Approximately one-third of the US population is presumed to have nonalcoholic fatty liver disease (NAFLD), with a similar prevalence reported in other parts of the world. Liver-related morbidity stems almost entirely from those individuals with nonalcoholic steatohepatitis (NASH). The prevalence of NASH in the United States is estimated to be 3%-5%, or roughly 9 million to 15 million persons, of whom up to 20% will develop cirrhosis. As a comparator, hepatitis C virus (HCV) infection, which is the leading indication for liver transplantation, had a prevalence of 1.6% between 1999 and 2002, representing approximately 4.1 million Americans.1 Although the incidence of HCV has plateaued and will possibly decrease over the long term, the incidence of NAFLD is on the rise. Thus, it is not surprising that NASH is predicted to surpass HCV as an indication for liver transplantation in the next 20 years.
The study by Bhala et al.2, in this issue of HEPATOLOGY, makes an important contribution to our understanding of the natural history of NASH. It is a large, multinational study that incorporates patients with advanced but compensated NASH or HCV. Somewhat not surprisingly, Bhala et al. showed that liver-related decompensation and incident hepatocellular carcinoma (HCC) were higher in patients with untreated or refractory HCV, compared to those with advanced NASH. This article also draws our attention to the development of HCC in patients with HCV or NASH who have not yet developed cirrhosis. There are currently sparse data addressing this in the literature, and this study further underscores the importance of better understanding the potential risk of HCC in patients without cirrhosis.
Several publications have provided insight into the natural history of NAFLD and NASH. In a large population study with a 15-year follow-up, NAFLD was independently associated with liver-related mortality as well as with cardiovascular (CV) disease and HCC, although the effect on the latter two conditions could not be dissociated from insulin resistance.3 To better put into perspective what will happen to our patients with advanced NASH, it is logical to compare it to HCV, a disease with a well-established natural history.
In a small prospective cohort study of Australian patients published in HEPATOLOGY nearly a decade ago, Hui and colleagues compared 23 patients with NASH-derived cirrhosis to 23 patients with untreated HCV-derived cirrhosis and 23 nonresponders with HCV-derived cirrhosis. The authors found that patients with NASH cirrhosis experienced less hepatic decompensation, but a similar mortality to their HCV cirrhosis counterparts.4 In this issue of HEPATOLOGY, Bhala et al. extend their findings to a multinational prospective cohort study that includes patients from Italy, the United States, the United Kingdom, and Australia. They investigated the long-term outcome of patients with NASH or HCV and advanced fibrosis (stage 3 or 4). They compared 247 patients with NASH to 264 patients with HCV (nonresponders or untreated) in the analysis and followed them for a mean of 85.6 and 74.9 months, respectively. The findings demonstrate that whereas the HCV cohort had more liver-related morbidity and incident HCC than the NASH cohort, rates of CV events and overall mortality were no different. Importantly, the current study differs from prior work in that it included patients with stage 3 fibrosis, in addition to those with compensated cirrhosis.
This is a timely study that sheds light on some aspects of the natural history of NAFLD. However, it has limitations that should be considered when interpreting the results. Given the heterogeneity of what we currently refer to as NASH, it is unlikely that any study would be generalizable to the entire NASH population. Ethnic differences in the susceptibility to develop NAFLD or progressive liver injury are well documented.5 For example, Hispanics and Asians (particularly the Indian subcontinent and southeast Asia), are at increased risk for advanced NASH, whereas African Americans are relatively protected despite the presence of similar metabolic risk factors.6-8 Although the current study is a multinational study from four countries (Australia, Italy, the United States, and the United Kingdom), 92% of the patients with NASH were Caucasian. Thus, as the authors concede, it is not clear whether these findings are applicable to other races. Although this is a potential weakness of the study, one could argue that given the heterogeneity of the NASH population at large, studies of ethnic-specific cohorts are important.
It is known that HCV treatment response deters the rate of decompensation and the development of HCC.9 Thus, the natural history of the HCV group chosen by Bhala et al. was at less risk of being influenced by external factors such as viral clearance. However, the disadvantage of this as a comparison group is that it is not representative of the population of patients with HCV-derived cirrhosis that one would encounter in clinical practice, which would include relapsers and those with sustained viral clearance. Patients in this cohort either did not respond to treatment or were untreated, namely representing the HCV subgroup with the poorest outcome. Were some patients not treated because they were sicker as suggested by a lower platelet count in the HCV-infected cohort? Moreover, compared to those that respond to treatment, HCV nonresponders have a higher risk of decompensation. Both NASH and HCV are common and often occur together. Unfortunately, data on concomitant fatty liver or insulin resistance was not systematically collected. The recent approval of the direct-acting antivirals for treatment of HCV raises many questions. Although the natural history of HCV as we know it is likely to change, drug interactions and side effects may limit broad use of direct-acting antivirals. Even so, with more patients achieving viral clearance, one could speculate that this will translate into less liver-related morbidity and mortality, including incident HCC in the years to come.
Although the natural evolution of disease in the HCV cohort was fairly predictable, the NASH cohort may have been influenced by several factors. In the article by Bhala et al., the use of metformin and statins was reported, but no mention was made of thiazolidenediones or vitamin E use. This is particularly relevant, because 50% of the NASH cohort had diabetes, and thiazolidenediones are commonly used in this setting. Furthermore, there are no data provided on the use of new medications or changes in body weight during the follow-up period. The short-term data on the effects of sustained weight loss, vitamin E, and pioglitazone are compelling, and future studies will be needed to determine their influence on the natural history of NASH.10, 11 While we await these data, we need to, at a minimum, consider the role of such factors as potentially altering the course of NASH for the better.
Given the observational nature of the study, the assessment of liver decompensation was left to the discretion of the investigators. Thus, the “development of varices” as a major outcome of hepatic morbidity could have been ascertained by screening endoscopy or by the development of a variceal hemorrhage. Screening practices for the detection of HCC or varices of the individual centers were not reported. Differences in how these endpoints were reported could be relevant, because the majority of patients with HCV came from Australia and Italy, and the majority of patients with NASH came from the United States and the United Kingdom, suggesting that different screening practices or definitions between the participating centers could have influenced outcome.
HCC does occur in patients with NASH-derived cirrhosis, and others have identified factors such as diabetes that confer an increased risk (1.3- to 2.4-fold) of HCC.12 Some have estimated the risk attributable to HCV to be four- to five-fold higher, and the data reported by Bhala and colleagues do support previous publications.13 Patients with advanced HCV unresponsive to treatment still comprise a large portion of the HCV-infected population, and this subgroup develops decompensated liver disease and HCC at an accelerated rate.14 However, in contrast, a recent prospective cohort study comparing patients with NASH-derived cirrhosis to an unselected group of patients with HCV-derived cirrhosis found that the annual incidence of HCC in patients with NASH compared to those with HCV was no different.15
The authors report equivalent CV outcomes in the two cohorts. Although this is intriguing, such an assertion may be premature. The data supporting the association between NASH and CV disease are fairly robust and have recently been reviewed.16 Prospectively collected data with well-defined cardiac endpoints and longer follow-up are needed to make a definitive statement about differences in CV outcomes between HCV and NASH.
Interestingly, despite comparing patients with advanced NASH to the subset of HCV patients with the worse prognosis, mortality rates were similar. These data suggest that over time, liver-related morbidity, including HCC and mortality due to NASH, may exceed that of the HCV population at large. The combination of the increasing burden of liver disease from NASH and more effective treatments for HCV (and NASH) are sure to change the landscape in how we approach our patients with cirrhosis arising from NASH or HCV. Future studies will be needed to redefine these dynamic populations and assess relative differences in risk as we enter this new era.