, and not cholestasis.Finally, a patient with suspected alcoholic liver disease had normal histology.Lifestyle recommendations were not significantly altered by liver biopsy results.The frequency of liver test monitoring was increased in six patients and decreased in seven others.In the majority of cases, no changes were made in medications for treating liver disease.In six patients with NASH, three with PSC and one with autoimmune hepatitis, investigational therapies were offered.In one patient with NASH, corticosteroids were discontinued, and, in another, ursodeoxycholic acid was stopped.The decision to stop these medications was not based on liver biopsy information, because the pre-and postbiopsy diagnoses were the same.The authors concluded that, even though a liver biopsy might change the final diagnosis in some patients with nonspecific elevations of liver enzymes, the resultant modifications in management were often minor, especially because no proven therapy exists for the majority of these patients.They recommended that, until effective therapy for NASH is available, patients should be told that liver biopsy is unlikely to alter their management.Another study also found that close clinical follow-up is the most cost effective strategy for asymptomatic patients with chronically elevated aminotransferases and negative tests for viral, autoimmune and metabolic markers of chronic liver disease (19).Finally, Goddard and Warnes (20) reviewed the literature and came to the same conclusion -most liver diseases for which a specific treatment is available can be diagnosed without a liver biopsy.They emphasized that, because of sampling error and observer error, cirrhosis may be missed in 20% of cases.They considered the risks and benefits of liver biopsy and concluded that this test can be omitted in most cases of asymptomatic elevations of liver enzymes.

RISKS OF LIVER BIOPSY
There is a less than 1% risk of serious complications in persons undergoing liver biopsy, but up to 5% may require hospitalization, rarely days or weeks later.Furthermore, the procedure fails to obtain tissue in 2% of cases (18,21,22).
A retrospective study of more than 9000 liver biopsies performed over a 20-year period revealed a fatality rate of 0.11%, and a 0.24% risk of nonfatal hemorrhage (23).The presence of malignancy and increased number of passes were independent risk factors for complications.The risks of fatal and nonfatal hemorrhage were 0.4% and 0.57%, respectively, in patients with malignancy, but only 0.04% and 0.16%, respectively, for those with nonmalignant disease.
A more recent study showed that ultrasonographic guidance decreased the rates of hospitalization and pain requiring treatment after percutaneous outpatient liver biopsy (21).There was also a strong trend toward reduction in the incidence of bleeding and hypotension not requiring hospitalization.

COSTS OF LIVER BIOPSY
It is difficult to draw general conclusions about the costs of liver biopsy because of the differences in charges among various institutions.In the United States, Medicare reimburses approximately US$113 for physician costs and US$800 for the facility fee.The global Medicare reimbursement for pathology services is approximately US$155.Ultrasoundguided liver biopsies, which may be safer (21), add to the cost of the procedure.If 20% to 25% of the general population had NAFLD or NASH, the total costs of diagnostic liver biopsy for these individuals would be astronomical.

SUMMARY
NAFLD is a diagnosis of exclusion, and the definitive diagnosis of NASH is based on liver biopsy.Most patients who present with asymptomatic elevation of aminotransferases, and who have no clinical signs or blood test markers of chronic liver disease, have NAFLD or NASH.Because there is no proven therapy for these conditions, the costs and risks of liver biopsy cannot be justified.