NASH: a hidden and silent fibroser finally revealed?
Article Outline
Non-alcoholic fatty liver disease (NAFLD), defined as an excessive accumulation of fat in hepatocytes, is an adaptive response of the liver to insulin resistance. It is an increasingly common condition [1] which has been brought to the frontline of the medical community by the exploding epidemic of obesity and diabetes. The natural progression of insulin resistance and endogenous noxious insults (such as free radical production, mitochondrial dysfunction, endotoxin to name just a few) which are at least in part related to the presence of excessive fat in the liver, can trigger, in certain individuals, the development of non-alcoholic steatohepatitis (NASH) [2]. NASH itself can induce a fibrogenic response that can result in cirrhosis, hepatocellular cancer and liver-related death.
Thus, the stage has been set for the potential of liver damage in the insulin resistance syndrome and the natural history of NAFLD/NASH has been grossly outlined by previous studies. It is now urgent to fully establish NAFLD/NASH as a complication of obesity and diabetes with the potential for serious disease beyond the circle of believers (liver disease specialists). The challenge of convincing the physicians primarily in charge of diabetes and obesity of the dire prognosis potentially carried by NASH is not trivial. Long-term follow-up studies from obese patients showing increased cirrhosis-related morbidity and mortality [3] and studies from cohorts of diabetic patients showing increased incidence of non-alcoholic chronic liver disease and of hepatocellular carcinoma [4] are highly valuable. Other important data could be provided by histological follow-up studies in NASH patients. Several such studies are now available.
In this issue of the Journal, Adams et al. from the Mayo Clinic present data on 103 patients undergoing repeat liver biopsy for NASH [5]. Half of these patients were treated by ursodeoxycholic acid or clofibrate between the two biopsies within trials that did not demonstrate a therapeutic benefit for these drugs. A quarter of patients received placebo and another quarter had the repeat liver biopsy performed as part of the clinical follow up. How close the drug treated patients would come to the natural history of histological course in NASH is debatable, mainly because the small sample size of these negative trials does not rule out an impact of these drugs on histological lesions. Besides, the time interval between the two biopsies is probably shorter for patients included in these 1-year trials than the mean 3.2 years of the whole population. However, with these restrictions in mind, this study provides some important new insight in the histological course of NAFLD. And the news does not look good for the liver.
First, in more than a third of patients fibrosis increased between the two biopsies and in a quarter of those biopsied more than 4 years apart, the progression is substantial, more than two stages on the Brunt score. Remarkably, some patients even progressed from no fibrosis to cirrhosis over 12 years although, regrettably, the authors could not rule out a false negative result for cirrhosis on the first biopsy by providing information on the presence of other signs of cirrhosis (low platelet count, ultrasonographic signs etc.). Second, worsening of fibrosis occurred even in patients with declining levels of transaminases thus making these markers unfit for follow-up. Third, the distinction between progressors and non-progressors could not be made at the time of the first biopsy on clinical or biological grounds, which unfortunately does not allow us to identify the subgroup of patients at high risk of fibrosis progression.
For all their methodological limitations, all these data are strikingly similar to other recently published reports. Fassio et al., in a concise but well described study of follow-up liver biopsies in 22 NASH patients, have shown that there are almost twice as many patients with fibrosis progression than with regression (32% vs. 18%) [6]. Interestingly, the proportion of patients with significant (>2 stages on the Ishak scale) variation of fibrosis was the same in both progressors and regressors, approximately half of the patients. Here again, progression of fibrosis occurred despite normalization of aminotransferase values and could not be confidently predicted by clinical or standard biological data or even by the other histological lesions present on the baseline biopsy. Particularly intriguing is the fact that fibrosis worsened in some patients despite their slight weight loss [6]. Harrison et al. [7], in a population of 22 NASH patients with repeat liver biopsies, report on worsening and improvement of fibrosis in the same proportion as Fassio et al. (32 and 18%, respectively) [6]. Worsening of fibrosis was significant as the proportion of those with bridging fibrosis doubled between the two biopsies [7].
While at first glance the results from these studies are coherent, a closer look reveals some inconsistencies with what we have already learned about liver fibrosis in NAFLD. Cross-sectional studies have unanimously identified age and insulin resistance (through its various phenotypic manifestations) as the strongest predictors of liver fibrosis in NASH [8], [9], [10]. Unexpectedly, none of these clearly correlate with fibrosis progression in longitudinal studies, and neither do dynamic changes of BMI correlate with changes in fibrosis scores. Even more worrisome, despite a strong association between necroinflammatory activity and fibrosis on cross sectional studies [9], [11], is that fibrosis progression in many patients occurs concurrently with an improvement in inflammation, ballooning and steatosis. We certainly still lack a conceptual framework for the understanding of fibrosis development in NAFLD. But could it be that some of the findings of these longitudinal studies are artifactual rather than real?
One reason for this could be sampling error of liver biopsy. Just like in any other liver disease [12], [13], [14], sampling error might be a significant problem in NASH [15], accounting for 20–30% of discordances within paired biopsies. If, for any histological lesion, the proportion of patients improving and worsening is the same, especially with a short time interval between biopsies, then sampling error might greatly account for the findings. This is just what seems to be happening in the most rigorous set of data on sequential liver biopsies available so far, the placebo arm of randomized controlled trials. Among 55 placebo-treated patients that had repeat liver biopsy in the US ursodeoxycholic acid trial [16], fibrosis improved in 22% of patients and worsened in 22%, inflammation improved in 23% and worsened in 20% and ballooning improved in 14% and worsened in 11%. The only ways to bypass the problem of sampling error would be to have an additional assessment of liver fibrosis independent of liver biopsy [17], [18] or to identify a distinct clinical and biological profile for progressors or regressors [19]. The first option was not available and the second option could not be achieved in the longitudinal studies under consideration. By carefully analyzing their data, Adams et al. [5] provide, however, an acceptable argument against the major confounding role of sampling error for their findings: the longer the time interval between biopsies, the greater the proportion of patients with worsening fibrosis.
There are nevertheless numerous other methodological shortcomings of these histological follow-up studies prompting caution in the interpretation of the results. One is the protocol itself of repeat liver biopsy. A prospective design such as in placebo treated arms of therapeutic trials has the caveat of too short a time interval for allowing progression to be seen in a slowly evolving disease. Longer time intervals can only be obtained in cohorts of patients with a clinical follow-up but then the problem becomes unknown biases in the decision to perform biopsies. If only those patients who were considered to have an aggressive course on clinical grounds underwent repeat liver biopsy, then the findings certainly would not apply to the whole cohort of NASH patients, most of whom have a slowly-evolving course. Another important problem is the lack of detailed information on the persistance of metabolic risk factors for NASH between the liver biopsies. As shown above, the mere measurement of BMI variation between biopsies is not enough. In the placebo arm of the urso trial a surprisingly high proportion of placebo treated patients (37%) had improved steatosis over a 2-year period while the BMI did not change [16]. Other parameters such as exercise, light or moderate alcohol consumption, control of lipid or serum glucose levels or control of arterial hypertension might be just as important in terms of impact on liver histology but impossible to assess from retrospective studies. The pathological evaluation itself brings its own set of challenges. For one thing, the current staging systems, which are mainly intended for diagnostic purposes, might not be best for detecting changes in fibrosis. Fassio et al. made the elegant observation that according to the scoring system used the end-result could be quite different: an increase in the fibrosis stage according to Ishak et al. [20] may go unnoticed when using the score of Brunt et al. [21]. Another important aspect lies with the histological definition of NASH. Here the studies differ because of the lack of consensual pathological criteria. Even when using rather stringent diagnostic criteria (hepatocyte ballooning and lobular inflammation in the context of steatosis), the overwhelming majority of the cohort from the Adams study had NASH and only three had bland steatosis [5]. Therefore their results apply to NASH patients rather than to the majority of NAFLD patients in which bland steatosis is a much more common finding and who are expected to have slower histological progression.
Despite all these shortcomings, what is the incremental advance in the understanding of the histological course of NASH gained from these recent studies? Liver fibrosis progression seems to be real in NASH, as it occurs at a higher rate than fibrosis regression, between a quarter and a third of patients over a 4 year or more time period. In some individuals this progression is of at least two fibrosis stages, which by all standards is clinically relevant. Unfortunately, we still cannot predict which patients are bound to have a progressive course by simple clinical or biological indexes or even by histological lesions on index biopsy. Whether bland steatosis does not progress as suggested by earlier histological studies [9], [22] and by clinical outcome data [11], [23] or if it does so at a lower pace, remains to be determined in large scale studies involving NAFLD instead of only NASH patients. Normalization of serum transaminases provides no reassurance as their evolution and that of fibrosis appear to be independent. Although major weight loss achieved through anti-obesity surgery does improve liver injury in NASH [24], minor weight loss (<10%) achieved through a soft-line dietary approach does not protect from fibrosis progression even when it is associated with normalization of serum transaminases. This casts a serious doubt over the prudent practical approach in the first line exploration of a suspected NASH patient. Many physicians delay histological exploration if minor weight loss with a decrease in transaminase values was achieved through dietary efforts. If weight loss is 10% or less in a patient that is still obese and retains all other risk factors for NASH, there may be no point in sticking to an expectant approach. Finally, when deciding to perform a follow-up liver biopsy, a 4-year or longer time interval between biopsies seems to be a reasonable approach.
Because of obvious limitations of the feasibility of liver biopsy in large scale studies, meaningful answers may only come from non-invasive serum markers [17], [18], [25] or imaging studies [26] when those will be fully validated. It is only then that the true nature of NASH might be revealed: a hidden fibroser that silently damages the liver.
Acknowledgements
Supported by the Association pour la Recherche sur les Maladies Hépatiques Virales (ARMHV).
References
- . The prevalence and etiology of elevated aminotransferase levels in the United States. Am J Gastroenterol. 2003;98:960–967
- . Pathogenesis of steatohepatitis. Best Pract Res Clin Gastroenterol. 2002;16:663–678
- . Is obesity a risk factor for cirrhosis-related death or hospitalization? A population-based cohort study. Gastroenterology. 2003;125:1053–1059
- . Diabetes increases the risk of chronic liver disease and hepatocellular carcinoma. Gastroenterology. 2004;126:460–468
- . The histological course of nonalcoholic fatty liver disease: a longitudinal study of 103 patients with sequential liver biopsies. J Hepatol. 2005;42:132–138
- . Natural history of nonalcoholic steatohepatitis: a longitudinal study of repeat liver biopsies. Hepatology. 2004;40:820–826
- . The natural history of nonalcoholic fatty liver disease: a clinical histopathological study. Am J Gastroenterol. 2003;98:2042–2047
- . Independent predictors of liver fibrosis in patients with nonalcoholic steatohepatitis. Hepatology. 1999;30:1356–1362
- Liver fibrosis in overweight patients. Gastroenterology. 2000;118:1117–1123
- . Nonalcoholic fatty liver disease: predictors of nonalcoholic steatohepatitis and liver fibrosis in the severely obese. Gastroenterology. 2001;121:91–100
- . Nonalcoholic fatty liver disease: a spectrum of clinical and pathological severity. Gastroenterology. 1999;116:1413–1419
- . Comparison of the histological lesions in tissue specimens taken from the right and left lobe of the liver in alcoholic liver disease. Gastroenterol Clin Biol. 1979;3:235–240
- Sampling variability and its influence on the diagnostic yield of percutaneous needle biopsy of the liver. Lancet. 1986;1:523–525
- Sampling error and intraobserver variation in liver biopsy in patients with chronic HCV infection. Am J Gastroenterol. 2002;97:2614–2618
- Sampling variability of liver biopsy in nonalcoholic fatty liver disease. Hepatology. 2004;40:237A
- Ursodeoxycholic acid for treatment of nonalcoholic steatohepatitis: results of a randomized trial. Hepatology. 2004;39:770–778
- Prediction of liver fibrosis in patients with features of the metabolic syndrome regardless of alcohol consumption. Hepatology. 2004;39:1639–1646
- Diagnostic value of biochemical markers (Fibrotest) for the prediction of liver fibrosis in patients with non-alcoholic fatty liver disease. Hepatology. 2003;38:729A
- Impact of pegylated interferon alfa-2b and ribavirin on liver fibrosis in patients with chronic hepatitis C. Gastroenterology. 2002;122:1303–1313
- Histological grading and staging of chronic hepatitis. J Hepatol. 1995;22:696–699
- . Nonalcoholic steatohepatitis: a proposal for grading and staging the histological lesions. Am J Gastroenterol. 1999;94:2467–2474
- . The natural history of nonalcoholic fatty liver: a follow-up study. Hepatology. 1995;22:1714–1719
- Long term prognosis of fatty liver: risk of chronic liver disease and death. Gut. 2004;53:750–755
- . Nonalcoholic fatty liver disease: improvement in liver histological analysis with weight loss. Hepatology. 2004;39:1647–1654
- Serum markers detect the presence of liver fibrosis: a cohort study. Gastroenterology. 2004;127:1704–1713
- Non‐invasive assessment of liver fibrosis by stiffness measurement in patients with chronic hepatitis C. in press Hepatology. 2005;
PII: S0168-8278(04)00547-1
doi:10.1016/j.jhep.2004.11.027
© 2004 Published by Elsevier Inc.
