Editorial: NAFLD related hepatocellular carcinoma - increasing or not? With or without cirrhosis?

To evaluate nonalcoholic fatty liver disease (NAFLD)-related hepatocellular carcinoma (HCC) trends, Pais et al analyzed 323 HCC cases referred for hepatic resection to a single hospital in Paris, France during 1995-2014 (1). They observed a proportional increase of NAFLD-related HCC and decline in HCV-related HCC. Nearly one-third of HCC cases occurred in the absence of significant fibrosis, largely related to NAFLD and hepatitis B (63.0% and 28.0%, respectively). Because the study population was derived from a surgical referral center and had early HCC with well-compensated liver disease the proportion of non-cirrhotic cases and consequently NAFLD and HBV were likely to be overrepresented.

Population attributable fraction (PAF) is the proportion of cases with disease (i.e., HCC) that can be avoided by removing the underlying risk factor (i.e., NAFLD). PAF is calculated using the prevalence (how common) and risk estimate (how strong). HCV and HBV are uncommon but strong HCC risk factors in the general population; their PAFs are less than that of NAFLD, a highly prevalent but weak risk factor (2). Nevertheless population based studies demonstrate that HCV remains the leading cause of HCC in the US. In a national study among liver transplant recipients, NAFLD-related HCC as a liver transplant indication increased most rapidly over 2000-2012 (3), but HCV remained the leading HCC etiology. In the US Veterans Administration (VA) hospital system the annual proportion of NAFLD-related HCC remained stable during 2005-2010 (7.5%-12.0%), while those of HCV-related HCC increased from 61.0% to 74.9% (4).

How do we reconcile a high PAF for NAFLD-related HCC with studies showing a relatively low proportion of NAFLD among newly diagnosed HCC? First, the evidence of causality and its magnitude (NAFLD to HCC) is weak and inconsistent (5). Second, PAF calculation assumes that every HCC arising in a patient with NAFLD is caused by NAFLD; an assumption that is not likely to hold given that other risk factors are likely to coexist. Third, PAF does not consider for temporal lag between risk factor acquisition and HCC development. So it is possible that it will take several decades for the cohorts affected with NAFLD to develop HCC in large numbers.

HCC occurring in the absence of cirrhosis is the conventional exception to the rule, and mainly occurring in up to 15% of HBV related cases in endemic areas. NAFLD-related HCC disrupts this paradigm of HCC risk to a greater degree than HBV. Among new HCC cases without advanced fibrosis or cirrhosis, NAFLD is constitutes the largest etiological proportion of cases (6). This entity poses a challenge to our clinical practice paradigms that are based on HCC risk mediated through cirrhosis, as well as our understanding of HCC pathogenesis.

The rising prevalence of NAFLD (7) coupled with viral hepatitis treatment advances may result in a future increase in NAFLD-related HCC and reduction in HCV- and HBV-related HCC (8,9). While the Pais study does not provide convincing evidence that this has yet come to pass it reminds us of the increasingly important and potentially paradigm shifting entity of NAFLD-related HCC.