PREVALENCE AND RISK FACTORS OF SIGNIFICANT FIBROSIS IN PATIENTS WITH NONALCOHOLIC FATTY LIVER WITHOUT STEATOHEPATITIS
Pelusi S, Cespiati A, Rametta R, Rosso C, Baselli G, Dongiovanni P, Fracanzani AL, Maggioni M, Craxi A, Nobili V, Fargion S, Bugianesi E, Romeo S, Pihlajamaki J, Petta S, Valenti L.
ABSTRACT
Background&Aims: In patients with nonalcoholic fatty liver, steatohepatitis (NASH) is the major
predictor of clinically significant fibrosis. However, liver disease may be observed in the absence of NASH, but the prevalence and specific risk factors for this condition have not systematically been assessed.
Methods: We examined a multicenter cross-sectional liver biopsy cohort of 1,738 (44.9% of whom
with severe obesity) and a prospective cohort with sequential biopsies of 118 European patients with NAFLD.
Results: In the cross-sectional cohort, 132/389 (33.9%) of patients with clinically significant fibrosis
had no NASH, with 39 (10.0%) having no inflammation. The dissociation between NASH and fibrosis was more marked in patients with severe obesity (p<0.005). Each major histological feature of NAFLD, namely steatosis, ballooning, and lobular inflammation was independently associated with clinically significant fibrosis (p<0.001), which was also influenced by age, adiposity, fasting hyperglycemia and the PNPLA3 I148M variant. In patients without, but not in those with NASH, clinically significant fibrosis was associated with steatosis grade and the PNPLA3 I148M variant, whereas age, fasting hyperglycemia, ballooning and inflammation were common risk factors. In the prospective cohort, 16/47 (34.0%) of patients with clinically significant fibrosis did not have NASH at baseline. During the follow-up, faster fibrosis progression rate was associated with fasting hyperglycemia and with more severe steatosis in patients without (p=0.016), but not in those with NASH.
Conclusion: In patients with nonalcoholic fatty liver, at a definite time point clinically significant
fibrosis was observed in almost one third of cases in the absence of NASH. In these individuals, fasting hyperglycemia, severe steatosis and likely mild inflammation/ballooning and the PNPLA3 I148M variant may help identifying those at higher risk.
INTRODUCTION
Nonalcoholic fatty liver disease (NAFLD), defined in the presence of increased hepatic fat content not explained by at risk alcohol intake, has becoming the leading cause of liver disease [1]. NAFLD is epidemiologically associated with excessive adiposity, insulin resistance, physical inactivity, and qualitative alterations of the diet and the microbiota [2-4]. However, only a minor proportion of affected individuals progress to advanced liver disease. The factors that drive disease progression are likely heterogeneous, encompassing multiple hits leading to inflammation and development of nonalcoholic steatohepatitis (NASH) [5], with genetic factors playing an important role [6].
NASH is characterized by fatty liver, associated with both hepatocellular damage and lobular inflammation [7]. NASH has been linked with faster average progression rate of liver fibrosis [8], and lobular inflammation may be a marker of more severe disease even in patients without NASH [9]. However, there is a huge variability, and fibrosis frequently progresses even in patients without baseline NASH [8, 10, 11]. NASH is characterized by both more severe steatosis, and hepatocellular damage and inflammation. In selected cohorts of patients with aggressive disease, ballooning and lobular inflammation predicted fibrosis evolution [12, 13]. On the other hand, changes in metabolic risk factors and insulin resistance are also associated with disease evolution [8, 10, 11]. Furthermore, hepatic fat has now been linked with liver fibrosis progression [14, 15]. These data are in line with new genetic evidence that excessive hepatic lipid accumulation, drives secondary inflammation and disease progression [16].
Although progressive NAFLD appears as a heterogeneous disease, with variable involvement of genetic defects [6] and metabolic factors [17] in triggering fibrogenesis independently of inflammation, therapeutic studies are presently focused on the resolution of NASH in patients with severe histological activity [18]. However, several questions still remain unanswered. What’s the burden of NAFLD progressed to clinically significant fibrosis, which is not associated with NASH? What are the risk factors for having severe disease in patients without NASH? And in those without significant histological inflammation?
These questions have implications for modelling of disease burden [19], for the design of therapeutic trials, and for the applicability of results to the clinical practice. Furthermore, they may be relevant for the possible identification of a subset of patients characterized by different mechanism of disease progression, that is less strictly related to inflammation.
The aim of the study was therefore to examine the prevalence of NAFLD with clinically significant fibrosis not associated with NASH and/or inflammation in a large multicenter cohort, and the risk factors for disease progression in this subgroup, which were validated in a prospective cohort.
PATIENTS AND METHODS
Study cohorts
Part of the cross-sectional Liver Biopsy Cohort (LBC) has previously been described [20, 21]. Briefly, a total of 1,738 individuals of European descent were consecutively enrolled from Italian and Finnish referral centers. Inclusion criteria were liver biopsy for suspected NASH or severe obesity, presence of NAFLD and availability of clinical data and consent. Individuals with increased alcohol intake (men >30 g/day; women >20 g/day), viral and autoimmune hepatitis or other causes of liver disease were excluded. The study conformed to Declaration of Helsinki and was approved by the Institutional Review Board of the Fondazione Ca’ Granda IRCCS of Milan. The clinical features of the overall cross-sectional LBC and after stratification for the enrolment criterium is presented in Table 1.
To validate the results prospectively, we took advantage of a previously described multicenter cohort of 118 patients with NAFLD with serial liver biopsies followed-up at tertiary referral centers in Italy [11]. Their clinical features have previously been described in details [11]. Briefly, they were mostly middle-aged men or post-menopausal women, overweight or obese, with a high prevalence of metabolic alterations defining metabolic syndrome and/or altered liver enzymes. Median follow-up was 36 months (interquartile range 24–77). Fibrosis progression rate (FPR) was calculated by taking the ratio between the difference of fibrosis stage and the time (months) between the baseline and follow-up biopsy, and it was treated as a continuous variable. Forty-nine (41.5%) had NASH, 46 (39.0%) steatosis plus inflammation, and 23 (19.5%) simple steatosis.
Histological evaluation
Slides were coded and read by one expert pathologist at each center, who was unaware of patients’ identity and history. A minimum 15mm-length of the biopsy specimen or the presence of at least 10 complete portal tracts was required [22]. Steatosis was graded based on the percentage of affected hepatocytes as 0: 0-4%, 1: 5-32%, 2: 33-65%, and 3: 66-100%. Disease activity was assessed according to the NAFLD Activity Score (NAS) with systematic evaluation of hepatocellular ballooning and lobular inflammation; fibrosis was also staged according to the recommendations of the NAFLD clinical research network [23]. NASH was diagnosed when steatosis, hepatocellular ballooning and lobular inflammation were all present. The scoring of liver biopsies was performed by independent pathologists unaware of patients’ status and genotype. The concordance between pathologists within this cohort was very good for fibrosis and good for steatosis with a coefficient of inter-observer agreement for fibrosis, steatosis grade, lobular inflammation and ballooning of 0.89, 0.76, 0.60 and 0.55 [24].
Genotyping
When DNA samples and consent were available, the study cohorts were genotyped for rs738409 C>G (PNPLA3 I148M) by TaqMan 5’-nuclease assays (Life Technologies, Carlsbad, CA), as previously described [11, 20, 21].
Statistical analysis
For descriptive statistics, continuous variables were presented as mean and standard deviation (for approximately normally distributed variables) or median and interquartile range (for highly skewed variables). Categorical variables were presented as number and proportion. Genetic
association were tested assuming an additive model. Analyses were performed using generalized linear models: linear regression models were fit to analyse continuous traits (FPR), logistic regression for binary traits (clinically significant fibrosis). Models were adjusted for confounding factors (including recruitment criterium), as specified. Variables with skewed distributions were logarithmically transformed before entering the models.
Statistical analyses were carried out using the JMP 12.0 (SAS Institute, Cary, NC, USA) and R statistical analysis software version 3.3.2 (http://www.R-project.org/). P-values <0.05 were considered statistically significant.
RESULTS
Clinical features of the cross-sectional Liver Biopsy Cohort
The clinical features of patients stratified according to the enrolment criterium is shown in Table 1. Most of patients were middle-aged adults, about half of them females, 60% obese. Clinical and metabolic features differed according to the enrolment criterium. Liver clinic as compared to severely obese patients had expectedly lower adiposity, were more frequently males, but less frequently affected by type 2 diabetes (T2D) and arterial hypertension, and had lower total cholesterol. Conversely, as expected they had more severe liver damage and were more frequently carriers of the PNPLA3 I148M risk variant. Of the overall cohort, 32.9% of patients had NASH, 22.4% clinically significant fibrosis.
Impact of NASH and inflammation on fibrosis in the cross-sectional cohort
The prevalence of clinically significant fibrosis in patients with and without NASH was 257/571, 45.0% vs. 132/1165, 11.3%, respectively (p=2*10-53). The frequency distribution of NASH, steatosis
associated with inflammation, and simple steatosis according to fibrosis stage is presented in Figure 1A, left panel. Expectedly, there was a definite trend for an increasing prevalence of NASH with worsening of fibrosis (p=4*10-83). Furthermore, while the majority of patients without fibrosis had
simple steatosis, most of patients who had developed fibrosis showed some sign of lobular inflammation. However, as much as 132/389, 33.9% of patients with clinically significant fibrosis had no NASH, and 39/389, 10% of patients with clinically significant fibrosis had no sign of histological inflammation.
NASH was more frequently observed in the liver clinic patients, but the proportion of patients with clinically significant fibrosis in the absence of detectable lobular inflammation at the time of biopsy was higher in severely obese individuals, being observed in 35/102, 19.2% of them (p=0.0002 for interaction; shown in Figure 1A, right panel). Similarly, NASH was more prevalent in patients with fasting hyperglycemia than in those with normal glucose tolerance, and lack of inflammation was observed in 4/24, 16.7% of cirrhotic patients with normal glucose tolerance (Figure 1B). Remarkably, despite the reduced sample size, there was no association between NASH and fibrosis in developmental age patients (Figure S1A), while the association between NASH and fibrosis was not significantly influenced by gender (Figure S1B), nor by carriage of the PNPLA3 I148M variant (Figure S1C).
Histological predictors of significant fibrosis development in the cross-sectional cohort
NASH was characterized by the presence of the most severe grade of steatosis, hepatocellular ballooning, and lobular inflammation, whereas steatosis plus inflammation had intermediate steatosis and inflammatory grade, but by definition no ballooning, and simple steatosis had the lowest grade of steatosis and by definition no inflammation (Figure 2, upper panel).
The clinical features of patients stratified by the presence of NASH or of lobular inflammation is presented in Table 2, upper panel. Despite patients with NASH or inflammation were less frequently severely obese, which was associated with lower prevalence of female gender, there was a progressive increase in the prevalence of fasting hyperglycemia (T2D or IFG) and higher triglycerides with increasing disease activity (from simple steatosis, to steatosis plus inflammation and next NASH), as well as a progressive increase in the frequency of homozygosity of the PNPLA3 I148M variant.
Expectedly, aminotransferases and the prevalence of more severe stages of fibrosis also increased with disease activity.
The independent predictors of clinically significant fibrosis in the overall cohort, and in patients stratified according to the modality of enrolment is presented in Table 3. Clinically significant fibrosis was independently associated with each feature of NAFLD activity (steatosis, hepatocellular ballooning and lobular inflammation) after adjustment for confounders, both in the overall and in the sub-cohorts, although it was most strictly related to lobular inflammation. Besides NASH features, the presence of fasting hyperglycemia and the PNPLA3 I148M variant increased the risk independently of each histological feature of the disease, the independent impact of the latter being larger in severely obese. Furthermore, clinically significant fibrosis was associated with older age and adiposity in the overall cohort and liver clinic, but not in severely obese individuals. Conversely, female sex was mainly protective in severely obese individuals, despite mean age was not different in females according to recruitment criterium (44.519 years in severely obese vs. 45.29.9 years in liver clinic patients, p=0.58).
Predictors of significant fibrosis according to NASH in the cross-sectional cohort
The frequency distribution of the histological grade of steatosis, hepatocellular ballooning and lobular inflammation between patients who had develop or not clinically significant fibrosis, stratified by the presence of NASH, steatosis plus inflammation, or simple inflammation, is shown in Figure 2, bottom panel. In patients with NASH, steatosis grade tended to be severe independently or fibrosis. On the other hand, steatosis was more severe in those who developed clinically significant fibrosis in the subgroups of patients with steatosis and inflammation and in those with steatosis alone. More severe ballooning was associated with clinically significant fibrosis in patients with NASH and simple steatosis, whereas more severe inflammation in those with NASH and steatosis plus inflammation.
The independent predictors of clinically significant fibrosis in patients stratified according to the presence of NASH is presented in Table 4. We focused on specific histological features, metabolic comorbidities and genetic factors in order to assess the factors potentially determining fibrosis development. In those with NASH, clinically significant fibrosis was associated with more severe ballooning and inflammation, and with older age and fasting hyperglycemia independently of these features. In those without, clinically significant fibrosis was associated with more severe steatosis, inflammation, and fibrosis, and with older age, faster hyperglycemia, and the PNPLA3 I148M variant independently of these features. Therefore, steatosis grade and the PNPLA3 I148M variant seemed to have a larger independent role in fibrosis progression in patients not fulfilling NASH diagnostic criteria at liver biopsy. Even at univariate analysis, the PNPLA3 I148M variant had a larger effect size on clinically significant fibrosis in patients without (estimate +0.552, 0.132, p=3*10-5), than in those with
NASH (estimate +0.239, 0.118, p=0.003). Results were similar when patients were further stratified by the presence of severe obesity (Table S1). In particular, steatosis grade and the PNPLA3 I148M variant were independently associated with clinically significant fibrosis in patients without, but not in those with NASH in both sub-cohorts.
When those patients without NASH were stratified by the presence of lobular inflammation, in those with inflammation clinically significant fibrosis was associated with steatosis and inflammatory grade, and with older age, male sex, fasting hyperglycemia and the PNPLA3 I148M variant independently of these features. In those without inflammation, clinically significant fibrosis was associated with more severe steatosis and ballooning, and with adiposity and the PNPLA3 I148M variant independently of these features.
Predictors of fibrosis evolution according to NASH in the prospective cohort
The clinical features at baseline of patients included in the prospective cohort stratified by liver disease activity (from simple steatosis, to steatosis plus inflammation and next NASH) is shown in Table 2, bottom panel. Factors associated with a progressive increase in disease activity were again fasting hyperglycemia (T2D or IFG), and homozygosity for the PNPLA3 I148M variant. Expectedly, the prevalence of more severe stages of fibrosis also increased with disease activity. Interestingly however, disease activity at baseline was not associated with fibrosis progression rate (FPR) during the follow-up.
The independent predictor of FPR in the overall prospective cohort and in patients stratified according to disease activity (from simple steatosis, to steatosis plus inflammation and next NASH) is shown in Table 5. In this analysis genetic factors were not considered because not available in the majority of patients. In the overall cohort, faster FPR was associated with fasting hyperglycemia, but not with more severe features of liver damage. After stratification for the presence of NASH, in patients with NASH faster FPR was associated with male sex and more severe ballooning. In those without NASH, faster FPR was again independently associated with the metabolic features of the disease, namely fasting hyperglycemia and steatosis grade.
In patients without baseline NASH who had fibrosis progression at follow-up, we observed a higher rate of conversion of histology to NASH as compared to non-progressors (6/15, 40.0% vs. 8/54, 14.8%; p=0.043), while the variation in BMI, liver enzymes, and incidence of type 2 diabetes were not significantly different (not shown).
DISCUSSION
In this study, we examined the prevalence and risk factors of NAFLD with clinically significant fibrosis in patients without a NASH diagnosis in a large European multicenter cross-sectional cohort, and validated the findings prospectively.
Strikingly, we found that more than one third of patients with clinically significant fibrosis had no NASH, and more than one in ten had no sign of histological inflammation. Although the association was influenced by severe obesity, in that individuals referred for this condition had a higher prevalence of clinically significant fibrosis not associated with NASH nor with inflammation, still 27% of patients referred to a liver clinic were found to have clinically significant fibrosis had no NASH, and 7% no inflammation. The fraction of patients without inflammation was maintained in those with cirrhosis, possibly due to burnt-out disease in a fraction of these individuals [25]. However, the prevalence of clinically significant fibrosis not associated with NASH was higher in patients with fibrosis stage 2 and 3 than in those with cirrhosis, and in patients without NASH clinically significant fibrosis was associated with more severe steatosis, which would not be in line with burnt-out disease. However, we cannot exclude that these individuals were experiencing a phase of temporary remission of disease activity at the time of biopsy, leading to reduced ballooning and inflammation, but without improvement in fibrosis. Notwithstanding, the prevalence of clinically significant fibrosis not associated with NASH observed in this study is several-fold higher than that estimated in recent epidemiological projections of the burden of NAFLD (at around 5%) [19], suggesting that the temporal trends of disease evolution may have been currently underestimated [19].
The higher prevalence of clinically significant fibrosis in the absence of NASH in severely obese patients may be explained by a direct impact of insulin resistance in disease progression independent of inflammation [17, 26]. Alternatively, it could be speculated that a more prolonged fasting before undergoing bariatric surgery may have been positively affected disease activity in this subgroup, partially masking the contribution of inflammation. Indeed, little is known concerning the impact of the duration of fasting as well as of the quality of diet in the period immediately preceding evaluation of liver damage (either by liver biopsy or by noninvasive approaches) on disease activity. Such information would allow a better interpretation of the results and standardization of the procedures.
Secondly, we found that each of the individual major features of the disease, namely steatosis, ballooning, and lobular inflammation, was independently associated with clinically significant fibrosis. These data suggest that they reflect different pathophysiological aspects of the disease, e.g. fat accumulation, oxidative stress and inflammation, which contribute to disease progression, and confirm that accumulation of lipotoxic species may have a role independent of inflammatory infiltrates [16]. Furthermore, we observed that the impact of fasting hyperglycemia and severe insulin resistance was independent of these histological features [17, 27], which may be mediated by glucose toxicity or alteration in insulin and other fibrogenic signaling pathways [17, 26, 28]. In addition, it is worth noting that the PNPLA3 I148M variant was associated with clinically significant fibrosis independently of the impact on histological fat and inflammation, in keeping with a direct effect on this mutation on fibrogenesis by acting on retinol/lipid metabolism in hepatic stellate cells [29-32].
An important observation is that the risk factors of clinically significant fibrosis were different according to the presence of NASH. Indeed, hepatocellular ballooning and inflammation, as well as fasting hyperglycemia and ageing predicted this outcome irrespective of the presence of NASH. On the other hand, an independent impact of steatosis grade, as well as of male sex, and of the PNPLA3
I148M variant could only be observed in patients without NASH. In keeping, prospective cohort studies showed that in patients with severe NASH disease progression was predicted by ballooning or inflammation [12, 13], while in unselected cohorts by steatosis grade [14, 15]. This may be explained by the fact that patients with NASH have already severe fat accumulation and activation of fibrogenic pathways independently of these risk factors. Nevertheless, evaluation of these additional risk factors may help selecting patients at risk of progression despite the absence of NASH.
Most importantly, the specific association between steatosis grade and fibrosis progression rate in patients without NASH was replicated in an independent prospective cohort, confirming the potential relevance of this histological feature of NAFLD as a prognostic marker in this subgroup. Here, we also confirmed that during follow-up fibrosis progression occurred in patients without baseline NASH in as much as 50% of cases, and in 10% of cases in patients without baseline histological inflammation. However, in 40% of cases without baseline NASH who had fibrosis progression, this was associated with development of NASH during follow-up, suggesting that they may be affected by a “remitting-relapsing” form of disease, which might have led to underestimation of the average disease activity at the time of baseline histological evaluation. Heterogeneity of liver damage distribution and sampling variability should also be considered as possible explanations.
Despite it included a large series of patients with validation in a prospective cohort, this study has limitations. These include the retrospective design, with lack of possibility to central revision of all slides. However, histological evaluations were carried out by expert pathologists at tertiary referral centers, receiving regular feedbacks from central histological readings in randomized controlled trials, and the concordance for staging fibrosis and steatosis in a training set of slides was good [24]. We do acknowledge that the concordance for evaluation of inflammation and ballooning, which are key for diagnosis NASH, was only moderate [24]. Notwithstanding, in the present study these histological features turned out to represent the most accurate predictors of clinically significant fibrosis in all subgroups, rendering unlikely that inaccurate staging may account for the present findings. Furthermore, this limitation was partially compensated by the large sample size, suggesting that results reflect the clinical challenges in identifying at risk patients in referral center. Finally, genetic data were not available in a subgroup, and results may not be applicable to non-European individuals.
In conclusion, in a large cross-sectional multicenter cohort of patients with NAFLD we found that at a definite time point clinically significant fibrosis is not associated with NASH in about one third of cases, and in one in ten is not even associated with signs of inflammatory activity. In addition, the dissociation between NASH and fibrosis is more marked in severely obese individuals. Importantly, the severity of steatosis and the PNPLA3 I148M variant predicted the presence of clinically significant fibrosis, and steatosis predicted fibrosis progression specifically in patients without NASH. These findings suggest that a higher than expected fraction of patients with NAFLD may evolve towards clinically advanced fibrosis, possibly due to fluctuating disease activity, which is not easily captured at a definite time point, and/or the existence of mechanisms of disease progression that may be partially different in this subgroup.
TABLES
Table 1. Demographic, anthropometric, and clinical features of the cross-sectional Liver Biopsy Cohort
(n=1,738). Overall (n=1,738) Liver clinic (n=958, 55.1%) Severe obesity (n=780, 44.9%) p value Sex, F 860 (49.4) 299 (31.2) 561 (71.9) <0.0001 Age, years 43.7±14.8 42.8±17.5 44.7 ± 10.6 0.99 BMI, Kg/m2 34.6±9.0 27.8±4.0 42.9±5.7 <0.0001 Obesity 1047 (60.2) 267 (27.9) 780 (100) <0.0001 T2D/IFG, yes 466 (26.8) 228 (23.8) 238 (30.5) 0.0017 Hypertension, yes 448 (33.3) 272 (29.3) 176 (42.5) <0.0001 Total cholesterol, mg/dl 191±44 195±44 185±44 <0.0001 Triglycerides, mg/dl 139±77 138±80 140±73 0.55 HDL cholesterol, mg/dl 48±14 48±14 48±14 0.35 ALT, IU/L 42 {24-69} 54 {34-80} 30 {19-48} <0.0001 AST, IU/L 29 {21-43} 34 {24-47} 22 {17-33} <0.0001 NASH, yes 571 (32.9) 422 (44.0) 149 (19.2) <0.0001
Clinically significant fibrosis, stage F2-F4
389 (22.4) 287 (30.0) 102 (13.1) <0.0001
Severe fibrosis, stage F3-F4 174 (10.0) 137 (14.3) 37 (4.8) <0.0001
PNPLA3, 148M/M * 234 (13.8) 168 (17.7) 66 (8.8) <0.0001
Values are reported as mean ± SD, median {IQR}, or number (%), as appropriate. BMI: body mass index. IFG: impaired fasting glucose. T2D: type 2 diabetes. * PNPLA3 I148M genotype was available in a subset of patients (n=1,698). Characteristics of participants were compared using linear regression model or logistic regression model (as required).
Table 2. Clinical features of the study cohorts stratified by the presence of NASH, steatosis associated
with lobular inflammation or simple steatosis.
Cross-sectional LBC cohort N= NASH 571 (32.8) Steatosis plus inflammation 465 (26.8) Simple steatosis 702 (40.4) p value Sex, F 230 (40.3) 197 (42.4) 433 (61.7) <0.0001 Age, years 44.7±17.0 40.8±16.8 44.7 ± 10.7 <0.0001 BMI, Kg/m2 32.2±7.7 33.5±9.2 37.3±9.0 <0.0001 Obesity 292 (51.1) 247 (53.1) 508 (72.4) <0.0001 T2D/IFG, yes 211 (37.0) 119 (25.6) 136 (19.4) <0.0001 Hypertension, yes 166 (34.2) 123 (30.1) 159 (35.4) 0.23 Total cholesterol, mg/dl 191±45 196±45 187±43 0.0036 Triglycerides, mg/dl 147±77 140±79 131±74 0.0017 HDL cholesterol, mg/dl 47±14 48±14 48±14 0.65 ALT, IU/L 59 {38-88} 48 {26-76} 30 {19-45} <0.0001 AST, IU/L 39 {28-54} 30 {21-46} 22 {17-31} <0.0001
Clinically significant fibrosis, stage F2-F4
257 (45.0) 93 (20.0) 39 (5.6) <0.0001
Severe fibrosis, stage F3-F4 139 (24.3) 20 (4.3) 15 (2.1) <0.0001
PNPLA3, 148M/M * 107 (19.0) 74 (16.2) 53 (7.8) <0.0001 Prospective cohort N= NASH 49 (41.5) Steatosis plus inflammation 46 (39.0) Simple steatosis 23 (19.5) p value Sex, F 21 (42.9) 15 (32.6) 9 (39.1) 0.58 Age, years 47.9±13.2 46.4±12.0 47.5 ± 12.5 0.80 BMI, Kg/m2 30.0±4.1 31.1±1.0 30.6±1.4 0.72 Obesity 24 (49.0) 19 (41.3) 10 (43.5) 0.75 T2D/IFG, yes 28 (58.3) 14 (30.4) 4 (17.4) 0.001 Hypertension, yes 20 (40.8) 10 (21.7) 8 (34.8) 0.13 LDL cholesterol, mg/dl 115±36 125±43 133±33 0.16 Triglycerides, mg/dl 136±70 142±86 109±47 0.21 HDL cholesterol, mg/dl 48±15 46±12 50±15 0.50 ALT, IU/L 57 {42-95} 51 {26-82} 35 {21-63} 0.78 AST, IU/L 37 {27-62} 34 {22-40} 24 {19-39} 0.52
Clinically significant fibrosis, stage F2-F4
31 (63.3) 15 (32.6) 1 (4.4) <0.0001
Severe fibrosis, stage F3-F4 19 (38.8) 8 (17.4) 0 <0.0001
FPR, stages/year -0.0110.055 +0.0050.040 +0.0020.011 0.16
Fibrosis progression, yes 15 (30.6) 12 (26.1) 3 (13.0) 0.13
Values are reported as mean ± SD, median {IQR}, or number (%), as appropriate. BMI: body mass index. IFG: impaired fasting glucose, T2D: type 2 diabetes, RAAA: renin angiotensin aldosterone axis, FPR: fibrosis progression rate. PNPLA3 I148M genotype was available in a subset of patients (* n=1,698; ** n=89). Characteristics of participants were compared using linear regression model or logistic regression model (as required).
Table 3. Independent predictors of clinically significant fibrosis (stage F2-F4) in the overall
cross-sectional LBC cohort and in patients stratified according to modality of enrolment (liver clinic vs. severe obesity).
OVERALL LIVER CLINIC SEVERE OBESITY
Estimate SE p value Estimate SE p value Estimate SE p value
Sex, F -0.143 0.078 0.061 +0.014 0.009 0.88 -0.400 0.106 0.0018 Age, years +0.037 0.005 7*10-14 +0.040 0.005 7*10-13 +0.014 0.011 0.23 BMI, Kg/m2 +0.038 0.015 0.009 +0.054 0.023 0.021 +0.022 0.020 0.24 T2D/IFG, yes +0.480 0.079 1*10-9 +0.503 0.098 8*10-8 +0.350 0.129 0.0067 Steatosis, grade +0.399 0.079 3*10-7 +0.338 0.098 6*10-4 +0.547 0.139 9*10-5 Ballooning, grade +0.608 0.103 4*10-9 +0.571 0.113 5*10-7 +0.776 0.266 0.0036
Lobular inflammation, grade +0.779 0.081 2*10-21 +0.733 0.098 2*10-13 +0.772 0.266 2*10-7
PNPLA3, I148M alleles +0.326 0.101 0.0001 +0.231 0.125 0.064 +0.550 0.183 0.0036
BMI: body mass index. IFG: impaired fasting glucose. T2D: type 2 diabetes. * PNPLA3 I148M genotype was available in a subset of patients (n=1,698). Characteristics of participants were compared using linear regression models. Results in the overall cohort were adjusted for enrolment criterium (liver clinic vs. severe obesity) and the covariates shown in the table.
Table 4. Independent predictors of clinically significant fibrosis (stage F2-F4) in the cross-sectional LBC
patients stratified according to disease activity: from simple steatosis, to steatosis associated with lobular inflammation, and next to NASH.
NASH NO
NASH
OVERALL Steatosis plus
inflammation
Simple steatosis
Estimate SE p value Estimate SE p value Estimate SE p value Estimate SE p value
Sex, F -0.074 0.110 0.51 -0.279 0.116 0.016 -0.306 0.145 0.035 -0.171 0.196 0.38 Age, years +0.042 0.007 2*10-9 +0.024 0.007 0.001 +0.023 0.008 0.004 +0.033 0.017 0.051 BMI, Kg/m2 +0.029 0.023 0.21 +0.035 0.020 0.087 +0.014 0.027 0.61 +0.070 0.030 0.021 T2D/IFG, yes +0.449 0.113 9*10-5 +0.387 0.113 6*10-4 +0.413 0.143 0.004 +0.318 0.196 0.10 Steatosis, grade +0.033 0.122 0.79 +0.585 0.111 9*10-8 +0.337 0.186 0.010 +0.694 0.183 2*10-4 Ballooning, grade +0.365 0.115 0.001 +1.342 0.432 0.002 - - - +1.556 0.459 7*10-4 Lobular inflammation, grade +0.710 0.109 8*10-11 +0.718 0.134 9*10-8 +0.596 0.140 2*10-5 - - -PNPLA3, I148M alleles +0.211 0.143 0.14 +0.442 0.149 0.003 +0.359 0.183 0.038 +0.568 0.248 0.022
BMI: body mass index. IFG: impaired fasting glucose. T2D: type 2 diabetes. * PNPLA3 I148M genotype was available in a subset of patients (n=1,698). Characteristics of participants were compared using linear regression models. All results were adjusted for enrolment criterium (liver clinic vs. severe obesity) and the covariates shown in the table.
Table 5. Independent predictors of fibrosis progression rate (FPR) in the prospective cohort patients
(n=118) stratified according to disease activity: from simple steatosis, to steatosis associated with lobular inflammation, and next to NASH.
OVERALL NASH NO NASH
Estimate SE p value Estimate SE p value Estimate SE p value
Sex, F +0.001 0.004 0.74 -0.014 0.007 0.032 +0.001 0.004 0.30
Age, years +0.001 0.001 0.35 +0.001 0.001 0.31 +0.001 0.001 0.93
T2D/IFG, yes +0.013 0.005 0.014 +0.009 0.009 0.38 +0.013 0.005 3.6*10-5
Steatosis, grade >1 +0.014 0.010 0.22 +0.005 0.015 0.89 +0.021 0.010 0.016
Ballooning, grade >0 -0.019 0.009 0.047 +0.036 0.014 0.012 +0.015 0.017 0.36
Lobular inflammation, grade >1 +0.010 0.010 0.55 +0.024 0.016 0.14 +0.001 0.001 0.66
BMI: body mass index. IFG: impaired fasting glucose. T2D: type 2 diabetes. Characteristics of participants were compared using linear regression models. All results were adjusted for the covariates shown in the table, and for baseline fibrosis stage and use of renin-angiotensin receptor axis modulators, which influenced FPR in the overall cohort [11].
SUPPLEMENTARY TABLES
Table S1. Independent predictors of clinically significant fibrosis (stage F2-F4) in the cross-sectional
LBC patients stratified according to disease activity (from simple steatosis, to steatosis associated with lobular inflammation, and next to NASH) and the presence of severe obesity.
Liver clinic Severe obesity
NASH No NASH NASH No NASH
Estimate SE p value Estimate SE p value Estimate SE p value Estimate SE p value
Sex, F -0.092 0.139 0.49 -0.073 0.157 0.064 -0.314 0.223 0.16 -0.461 0.163 0.004 Age, years +0.045 0.008 1*10-8 +0.028 0.008 0.001 +0.001 0.019 0.96 +0.020 0.015 0.18 BMI, Kg/m2 +0.098 0.034 0.004 -0.015 0.037 0.69 -0.054 0.047 0.25 +0.054 0.024 0.030 T2D/IFG, yes +0.515 0.139 2*10-4 +0.477 0.164 0.003 +0.356 0.222 0.10 +0.320 0.165 0.053 Steatosis, grade +0.072 0.146 0.62 +0.409 0.152 0.007 +0.147 0.255 0.57 +0.821 0.175 3*10-6 Ballooning, grade +0.284 0.129 0.028 +1.239 0.432 0.005 +0.592 0.283 0.037 +0.523 0.232 0.024 Lobular inflammation, grade +0.634 0.131 1*10-6 +0.733 0.168 1*10-5 +0.751 0.220 7*10-4 +0.743 0.232 0.001 PNPLA3, I148M alleles +0.078 0.163 0.65 +0.437 0.196 0.026 +0.359 0.183 0.038 +0.497 0.239 0.024
BMI: body mass index. IFG: impaired fasting glucose. T2D: type 2 diabetes. * PNPLA3 I148M genotype was available in a subset of patients (n=1,698). Characteristics of participants were compared using linear regression models.
FIGURES LEGENDS
Figure 1. Prevalence of NASH and of lobular inflammation not fulfilling NASH criteria, vs. simple
steatosis according to fibrosis stage in the overall cohort (n=1,738) and in patients stratified by recruitment criterium (referral for liver disease or for severe obesity) (A). The same analysis was conducted in patients stratified by the presence of fasting hyperglycemia (type 2 diabetes: T2D or impaired fasting glucose: IFG) (B).
Figure 2. Frequency distribution of severity of steatosis, hepatocellular ballooning, and lobular
inflammation according to diagnosis of NASH, steatosis plus inflammation and simple steatosis in the overall cohort and in patients stratified according to the presence of clinically significant fibrosis.
SUPPLEMENTARY FIGURES LEGENDS
Figure S1. Prevalence of NASH and of lobular inflammation not fulfilling NASH criteria, vs. simple
steatosis according to fibrosis stage in patients stratified by developmental vs. adult age (A), sex (B), and PNPLA3 I148M variant status (available in n=1,698) (C).
REFERENCES
[1] Younossi ZM, Koenig AB, Abdelatif D, Fazel Y, Henry L, Wymer M. Global epidemiology of nonalcoholic fatty liver disease-Meta-analytic assessment of prevalence, incidence, and outcomes. Hepatology 2016;64:73-84.
[2] Dongiovanni P, Valenti L. A Nutrigenomic Approach to Non-Alcoholic Fatty Liver Disease. Int J Mol Sci 2017;18.
[3] Valenti L, Bugianesi E, Pajvani U, Targher G. Nonalcoholic fatty liver disease: cause or consequence of type 2 diabetes? Liver Int 2016;36:1563-1579.
[4] Loomba R, Seguritan V, Li W, Long T, Klitgord N, Bhatt A, et al. Gut Microbiome-Based
Metagenomic Signature for Non-invasive Detection of Advanced Fibrosis in Human Nonalcoholic Fatty Liver Disease. Cell Metab 2017;25:1054-1062 e1055.
[5] Tilg H, Moschen AR. Evolution of inflammation in nonalcoholic fatty liver disease: the multiple parallel hits hypothesis. Hepatology 2010;52:1836-1846.
[6] Eslam M, Valenti L, Romeo S. Genetics and epigenetics of NAFLD and NASH: Clinical impact. J Hepatol 2018;68:268-279.
[7] Kleiner DE, Brunt EM, Van Natta M, Behling C, Contos MJ, Cummings OW, et al. Design and validation of a histological scoring system for nonalcoholic fatty liver disease. Hepatology
2005;41:1313-1321.
[8] Singh S, Allen AM, Wang Z, Prokop LJ, Murad MH, Loomba R. Fibrosis progression in nonalcoholic fatty liver vs nonalcoholic steatohepatitis: a systematic review and meta-analysis of paired-biopsy studies. Clin Gastroenterol Hepatol 2015;13:643-654.e641-649; quiz e639-640. [9] Pais R, Charlotte F, Fedchuk L, Bedossa P, Lebray P, Poynard T, et al. A systematic review of follow-up biopsies reveals disease progression in patients with non-alcoholic fatty liver. J Hepatol 2013;59:550-556.
[10] McPherson S, Hardy T, Henderson E, Burt AD, Day CP, Anstee QM. Evidence of NAFLD
progression from steatosis to fibrosing-steatohepatitis using paired biopsies: implications for prognosis and clinical management. J Hepatol 2015;62:1148-1155.
[11] Pelusi S, Petta S, Rosso C, Borroni V, Fracanzani AL, Dongiovanni P, et al. Renin-Angiotensin System Inhibitors, Type 2 Diabetes and Fibrosis Progression: An Observational Study in Patients with Nonalcoholic Fatty Liver Disease. PLoS One 2016;11:e0163069.
[12] Sanyal A, Harrison S, Ratziu V, Abdelmalek M, Diehl AM, Caldwell S, et al. Changes in fibrosis, but not the NAFLD Activity Score (NAS), are associated with disease progression in patients with nonalcoholic steatohepatitis (NASH) and advanced fibrosis. J Hepatol 2017;66:S2.
[13] Ratziu V, Wong VW-S, Lanthier N, Boursier J, Alazawi W, Kluwe J, et al. Hepatic fibrosis is associated with histological activity in nonalcoholic steatohepatitis: an analysis from a large database of screening biopsies in the CENTAUR trial. J Hepatol 2017;66:S113.
[14] Ajmera V, Park CC, Caussy C, Singh S, Hernandez C, Bettencourt R, et al. Magnetic Resonance Imaging Proton Density Fat Fraction Associates With Progression of Fibrosis in Patients With
Nonalcoholic Fatty Liver Disease. Gastroenterology 2018.
[15] McPherson S, Pais R, Valenti L, Schattenberg J, Dufour JF, Tsochatzis E, et al. Further delineation of fibrosis progression in NAFLD: evidence from a large cohort of patients with sequential biopsies. . J Hepatol 2017;64:S593.
[16] Dongiovanni P, Stender S, Pietrelli A, Mancina RM, Cespiati A, Petta S, et al. Causal relationship of hepatic fat with liver damage and insulin resistance in nonalcoholic fatty liver. J Intern Med
2018;283:356-370.
[17] Dongiovanni P, Meroni M, Baselli GA, Bassani GA, Rametta R, Pietrelli A, et al. Insulin resistance promotes Lysyl Oxidase Like 2 induction and fibrosis accumulation in non-alcoholic fatty liver disease. Clinical science 2017;131:1301-1315.
[18] Sanyal AJ, Brunt EM, Kleiner DE, Kowdley KV, Chalasani N, Lavine JE, et al. Endpoints and clinical trial design for nonalcoholic steatohepatitis. Hepatology 2011;54:344-353.
[19] Estes C, Anstee QM, Teresa Arias-Loste M, Bantel H, Bellentani S, Caballeria J, et al. Modeling NAFLD Disease Burden in China, France, Germany, Italy, Japan, Spain, United Kingdom, and United States for the period 2016-2030. J Hepatol 2018.
[20] Dongiovanni P, Petta S, Maglio C, Fracanzani AL, Pipitone R, Mozzi E, et al. Transmembrane 6 superfamily member 2 gene variant disentangles nonalcoholic steatohepatitis from cardiovascular disease. Hepatology 2015;61:506-514.
[21] Mancina RM, Dongiovanni P, Petta S, Pingitore P, Meroni M, Rametta R, et al. The MBOAT7-TMC4 Variant rs641738 Increases Risk of Nonalcoholic Fatty Liver Disease in Individuals of European Descent. Gastroenterology 2016;150:1219-1230 e1216.
[22] Colloredo G, Guido M, Sonzogni A, Leandro G. Impact of liver biopsy size on histological evaluation of chronic viral hepatitis: the smaller the sample, the milder the disease. J Hepatol 2003;39:239-244.
[23] Kleiner DE, Brunt EM, Van Natta M, Behling C, Contos MJ, Cummings OW, et al. Design and validation of a histological scoring system for nonalcoholic fatty liver disease. Hepatology
2005;41:1313-1321.
[24] Petta S, Valenti L, Marra F, Grimaudo S, Tripodo C, Bugianesi E, et al. MERTK rs4374383 polymorphism affects the severity of fibrosis in non-alcoholic fatty liver disease. J Hepatol 2016;64:682-690.
[25] van der Poorten D, Samer CF, Ramezani-Moghadam M, Coulter S, Kacevska M, Schrijnders D, et al. Hepatic fat loss in advanced nonalcoholic steatohepatitis: are alterations in serum adiponectin the cause? Hepatology 2013;57:2180-2188.
[26] Valenti L, Mendoza RM, Rametta R, Maggioni M, Kitajewski C, Shawber CJ, et al. Hepatic Notch Signaling Correlates With Insulin Resistance and Nonalcoholic Fatty Liver Disease. Diabetes
2013;62:4052-4062.
[27] Marchesini G, Bugianesi E, Forlani G, Cerrelli F, Lenzi M, Manini R, et al. Nonalcoholic fatty liver, steatohepatitis, and the metabolic syndrome. Hepatology 2003;37:917-923.
[28] Paradis V, Perlemuter G, Bonvoust F, Dargere D, Parfait B, Vidaud M, et al. High glucose and hyperinsulinemia stimulate connective tissue growth factor expression: a potential mechanism involved in progression to fibrosis in nonalcoholic steatohepatitis. Hepatology 2001;34:738-744. [29] Pirazzi C, Valenti L, Motta BM, Pingitore P, Hedfalk K, Mancina RM, et al. PNPLA3 has retinyl-palmitate lipase activity in human hepatic stellate cells. Hum Mol Genet 2014;23:4077-4085.
[30] Mondul A, Mancina RM, Merlo A, Dongiovanni P, Rametta R, Montalcini T, et al. PNPLA3 I148M Variant Influences Circulating Retinol in Adults with Nonalcoholic Fatty Liver Disease or Obesity. J Nutr 2015;145:1687-1691.
[31] Pingitore P, Dongiovanni P, Motta BM, Meroni M, Lepore SM, Mancina RM, et al. PNPLA3 overexpression results in reduction of proteins predisposing to fibrosis. Hum Mol Genet
[32] Bruschi FV, Claudel T, Tardelli M, Caligiuri A, Stulnig TM, Marra F, et al. The PNPLA3 I148M variant modulates the fibrogenic phenotype of human hepatic stellate cells. Hepatology
