Nonalcoholic fatty liver disease: Fibrosis portends a worse prognosis

The opinion or assertions contained herein are the private views of the authors and are not to be construed as official or reflecting the view of the US Department of the Army or the US Department of Defense. See Article on Page 1547 Advances in the treatment of chronic hepatitis C have led to a shift in focus to nonalcoholic fatty liver disease (NAFLD) as the fastest growing chronic liver disease of the Western world. More sedentary lifestyles, diets higher in processed carbohydrates, and burgeoning waistlines all have contributed to the obesity and diabetes epidemic and, subsequently, the meteoric rise in NAFLD prevalence. It is both fortunate and unfortunate that NAFLD is relatively easy to diagnose: steatosis on imaging with or without elevated liver‐associated enzymes and a negative serologic evaluation for other chronic liver disease picks up the majority of the 30%‐40% of the general population with NAFLD.1 However, the optimal management of NAFLD remains to be fully delineated as diet, weight loss, and exercise are difficult to achieve and sustain in the vast majority of NAFLD patients and Food and Drug Administration–approved therapies remain elusive. Furthermore, there appears to be a wide disparity in long‐term outcomes depending on the severity of disease. Once a diagnosis of NAFLD is made, liver biopsy remains the gold standard second step in the diagnostic algorithm to evaluate for nonalcoholic steatohepatitis (NASH). Pragmatically, physicians who see NAFLD patients are left in a quandary. If there is no Food and Drug Administration–approved therapy for NASH, why differentiate given the cost and potential risk of complications, however small they may be? Conventional teaching has suggested that those patients who meet histopathologic criteria for NASH are at the greatest risk of progression to cirrhosis as well as the development of hepatocellular carcinoma. However, this group of patients makes up only about 25% of those with NAFLD, and the vast majority of patients will have non‐NASH NAFLD. While most authorities in the field advocate a “NASH” or “non‐NASH NAFLD” expert pathologist read in order to help identify those patients at greatest risk for progressive disease, the NAFLD activity score (NAS) has also been used by some in clinical practice.5 The NAS, which uses a sum score of steatosis, ballooning, and lobular inflammation, was originally developed as a research tool and has been used in most large randomized‐controlled trials to measure “improvement” in disease with a particular therapy. A score greater than 4 is thought of as “probable NASH” but is not diagnostic, and NASH can occur with a lower NAS. Similarly, a score of 5 or 6 can be seen in patients who do not meet histopathologic criteria for NASH. Neither the NASH/non‐NASH NAFLD approach (just defining the presence or absence of disease) nor the NAS necessarily takes into account hepatic fibrosis, which others have suggested is at least equally important in determining both the prognosis and the success of a particular therapy.6 This has left many questions unanswered and placed clinicians in a difficult position both when deciding whether or not to perform a biopsy and, if a biopsy is done, what to do with the information. In this issue of Hepatology, Ekstedt et al.7 provide some insights with their large natural history study of 229 biopsy‐proven NAFLD patients followed for a mean of 26.4 years (±5.6, range 6‐33). One caveat that should be mentioned in this study, however, is that while patients were screened for other chronic liver diseases, it appears that only about 50% of the patients had hepatitis C excluded, so it is possible that a minority of patients with coexisting chronic hepatitis C are included in this cohort. There are a few key findings from this study that deserve to be highlighted. First, this study reiterates that NAFLD with fibrosis portended a worse prognosis, while the NAS is not helpful in natural history determination. Increased fibrosis stages 3 and 4, regardless of the NAS, had substantially increased mortality with a hazard ratio of 3.3 (95% confidence interval 2.27‐4.76, P < 0.001). Conversely, a high NAS in the absence of advanced fibrosis did not predict outcomes. Patients with a NAS of 5‐8 in the absence of advanced fibrosis (stage 0‐2) did not have increased overall mortality, with a hazard ratio of 1.41 (95% confidence interval 0.97‐2.06, P = 0.07). Those with a NAS 0‐4 and fibrosis stage 0‐2 were also not shown to have increased mortality compared to the reference population (hazard ratio = 1.13, 95% confidence interval 0.79‐1.60, P = 0.51). The caveat to this was two patients with a NAS of 0‐4 who died due to cirrhosis‐related complications, although both patients had stage 2 fibrosis. The authors explained these findings in the discussion, suggesting that the NAS was overly reliant on hepatic steatosis, to which it gives equal importance alongside ballooning and lobular inflammation. This is likely correct but also highlights a notable weakness of this study in that the NAS was used exclusively to stratify patients rather than a histopathologic diagnosis of NASH versus non‐NASH NAFLD. It is unclear what portion of the patients with a NAS of 0‐4 or 5‐8 and minimal (0‐2) fibrosis had NASH, and perhaps this can account for the lack of increased mortality. Second, it is notable that cardiovascular death was the most common cause of death in this population (43%), which exceeded the sum of the next five most common causes of death: nongastrointestinal malignancy (19%), hepatocellular carcinoma (5%), infection (5%), gastrointestinal malignancy (4%), and cirrhosis (4%). While this is consistent with previous work,8 what is interesting is that cardiovascular mortality was only seen in those with advanced fibrosis. Third, hepatocellular carcinoma occurred relatively infrequently in this long‐term follow‐up study, with only five total cases. However, all of these cases occurred in the setting of cirrhosis. This is in contrast with other data, mainly out of Asian countries, demonstrating that up to 50% of hepatocellular carcinoma in NAFLD occurs in patients without cirrhosis.9 This may reflect ethnicity and/or genetic predisposition differences between these two populations. Finally, it appears from this study that about 17% of the patients were lean, as defined by a body mass index of less than 25. It would be interesting to know if this group of patients had a similar natural history or whether they behaved differently. Unfortunately, this study does not address this group of patients separately. Where does this study leave clinicians and researchers? This study raises serious concerns for the ability of the NAS to predict long‐term outcomes and thus should probably not be used as a primary endpoint variable in clinical trials. While this study did not assess for this, determining NASH from non‐NASH NAFLD is still likely very relevant as patients with non‐NASH NAFLD seldom have significant fibrosis. However, a greater emphasis should be placed on NASH patients with moderate to advanced fibrosis as this study suggests that stage 2 or greater fibrosis infers the greatest risk to the NAFLD patient. While a liver biopsy is currently required to make the diagnosis, serum biomarkers and noninvasive imaging modalities are rapidly improving and may be able to distinguish NASH with moderate to advanced fibrosis from non‐NASH NAFLD in the near future. Until that time, it is still reasonable to biopsy individuals with multiple risk factors for NASH, focusing on the presence or absence of steatohepatitis as well as fibrosis stage to select patients best suited for clinical treatment trials or intensive lifestyle‐modification programs.