Natural history of NASH and HCC

Widespread unhealthy dietary habits associated with a sedentary lifestyle have made NAFLD the most frequent chronic liver disease worldwide, with a global prevalence of ~25%. Although NAFLD is mainly considered to be a benign disease, it can progress to severe liver fibrosis and hepatocellular carcinoma (HCC), with the latter found in non‐cirrhotic livers in about 40% of cases. Factors favouring the progression of liver disease in NAFLD are only partially understood. Male sex, older age and Caucasian ethnicity have frequently been identified as factors accelerating the progression of fibrosis in NAFLD, although data are not consistent. Host genetic variants appear to be very important, especially in the gene coding for the patatin‐like phospholipase domain‐containing 3 (PNPLA3), and they may also play a role in the development of HCC, independent of activity and the extent of liver damage. However, the most important factors affecting disease progression are found in the metabolic syndrome, that is, obesity, type 2 diabetes and arterial hypertension. This mini‐review will discuss the contribution of these factors to NAFLD‐associated morbidity, emphasizing the importance of preventive measures such as physical activity and weight control in view of the current pandemic of the metabolic syndrome.


| INTRODUC TI ON
In a recent review 1 pooling data from 86 studies including 8.5 million subjects from 22 countries, the global estimated prevalence of non-alcoholic fatty liver disease (NAFLD) is 25.24%, with a peak prevalence in the Middle East and Latin America, and the disease increases with age. Thus, NAFLD is the most prevalent liver disease worldwide. Moreover 51.34% of patients are obese, and 22.51% have diabetes. 1 This short review will focus on the risk factors of the progression of NAFLD to advanced fibrosis and HCC.

| THE PROG RE SS I ON OF LIVER FIB ROS IS IN NON -ALCOHOLI C FAT T Y LIVER D IS E A S E
NAFLD is a slowly progressive disease, which does not lead to endstage liver disease in most cases. It is difficult to estimate the rate of progression of NAFLD to severe fibrosis because of the high risk of selection bias when liver biopsies are performed in these patients, as well as the different criteria defining NAFLD in different cohorts. NAFLD may progress to non-alcoholic steatohepatitis (NASH), 2,3 which poses a significant risk of progression to hepatic

| Risk factors of progression
The risk factors for the progression of NAFLD to liver fibrosis are not fully understood. Although age has been associated with increasingly severe liver fibrosis in NASH patients, this is probably related to accumulating metabolic alterations occurring in the elderly with a long duration of liver disease. In a cross-sectional study in 432 patients, 26.8% with NASH and 17.4% with moderate to severe liver fibrosis, the independent predictors of moderate to severe fibrosis were male sex, Caucasian ethnicity, type 2 diabetes and increased liver enzymes, but not age. 5 The increased risk of fibrosis affects not only men, but also postmenopausal women and those who have undergone premature menopause. 6 Time from menopause is also directly associated with an increased likelihood of more severe liver fibrosis, after several adjustments (OR for each 5-year unit = 1.2, 95% CI 1.1-1.3, P = .002). 6 Ethnicity is also associated with NASH and fibrosis.
However, although Hispanic ethnicity may predispose to NASH, 7 it is unclear whether this is also associated with increased progression of liver disease, 8 which has been suggested in Caucasians. 5 The differences among ethnic groups may be partly explained by genetic polymorphisms, most importantly an isoleucine to methionine protein variant at position 148 of the patatin-like phospholipase domaincontaining 3 (PNPLA3) factor, a protein expressed in the liver and involved in lipid metabolism. 9 PNPLA3 variants are associated with the progression of liver fibrosis and cirrhosis independent from liver inflammation and NASH. 10 Another rarer polymorphism that occurs in TM6SF2 has also been shown to be associated with advanced liver fibrosis and cirrhosis independent from age, BMI, diabetes and PNPLA3 genotype. 11 Higher rates of fibrosis progression are observed in patients who are obese, or with type 2 diabetes. 2,3,12,13 In an important, well-designed prospective study in 52 patients with histologically proven NAFLD, who systematically underwent a biopsy after 36 months, 2 fibrosis progressed in 14 (27%), was stable in 25 (48%) and regressed in 13 (25%). It is important to note that these authors identified a decrease in BMI and waist circumference as independent predictors of stable liver disease activity and fibrosis. 2 Clinical and laboratory factors associated with the progression of fibrosis were also critically assessed in a meta-analysis of seven studies with paired biopsies. 14 Interestingly, this analysis found that the presence of arterial hypertension and low aspartate to alanine aminotransferase ratio at baseline liver biopsy were associated with the progression of fibrosis, but not age, ethnicity or diabetes. Higher grades of steatosis also seemed also to be more likely to lead to disease progression, but this was only based on two studies, where, oddly, the baseline severity of necroinflammation was not predictive of fibrosis progression.
The authors correctly emphasized the heterogeneity among studies, and the difficulty of firmly establishing the independent factors of fibrosis progression in NAFLD patients, where several confounders, especially related to lifestyle, may play a role.

| Weight reduction has a beneficial role
Reduction in body weight has been identified as a major predictor of stability or improvement in liver lesions, including the stage of fibrosis. [15][16][17] This has also been reported in patients undergoing bariatric surgery to manage obesity, 18 although in these cases the mechanisms leading to improved liver lesions may also involve changes in intestinal hormone secretion.
A large prospective study analysed the effects of lifestyle changes to reduce body weight in 293 patients with histologically proven NASH, followed for 1 year. 17 Paired liver biopsies were collected at the beginning and end of the observation period in 261 patients. Improvement or even the resolution of NASH was more frequent in those who lost weight. In particular, NASH resolved in 90% of the patients who lost ≥10% of their weight and fibrosis regressed in 45%. However, a subsequent multicentre cross-sectional study of 1058 biopsy-proven NAFLD patients, emphasized the Keypoints • Non-alcoholic fatty liver disease (NAFLD) is the most frequent chronic liver disease worldwide with an estimated global prevalence of ~25%.
• Factors leading to the progression of NAFLD are only partially understood, a limitation that is particularly serious when considering that 40%-50% of HCC cases occur in non-cirrhotic livers, raising the question of whom and how to screen for this complication.
importance of distinguishing obesity from metabolic status. 19 A metabolically healthy status was defined in that study by the absence of diabetes, low HDL, hypertriglyceridemia and arterial hypertension. The number of altered metabolic factors determined the risk of NASH and significant fibrosis. Interestingly, the latter was more frequently observed in the presence of adverse metabolic conditions in both obese and non-obese patients. Patients who were not obese but metabolically unhealthy more often had significant liver fibrosis than healthy obese patients (31.7% vs 11.4%, P < .0001). The authors mentioned that metabolically healthy obese patients are not entirely healthy and emphasized that the greatest impact on NASH and liver fibrosis is determined by a metabolically unhealthy status, which should be the real focus of patient counselling.

| PROG RE SS I ON TO H CC
NAFLD may progress to hepatocellular carcinoma (HCC). According to a large, recent meta-analysis, 1 the incidence of HCC among persons with simple NAFLD is very low, that is, 0.44 per 1000 personyears (range, 0.29-0.66), which is much less than that commonly reported for chronic hepatitis B or C. In patients with NASH, on the other hand, the annual HCC incidence rate increases by more than 10-fold, that is, 5.29 per 1000 person years (95% CI: 0.75-37.56), which is remarkable but still less than that reported for other chronic liver disorders. These incidence rates should be considered in relation to two major epidemiological observations which have a significant public health impact. Firstly, the global prevalence of NAFLD is higher than that of any other chronic liver disorder such as chronic viral hepatitis B or C. 1   in those with hepatitis B virus (HBV), and 11.1% in those with alcoholrelated liver disease. 27 The lack of cirrhosis in a significant proportion of NAFLD patients has been extensively confirmed and may occur in about 30%-50% of patients. [28][29][30][31] The evidence was particularly convincing in a multicentre observational prospective study performed in Italy, 31 which not only confirmed that ~50% of patients with NASH and HCC were non-cirrhotic, but that less than half of them had been identified during routine screening programmes, raising the troublesome issue of how to identify curative HCC. It is important to note that HCC in NAFLD patients were larger at diagnosis and more frequently presented with an infiltrative pattern. Despite this, after careful patient propensity score matching, the survival rates of NAFLD HCC were similar to those observed in patients with HCV infection.

| Risk factors of HCC
Nevertheless, the authors emphasize the importance of identifying patients with NAFLD without cirrhosis who could benefit from surveillance, and respond to curative therapies.
Several additional studies have assessed the risk factors associated with the development of HCC in NAFLD/NASH which may be relevant to design targeted screening strategies. The features associated with the metabolic syndrome besides age and male sex, such as diabetes, arterial hypertension and increased BMI, are strongly and frequently among the independent risk factors. One study assessed whether obesity is an independent factor of HCC in patients with advanced cirrhosis undergoing liver transplantation. 32 In the 19 271 patients included in the study with an incidence of HCC of 3.4% (n = 659), obesity was an independent factor for HCC in patients with alcohol-related cirrhosis (OR 3.2; 95% CI, 1.5-6.6; P = .002) and cryptogenic cirrhosis (OR, 11.1; 95% CI, 1.5-87.4; P = .02), but not in those with HBV, HCV or other liver disorders, suggesting that closer surveillance is needed in the former.
Diabetes has also been repeatedly reported to increase the risk of HCC in patients with NASH. In a landmark study from the Mayo Clinic, 33

| Physical activity has a beneficial effect
Because the metabolic syndrome has a confirmed, negative impact on the risk of HCC, one could hypothesize that the level of physical activity should have a beneficial effect. Indeed, this association was first reported in a mouse model characterized by the spontaneous development of NASH and HCC. 35 Mice fed standard chow were randomized to an exercise (motorized treadmill) or sedentary routine for 32 weeks. At the end of the observation period, the exercising mice had fewer, smaller liver tumours, although no effect was identified on steatosis or NASH. Exercise resulted into increased phosphorylation of AMPK and its substrate raptor, leading to decreased mTOR activity. These experimental observations have been elegantly confirmed by recent data in humans. A landmark multinational cohort study (the European Prospective Investigation into Cancer and Nutrition Cohort, the EPIC Study) assessed the impact of vigorous physical activity on different types of liver cancer in more than 470 000 persons followed for a median of 14.9 months. 36 The multivariate-adjusted HR for HCC was 0.55 (95% CI 0.38-0.80) in active compared to inactive participants. Waist circumference and BMI explained respectively ~40% and 30% of the association. Vigorous physical activity for at least 2 h/wk was also found to be beneficial to the risk of HCC (HR 0.50, 95% CI 0.33-0.76) compared to no vigorous activity, after taking into consideration potential confounders. Interestingly, the presence and level of physical activity was not correlated to the risk of other liver cancers, such as intrahepatic bile duct cancers or non-gallbladder extrahepatic bile duct cancers.
Similarly, treatment of diabetes can reduce the risk of HCC.
Indeed, this has been shown in several studies, especially in Asia.
A nationwide study performed in Taiwan analysed 47 820 diabetic patients. 37  Research Database showed a clear dose-dependent reduction in the risk of HCC in diabetics taking metformin. The risk reduction was greater than that in patients taking thiazolidinediones (51% vs 44% reduction). 37 Interestingly, statins also seem to protect from HCC. A nested case-control study from Korea in patients with newly diagnosed diabetes reported an adjusted OR of 0.36 (95% CI 0.22-0.60) in statin users vs non-users. Risk reduction was accentuated with an increase of cumulative defined daily doses. 38

| The role of genetics
Finally, genetics seems to play a role in the risk of HCC in NAFLD patients. The association of PNPLA3 variants and severe liver lesions has been confirmed in several studies. Heterogeneity at rs738409 has also been associated with the risk of HCC in patients with NAFLD. 39 Variant frequencies were significantly different between 100 NAFLD-HCC cases (CC = 28, CG = 43, GG = 29) and 275 NAFLDcontrols (CC = 125, CG = 117, GG = 33). After adjustment for age, gender, diabetes, BMI and cirrhosis, each copy of the rs738409 G variant led to an additive risk for HCC with an OR of 2.26. The risk of HCC among GG homozygotes was five-fold vs wild type CC. These important results suggest that genetic variants could help stratify at risk patients in whom strict surveillance for HCC could be beneficial, independent of the presence of cirrhosis. 40 More data and prospective studies are clearly needed to further confirm this hypothesis.

ACK N OWLED G EM ENT
The author thanks Dr Nicolas Goossens for critically reading the manuscript.

CO N FLI C T O F I NTE R E S T
The author has received research funding from Gilead Sciences, and advisor fees from Gilead Sciences, AbbVie and Merck.