14783231, 2021, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/liv.14898 by Readcube (Labtiva Inc.), Wiley Online Library on [14/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Received: 20 January 2021 | Revised: 1 April 2021 | Accepted: 6 April 2021 DOI: 10.1111/liv.14898 ORIGINAL ARTICLE Definite and indeterminate nonalcoholic steatohepatitis share similar clinical features and prognosis: A longitudinal study of 1893 biopsy-­proven nonalcoholic fatty liver disease subjects Javier Ampuero1,2,3 | Rocío Aller4 | Rocío Gallego-­Durán2,3 | Javier Crespo5 | Javier Abad6 | Águeda González-­Rodríguez7 | Judith Gómez-­Camarero8 | Joan Caballería3,9 | Oreste Lo Iacono10 | Luis Ibañez3,11 | Javier García-­Samaniego3,12 | Rosa Martín-­Mateos3,13 | Rubén Francés3,14 | Conrado Fernández-­Rodríguez15 | Moisés Diago16 | Germán Soriano3,17 | Raúl J. Andrade3,18 | Raquel Latorre19 | Francisco Jorquera3,20 | Rosa M. Morillas3,21 | Desam Escudero22 | Pamela Estévez23 | Manuel Hernández-­Guerra24 | Salvador Augustín25 | María Jesús Pareja-­Megia26 | Jesús M. Banales3,27 | Patricia Aspichueta28,3 | Salvador Benlloch3,29 | 30 31 32 José Miguel Rosales | Javier Salmerón | Juan Turnes | 1,2,3 Manuel Romero-­Gómez | on behalf of HEPAmet Registry 1 Hospital Universitario Virgen del Rocío, Universidad de Sevilla, Seville, Spain 2 SeLiver Group, Instituto de Biomedicina de Sevilla, Seville, Spain 3 CIBERehd, Madrid, Spain 4 Centro de Investigación de Endocrinología y Nutrición, Hospital Clínico Universitario de Valladolid, Universidad de Valladolid, Valladolid, Spain 5 Digestive Department, Hospital Universitario Marqués de Valdecilla, Santander, Spain 6 Digestive Department, Hospital Universitario Puerta de Hierro, Madrid, Spain 7 Liver Research Unit, Hospital Universitario Santa Cristina Instituto de Investigación Sanitaria Princesa, Madrid, Spain 8 Digestive Department, Hospital Universitario de Burgos, Burgos, Spain 9 Liver Unit, Hospital Clínic, Institut d'Investigacions Biomèdiques August Pi i Sunyer (IDIBPAS), Barcelona, Spain 10 Digestive Department, Hospital Universitario Tajo, Aranjuez, Spain 11 Hospital General Universitario Gregorio Marañón, Instituto de Investigación Sanitaria Gregorio Marañón, Madrid, Spain 12 IdiPAZ, Hospital Universitario La Paz, Madrid, Spain 13 Digestive Department, Hospital Universitario Ramón y Cajal, Madrid, Spain 14 Hospital General Universitario de Alicante, Universidad Miguel Hernández, Elche, Spain 15 Hospital Universitario Fundación de Alcorcón, Universidad Rey Juan Carlos, Móstoles, Spain 16 Digestive Department, Hospital General Universitario de Valencia, Valencia, Spain 17 Digestive Department, Hospital de la Santa Creu i San Pau, Barcelona, Spain 18 Unidad de Gestión Clínica de Enfermedades Digestivas, Instituto de Investigación Biomédica de Málaga-­IBIMA, Hospital Universitario Virgen de la Victoria, Universidad de Málaga, Málaga, Spain 19 Digestive Department, Hospital Universitari Son Llátzer, Mallorca, Spain 20 Servicio de Aparato Digestivo, Complejo Asistencial Universitario de León, IBIOMED, León, Spain 21 Digestive Department, Hospital Germans Trias i Pujol, Badalona, Spain Abbreviations: AHT, arterial hypertension; BMI, body mass index; MAFLD, metabolic-­associated fatty liver disease; NAFL, nonalcoholic fatty liver; NAFLD, nonalcoholic fatty liver disease; NASH, nonalcoholic steatohepatitis; T2DM, type 2 diabetes mellitus. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2021 The Authors. Liver International published by John Wiley & Sons Ltd. 2076 |  wileyonlinelibrary.com/journal/liv Liver International. 2021;41:2076–2086. | 14783231, 2021, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/liv.14898 by Readcube (Labtiva Inc.), Wiley Online Library on [14/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License AMPUERO et al 2077 22 Hospital Clínico Universitario de Valencia, Universitat de València, Valencia, Spain 23 Digestive Department, Complejo Hospitalario Universitario de Vigo, Vigo, Spain 24 Digestive Department, Hospital Universitario de Canarias, Santa Cruz de Tenerife, Spain 25 Digestive Department, Hospital Vall d'Hebrón, Barcelona, Spain 26 Pathology Department, Hospital Universitario Virgen de Valme, Sevilla, Spain 27 Department of Liver and Gastrointestinal Diseases, Biodonostia Research Institute—­Donostia University Hospital, University of the Basque Country (UPV/ EHU), Ikerbasque, San Sebastian, Spain 28 Department of Physiology, Faculty of Medicine and Nursing, Biocruces Research Institute, University of Basque Country UPV/EHU, Leioa, Spain 29 Digestive Department, Hospital Universitari i Politecnic La Fe, Valencia, Spain 30 Digestive Department, Agencia Sanitaria Costa del Sol, Marbella, Spain 31 Digestive Department, Hospital Universitario San Cecilio, Granada, Spain 32 Digestive Department, Complejo Hospitalario de Pontevedra, Pontevedra, Spain Correspondence Javier Ampuero, Digestive Disease Abstract Department and CIBERehd, Virgen del Background and Aim: Histological score systems may not fully capture the essential Rocio University Hospital, Avenida Manuel Siurot s/n, 41013 Sevilla, Spain. nonalcoholic steatohepatitis (NASH) features, which is one of the leading causes of Email: jampuero-ibis@us.es screening failure in clinical trials. We assessed the NASH distribution and its compo- Funding information nents across the fibrosis stages and their impact on the prognosis and their relation- This project has been partially funded by ship with the concept of metabolic-­associated fatty liver disease (MAFLD). the ‘Consejería de Salud de la Junta de Andalucía’ (PI-­0 075-­2014) and the ‘Spanish Methods: Spanish multicenter study including 1893 biopsy-­proven nonalcoholic fatty Ministry of Economy, Innovation and liver disease (NAFLD) patients from HEPAmet registry. NASH was diagnosed by NAS Competition, Instituto de Salud Carlos III’ (PI19/01404, PI16/01842, PI17/00535 and score ≥4 (including steatosis, ballooning and lobular inflammation) and fibrosis by GLD19/00100). The funders have not had Kleiner score. The presence of MAFLD was determined. Progression to cirrhosis, first any role in the design, analysis, writing or interpretation of this project. episode of decompensated cirrhosis and death were collected during the follow-­up (4.7 ± 3.8 years). Handling Editor: Salvatore Petta Results: Fibrosis was F0 34.3% (649/1893), F1 27% (511/1893), F2 16.5% (312/1893), F3 15% (284/1893) and F4 7.2% (137/1893). NASH diagnosis 51.9% (982/1893), and its individual components (severe steatosis, ballooning and lobular inflammation), in- creased from F0 (33.6%) to F2 (68.6%), and decreased significantly in F4 patients (51.8%) (P = .0001). More than 70% of non-­NASH patients showed some inflam- matory activity (ballooning or lobular inflammation), showing a similar MAFLD rate than NASH (96.2% [945/982] vs. 95.2% [535/562]) and significantly higher than non- alcoholic fatty liver (NAFL) subjects (89.1% [311/349]) (P < .0001). Progression to cirrhosis was similar between NASH (9.5% [51/539]) and indeterminate NASH (7.9% [25/316]), and higher than steatosis (5% [14/263]) (logRank 8.417; P = .015). Death and decompensated cirrhosis were similar between these. Conclusions: The prevalence of steatohepatitis decreased in advanced liver disease. However, most of these patients showed some inflammatory activity histologically and had metabolic disturbances. These findings should be considered in clinical trials whose main aim is to prevent cirrhosis progression and complications, liver transplant and death. KEYWORDS ballooning, fatty liver disease, inflammation, metabolic-­associated fatty liver disease, natural coursesteatohepatitis, steatosis | 14783231, 2021, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/liv.14898 by Readcube (Labtiva Inc.), Wiley Online Library on [14/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 2078 AMPUERO et al Key points • The prevalence of definite NASH histological criteria decreases in advanced liver disease, precluding these subjects from receiving new therapeutic options. • Patients with NASH and not-­NASH specific inflammatory activity share similar clinical fea- tures and prognosis, and they are different from those with simple steatosis. • Most of the patients with significant fibrosis and cirrhosis without a well-­defined NASH show metabolic disturbances. • This is the first study in clinical practice reinforcing the renaming of MAFLD. 1 | I NTRO D U C TI O N ballooning and lobular inflammation) across the fibrosis stages and link them with prognostic outcomes to determine if NAS score ≥4 Nonalcoholic fatty liver disease (NAFLD) includes a spectrum of his- reflects a reliable scenario in clinical practice as inclusion criterium tological features that range from simple steatosis to nonalcoholic in NAFLD clinical trials. steatohepatitis (NASH) and, finally, cirrhosis.1 Splitting NAFLD into three successive stages (steatosis, non-­cirrhotic NASH and cirrhotic NASH) has provided a convenient conceptual framework for improv- 2 | M E TH O DS ing the diagnostic methods and identifying areas of potential future drug development. 2 However, the diagnosis of NAFLD is challeng- 2.1 | Selection of patients ing sometimes, so we need to consider a more pragmatic approach to targeting these individuals. In this scenario, the term “metabolic-­ This is an observational study of 1893 patients with biopsy-­confirmed associated fatty liver disease” (MAFLD) has been recently pro- NAFLD who had been enrolled and prospectively followed up from posed.3 The diagnosis of MAFLD is based on recognizing underlying the Spanish HEPAmet Registry. This registry is governed by the alterations in metabolism,4 beyond the histological classification of Spanish Association for the Study of the Liver and the Network of NAFLD. MAFLD is defined by the presence of steatosis (by histol- Biomedical Research Centre for the Study of the Liver and Digestive ogy or imaging) and overweight or at least two metabolic risk fac- Diseases (CIBERehd). Data monitoring is a fundamental element of tors. Notably, steatosis could be absent in the case of advanced liver the registry, ensuring data procurement accuracy and minimization disease.5 of bias. Subjects enrolled in NASH clinical trials are usually required to Patients underwent a liver biopsy according to the routine deci- have a NAS ≥ 4 and a fibrosis stage from F1 to cirrhosis by liver bi- sions in the clinical practice (eg, presence of fatty liver by imaging, opsy. However, up to half of the screened individuals fail to meet increased aminotransferase levels, based on non-­invasive tests and these eligibility criteria.6 We should consider that NAS score was suspected advanced liver disease by imaging or laboratory tests) or developed as a tool to measure NAFLD changes during therapeu- at the time of bariatric surgery. The inclusion criterion was biopsy-­ tic trials instead of as a surrogate for the histologic diagnosis of proven NAFLD, irrespective of the existence of NASH or fibrosis NASH.7 Notably, previous studies have reported that 20%–­4 0% of stage. Exclusion criteria were significant alcohol intake (≥30 g daily the NAFLD population does not display a definite NASH histolog- for men and ≥20 g daily for women) and evidence of concomitant ical diagnosis,8,9 making these individuals ineligible for enrollment. liver disease (ie, viral or autoimmune hepatitis, HIV, drug-­induced Consequently, the large number of suboptimal biopsies adds sig- fatty liver, hemochromatosis or Wilson's disease). The study was nificantly to the cost and duration of clinical trials. However, liver performed in agreement with the Declaration of Helsinki and ap- fibrosis is considered the strongest predictor of adverse clinical out- proved by the Ethics and Clinical Research Committee of every cen- comes.10 Besides, treatment goals for patients with advanced liver tre. All patients were informed of the nature of the study and gave disease are to halt or slow fibrosis progression, prevent clinical de- their written consent to participate. compensation, reduce the need for liver transplantation and improve survival.11 The liver biopsy must keep being essential in the management of 2.2 | Clinical assessment 12 NAFLD, but we should avoid converting this tool into a barrier to diagnose, treat or identify at-­risk patients. For instance, the develop- Demographic characteristics, anthropometric measurements and ment of non-­invasive tests and imaging-­based methods has allowed laboratory tests (alanine aminotransferase (ALT), aspartate ami- a wide expansion of liver fibrosis assessment in at-­risk patients.13-­16 notransferase (AST), gamma glutamyl transferase (GGT), triglyc- Thus, the liver biopsy should be relocated to maximize its advan- erides, cholesterol, HDL-­c, LDL-­c, fasting glucose, HbA1c, insulin, tages in selected patients. In a Spanish large cohort of individuals creatinine and albumin) were recorded at the same time of liver bi- with biopsy-­proven NAFLD followed-­up during 4.7 years, we aimed opsy. An overnight (12-­h) fasting blood sample was taken for rou- to assess the prevalence of NASH and its components (steatosis, tine biochemical analyses. Homeostatic model assessment (HOMA) | 14783231, 2021, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/liv.14898 by Readcube (Labtiva Inc.), Wiley Online Library on [14/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License AMPUERO et al 2079 was calculated based on insulin and glucose (fasting insulin (mIU/ 2.4 | Objectives ml) × fasting glucose (mg/ml)/405). We defined the following metabolic risk factors at baseline17: We assessed if NAS score ≥4 as inclusion criterion in NAFLD clinical (a) type 2 diabetes mellitus (T2DM), defined by fasting blood trials reflects a reliable scenario in clinical practice, due to some cases glucose ≥126 mg/dl, or Hb1Ac >6.5%, or use of blood glucose-­ with higher NAS score could not have findings of definite NASH and lowering agents; (b) arterial hypertension (AHT), determined by other cases of lower NAS score do. 20 To achieve this goal, secondary blood pressure ≥130/85 mm Hg or use of blood pressure-­lowering aims were (a) to analyze the presence and the distribution of NAS agents; (c) low HDL, defined by HDL-­c < 40 mg/dl in men or score ≥4 and of its single components (steatosis, inflammation and <50 mg/dl in women or use of lipid-­lowering agents; (d) hyper- ballooning) across the fibrosis stages; (b) to analyze the histological triglyceridemia, characterized by triglyceride levels ≥150 mg/dl features of patients with apparent lack of NASH (NAS < 4), who are or use of lipid-­lowering agents; (e) overweight and obesity, de- usually excluded from clinical trials; (c) to perform an exploratory termined by a body mass index (BMI) >25 and 30 kg/m2 , respec- analysis to determine the prognosis of three histological patterns tively; (f) insulin resistance, defined by HOMA > 2.5 in absence of based on NAS score and liver inflammation (NASH vs. indeterminate T2DM. Metabolic syndrome was considered when at least three NASH vs. NAFL) in terms of progression to cirrhosis, the appearance single components were present, while MAFLD was defined ac- of the first episode of decompensated cirrhosis and death. cording to the recent consensus. 3 Follow-­up was defined, depending on the endpoint, as the time from the liver biopsy to the progression to cirrhosis (for non-­ 2.5 | Statistical analyses cirrhotic patients), the first event of cirrhosis complication (asci- tes, hepatic encephalopathy and variceal bleeding) or death. They Data were reported as the mean ± standard deviation for normal were collected whenever the investigator received notice about and median (interquartile range) for non-­normal continuous vari- the event. In the case of no event, patients were censored at ables, while frequency was used for discrete variables. In the uni- 10 years of follow-­up or at the end of the study period (May 2020). variate comparisons, we used the Student's t test and ANOVA with In non-­cirrhotic patients at baseline, the progression to cirrhosis Bonferroni adjustments for continuous samples and chi-­square test was considered when typical findings of cirrhosis were observed or Fisher's exact test for the qualitative ones. Non-­parametric al- during the follow-­up in a second liver biopsy (3/90), abdominal ul- ternatives (Mann–­Whitney U and Kruskal–­Wallis tests) were used trasound (38/90) or transient elastography (>16 kPa) (36/90), or for non-­normal distributions. Survival analysis was assessed by the when an episode of decompensated cirrhosis occurred (13/90) Kaplan–­Meyer method, and differences between patient subgroups (TARGET-­N ASH criteria18). were evaluated by the log-­rank test. Values were supposed to be statistically significant when P < .05. On the other hand, the annual rate of progression to cirrhosis 2.3 | Histological assessment and the appearance of the first episode of decompensated cirrhosis and death were computed by dividing the number of patients with The diagnosis of NAFLD was based on histological criteria. All liver the defined event by the number of person years of which patients biopsies were assessed by experienced hepato-­pathologists (leaded were followed. We multiply rates by 100 to transform in cases per by MJPM), associated with the LITMUS histopathologists group,19 100 person years. who were blinded regarding the patient's evaluation and clinical The method used for missing data was complete-­case analysis data. Samples of <15-­mm length or <10 portal tracts were consid- because statistical packages excluded individuals with any missing ered not suitable for diagnosis and were excluded. Several histo- value. STATA (12.0, STATA Corporation, College Station, TX, USA) logical aspects were measured. Steatosis, lobular inflammation and statistical package was used in all analyses and GraphPad Prism (ver- hepatocyte balloon degeneration were systematically assessed ac- sion 6.0; GraphPad Software, Inc, La Jolla, CA) for graphics. cording to the NASH CRN Scoring System: (a) Steatosis was rated as Grade 0 (<5%), Grade 1 (5%–­33%), Grade 2 (33%–­66%) and Grade 3 (>66%); (b) hepatocyte ballooning was considered as 0 (none), 1 3 | R E S U LT S (mild-­few) and 2 (moderate-­marked); (c) lobular inflammation was rated as 0 (none), 1 (<2 foci/20 optical field), 2 (2-­4 foci/20 optical field) 3.1 | Baseline features of the study population and 3 (>4 foci/20 optical field). Although NASH CRN does not define exactly NASH by using the NAS score, 20 we determined NASH ac- The baseline features of the study cohort are stated in Tables 1 cording to NAS score ≥4 (with at least 1 point each in inflammation and S1. Briefly, steatohepatitis (NAS ≥ 4) was present in 51.9% 7 and ballooning). This threshold was based on the inclusion criteria (982/1893) of the overall population. According to liver fibrosis: F0 of most Food and Drug Administration (FDA)-­approved clinical trials 34.3% (649/1893), F1 27% (511/1893), F2 16.5% (312/1893), F3 to develop NAFLD therapeutic drugs. The severity of fibrosis was 15% (284/1893) and F4 7.2% (137/1893). The 41.6% (788/1893), staged from 0 to 4. 36.8% (696/1893) and 21.6% (409/1893) showed Grades 1, 2 and | 14783231, 2021, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/liv.14898 by Readcube (Labtiva Inc.), Wiley Online Library on [14/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 2080 AMPUERO et al 3 of steatosis, respectively. On the other hand, lobular inflammation 67% (1269/1893) of subjects (47.6% showed mild-­few and 19.4% was seen in 73.5% (1391/1893) of patients (50.5% had mild and 23% moderate-­marked). The histological activity in patients, stratified by moderate-­to-­severe inflammation), while ballooning was present in fibrosis stage, is shown in Table 2. TA B L E 1 Baseline features of the overall population Overall cohort 3.2 | Histological features depending on the Characteristic (n = 1893) fibrosis stage Male sex 48% (909/1893) Age; years ± SD 51.8 ± 12.3 We analyzed the distribution of severe steatosis (Grade 3), bal- BMI ± SD (kg/m ) 2 35.3 ± 8.8 looning and lobular inflammation according to the fibrosis stage Obesity (BMI > 30 kg/m ) 2 61.5% (1164/1893) (Figure 1A). All these three single components of the histological Arterial hypertension 48.7% (922/1893) definition of steatohepatitis were less frequent in patients with F0. The prevalence of these features, according to the fibrosis stage, Type 2 diabetes mellitus 36.7% (695/1893) showed an inverted U-­shaped curve. Particularly, severe steatosis, Glucose ± SD (mg/dl) 118 ± 43 ballooning and lobular inflammation increased from F1 to F2, while HOMA-­IR ± SD 5.5 ± 5.2 their percentage was similar in F3 and decreased in F4 patients. HOMA > 2.5 (in absence of T2DM) 14.4% (273/1893) According to the NAS score, the same trend was observed for the Total cholesterol ± SD (mg/dl) 194 ± 45 distribution of NASH-­specific inflammation, which was significantly HDL-­c ± SD (mg/dl) 49 ± 15 lower in cirrhotic than in F2 and F3 patients (Figure 1B). Besides, we LDL-­c ± SD (mg/dl) 117 ± 38 analyzed the distribution of the following histological patterns by Triglycerides ± SD (mg/dl) 166 ± 116 fibrosis stage: steatohepatitis (NAS ≥ 4) versus indeterminate NASH Hypertriglyceridemia 48.1% (910/1893) (at least, presence of ballooning or lobular inflammation) versus AST ± SD (IU/L) 43 ± 38 NAFL (lack of ballooning and lobular inflammation); in the overall co- ALT ± SD (IU/L) 60 ± 53 hort (Figure 2A) and depending on the presence of diabetes mellitus GGT ± SD (IU/L) 102 ± 57 (Figure 2B), obesity (BMI > 30) (Figure 2C), the indication of the liver biopsy (Figure 2D) and the sex (Figure 2E). Phosphatase alkaline ±SD (IU/L) 95 ± 53 On the other hand, the distribution of steatohepatitis, as well as Bilirubin ± SD (mg/dl) 0.68 ± 0.4 their individual components, is shown in Table S2, according to the Albumin ± SD (g/dl) 4.39 ± 0.4 presence of MAFLD clinical criteria. Briefly, the presence of meta- Creatinine ± SD (mg/dl) 0.83 ± 0.2 bolic derangement influenced a higher inflammatory activity in the Platelet count ± SD (×109/L) 234 ± 71 liver only in patients with F0: ballooning 51.6% versus 22.5%, lobu- INR ± SD 1.03 ± 0.1 lar inflammation 57.5% versus 32.5% and NASH 35% versus 12.5% Abbreviations: BMI, body mass index; SD, standard deviation. (P < .05 for all comparisons). TA B L E 2 Histological activity, including steatohepatitis, steatosis, ballooning and lobular inflammation, stratified by fibrosis stage Fibrosis stage F0 F1 F2 F3 F4 NASH (NAS > 4) 33.6% (218/649) 56.2% (287/511) 68.6% (214/312) 67.6% (192/284) 51.8% (71/137) Steatosis Grade 1 50.5% (328/649) 37.4% (191/511) 34.3% (107/312) 34.9% (99/284) 46% (63/137) Grade 2 34.4% (191/649) 38.7% (198/511) 36.2% (113/312) 39.8% (113/284) 35.8% (49/137) Grade 3 15.1% (98/649) 23.9% (122/511) 29.5% (92/312) 25.4% (72/284) 18.2% (25/137) Ballooning 0 50.2% (326/649) 25.2% (129/511) 18.6% (58/312) 23.2% (66/284) 32.8% (45/137) 1 39% (253/649) 54% (276/511) 51% (159/312) 52.5% (149/284) 46.7% (64/137) 2 10.8% (73/649) 20.7% (106/511) 30.4% (95/312) 24.3% (69/284) 20.4% (28/137) Lobular inflammation 0 44.1% (286/649) 20.9% (107/511) 11.9% (37/312) 15.1% (43/284) 21.2% (29/137) 1 43.8% (284/649) 56.2% (287/511) 53.5% (167/312) 51.4% (146/284) 52.6% (72/137) 2 10.3% (67/649) 18.6% (95/511) 28.8% (90/312) 30.3% (86/284) 22.6% (31/137) 3 1.8% (12/649) 4.3% (22/511) 5.8% (18/312) 3.2% (9/284) 3.6% (5/137) Abbreviation: NASH, nonalcoholic steatohepatitis. | 14783231, 2021, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/liv.14898 by Readcube (Labtiva Inc.), Wiley Online Library on [14/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License AMPUERO et al 2081 F I G U R E 1 Histological features depending on the fibrosis stage. A, Prevalence of severe steatosis (Grade 3)*, ballooning (>1)**, and lobular inflammation (>1)***. B, NAS score. *P values for steatosis. F0 versus F4 0.356; F1 versus F4 0.162; F2 versus F4 0.012; F3 versus F4 0.104. **P values for ballooning. F0 versus F4 0.0002; F1 versus F4 0.074; F2 versus F4 0.0009; F3 versus F4 0.036. ***P values for inflammation. F0 versus F4 <0.0001; F1 versus F4 0.953; F2 versus F4 0.010; F3 versus F4 0.036 (A) (B) (C) (D) (E) F I G U R E 2 Distribution of definite nonalcoholic steatohepatitis (NASH), indeterminate NASH, and nonalcoholic fatty liver (NAFL). A, Overall cohort. B, Depending on the presence of type 2 diabetes mellitus (T2DM). C, Depending on the presence of obesity. D, Indication of the liver biopsy. E, Comparison between men and women | 14783231, 2021, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/liv.14898 by Readcube (Labtiva Inc.), Wiley Online Library on [14/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 2082 AMPUERO et al 3.3 | Histological features in patients without NASH 3.5 | Prognosis according to the histological pattern Despite 31.4% of patients showing F2 had no NASH, 83.7% of After the exclusion of patients who underwent liver biopsy during these patients showed some inflammatory component (ballooning bariatric surgery (n = 539) and were enrolled in clinical trials (n = 120), or lobular inflammation) in the liver biopsy (Figure 3A). Also, a third 1234 subjects (65.2% of the overall cohort) were included in the sur- of patients with advanced fibrosis were not diagnosed with NASH. vival analysis. During the follow-­up (mean time 4.7 ± 3.8 years), 2.8% However, up to 71.7% of them showed some inflammatory degree in (34/1234) of patients died, 1.9% (24/1234) suffered from a first de- the liver (Figure 3B). In cirrhotic patients, the percentage of patients compensation (10.3% (12/116) among baseline cirrhotic patients), without steatohepatitis was 48.2%, although up to 68.2% of them 8.1% (90/1118) of non-­cirrhotics suffered progression to cirrhosis showed at least ballooning or lobular inflammation (Figure 3C). and 1.1% (13/1234) developed hepatocellular carcinoma. Fibrosis stage was the main histological determinant on the prognosis of the overall cohort (mortality was F0 1.8% [7/392] vs. 3.4 | Baseline features according to the F1 1.6% [5/315] vs. F2 3.1% [6/196] vs. F3 2.8% [6/215] vs. F4 9.6% histological pattern [10/116]; logRank 22.369, P = .0001). On the other hand, despite having more patients with advanced liver disease, NASH patients According to the histological pattern, the baseline features of the had not a worse survival (2.7% [16/598]) than those with inde- overall population are shown in Table 3. Briefly, both NASH and in- terminate NASH (1.4% [5/355]) or NAFL (4.6% [13/281]) (logRank determinate NASH groups showed similar rates for metabolic factors 3.175; P = .204). Also, the three groups suffered a similar rate such as obesity, diabetes mellitus and arterial hypertension, as well of first cirrhosis decompensation during the study (2% [12/598] as they were not significantly different for liver tests like bilirubin, vs. 1.7% [6/355] vs. 2.1% [6/281]; P = .806). Nevertheless, the albumin or platelet levels. Instead, NASH patients showed signifi- incidence of progression to cirrhosis was significantly higher for cantly higher AST and ALT levels. Also, MAFLD criteria were pre- both NASH patients (9.5% [51/539]; HR 2.35 [95% CI 1.30-­4.26]) sent in 96.2% (945/982) of NASH patients and 95.2% (535/562) in and those with indeterminate NASH (7.9% [25/316]; HR 1.90 indeterminate NASH, while it was 89.1% (311/349) in patients with [95% CI 1.01-­3 .66] [p = ns between them]) compared to subjects NAFL (P < .0001). On the other hand, the distribution of MAFLD by with NAFL (5% [14/263]; HR 1.00 [reference]) (logRank 8.417; histological pattern and fibrosis stage is shown in Figure S1. P = .015) (Figure 4A). Consequently, the annual incidence rate of (A) 100 (B) 100 90 90 80 80 70 70 60 60 50 50 40 40 30 67.6% 32.4% 30 68.6% 31.4% 20 20 10 10 0 0 Simple steatosis Simple steatosis Steatosis and lobular inflammation Steatosis and lobular inflammation Steatosis and ballooning Steatosis and ballooning Steatosis and lobular inflammation and ballooning Steatosis and lobular inflammation and ballooning 100 90 (C) 80 70 60 50 40 51.8% 48.2% 30 20 10 0 Simple steatosis Steatosis and lobular inflammation Steatosis and ballooning Steatosis and lobular inflammation and ballooning F I G U R E 3 Histological features in patients without a definite nonalcoholic steatohepatitis (NASH) diagnosis. A, Significant fibrosis. B, Advanced fibrosis. C, Cirrhosis | 14783231, 2021, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/liv.14898 by Readcube (Labtiva Inc.), Wiley Online Library on [14/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License AMPUERO et al 2083 TA B L E 3 Baseline features of the overall cohort according to the histological pattern Indeterminate P value for the Characteristic NASH NASH NAFL trend Male sex 44.6% (438/982)a,b 51.6% (290/562) 51.9% (181/349) .008 Age; years ± SD 51.9 ± 12.4 52.1 ± 12.4 50.8 ± 11.9 .243 BMI ± SD (kg/m2) 35.9 ± 8.7a 35.2 ± 8.7 34 ± 9 .005 2 a c Obesity (BMI ≥30 kg/m ) 70.9% (640/903) 67.5% (349/517) 59.7% (175/293) .002 Arterial hypertension 51.8% (515/982)a 48.2% (271/562)c 40.7% (142/349) .002 a c Type 2 diabetes mellitus 40.4% (397/982) 37.4% (210/562) 25.2% (88/349) .0001 Glucose ± SD (mg/dl) 121 ± 44a 117 ± 46 112 ± 37 .005 a c HbA1c (%) 6.5 ± 1.5 6.4 ± 1.4 6 ± 1.3 .017 Total cholesterol ± SD (mg/dl) 194 ± 47 193 ± 44 194 ± 43 .913 b HDL-­c ± SD (mg/dl) 48 ± 15 51 ± 17 49 ± 13 .003 LDL-­c ± SD (mg/dl) 116 ± 39 118 ± 37 120 ± 36 .346 Triglycerides ± SD (mg/dl) 177 ± 120 a,b 160 ± 125 147 ± 75 .0001 Hypertriglyceridemia 53% (520/982)a,b 46.3% (260/562)c 37.2% (130/349) .0001 a,b AST ± SD (IU/L) 48 ± 46 38 ± 25 36 ± 28 .0001 ALT ± SD (IU/L) 67 ± 61a,b 53 ± 40 54 ± 46 .0001 Bilirubin ± SD (mg/dl) 0.67 ± 0.4 0.70 ± 0.4 0.70 ± 0.4 .252 Albumin ± SD (g/dl) 4.39 ± 0.4 4.37 ± 0.5 4.41 ± 0.4 .520 Creatinine ± SD (mg/dl) 0.82 ± 0.3 0.83 ± 0.2 0.83 ± 0.2 .682 Platelet count ± SD (×109/L) 233 ± 70 231 ± 71 240 ± 74 .209 INR ± SD 1.03 ± 0.1 1.04 ± 0.2 1.03 ± 0.2 .830 a c MAFLD criteria 96.2% (945/982) 95.2% (535/562) 89.1% (311/349) .0001 Cirrhosis 7.2% (71/982) 8% (45/562) 6% (21/349) .530 Abbreviations: BMI, body mass index; MAFLD, metabolic-­associated fatty liver disease; NAFL, nonalcoholic fatty liver; NASH, nonalcoholic steatohepatitis; SD, standard deviation. a p<0.05 for comparison between NASH and indetermined NASH. b p<0.05 for comparison between NASH and NAFL. c p<0.05 for comparison between indetermined NASH and NAFL. progression to cirrhosis was higher for patients with NASH and NASH patients shared a similar rate of metabolic risk factors (in- indeterminate NASH than those with NAFL (Figure 4B). cluding obesity, diabetes, arterial hypertension and dyslipidemia) On the other hand, the impact of the histological pattern on and liver tests (albumin, bilirubin and platelets), although those with both mortality and decompensated cirrhosis was not influenced by NAS ≥ 4 presented higher levels of AST and ALT. Besides, these sub- MAFLD despite showing higher percentages of these outcomes in groups showed a significantly higher proportion of MAFLD criterion dysmetabolic patients (P > .05). According to progression to cirrho- than simple steatosis, focusing this term on patients with a higher sis, MAFLD also did not modify the impact of patients with NASH likelihood of some hepatic inflammatory grade. Interestingly, both or indeterminate NASH. By contrast, patients showing NAFL and NASH and indeterminate NASH showed similar progression rates MAFLD showed higher progression to cirrhosis than those metabol- to cirrhosis, decompensated cirrhosis and mortality. Therefore, our ically healthy patients (6.1% [14/231] vs. 0% [0/32]; logRank 3.677, findings cast doubt on current selection criteria for clinical trials P = .049) (Figure 4C). because many subjects, who would benefit from an experimental treatment to halt the progression of the liver disease, could be ex- cluded from clinical trials due to the absence of one single compo- 4 | D I S CU S S I O N nent of definite NASH. A relatively high proportion of our patients did not show find- Our study observed that the prevalence of NASH, defined by ings related to definite NASH (notably, half of the cirrhotic patients NAS ≥ 4 (with at least 1 point each in inflammation and ballooning), and a third of those with advanced fibrosis could not be diagnosed decreased in advanced liver disease, particularly in cirrhotic patients. from NASH). A recent Italian study demonstrated that up to 33% of Also, more than 70% of patients with NAS < 4 showed some inflam- patients with NAFLD-­related significant fibrosis did not show NASH matory degree in the liver biopsy. Both NASH and indeterminate (up to 10% had no inflammation), although the authors did not assess | 14783231, 2021, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/liv.14898 by Readcube (Labtiva Inc.), Wiley Online Library on [14/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 2084 AMPUERO et al (A) (B) (C) F I G U R E 4 A, Kaplan–­Meier curve to show the impact of the histological pattern on the progression to cirrhosis. B, Annual incidence rate of progression to cirrhosis depending on histological pattern and fibrosis stage (cases per 100 person-­years). C, Kaplan–­Meier curve to show the impact of metabolic-­associated fatty liver disease (MAFLD) on the progression to cirrhosis in patients with nonalcoholic fatty liver (NAFL) the distribution of NAS single components across the different fi- Indeterminate NASH, a subclassification recognized by the NASH brosis stages. 21 Another French study defined histologically severe CRN system, is characterized by steatosis with inflammation and disease as a SAF activity score ≥3 or advanced fibrosis (not requir- hepatocyte injury but atypical of definite steatohepatitis, and it ing steatohepatitis), assuming that patients with NAFLD-­related can be associated with varying amounts of fibrosis. 5 It has been advanced liver disease could not have some definite elements of documented that fibrosis can progress even in patients without NASH. 22 Besides, steatofibrosis demonstrated to have higher overall baseline steatohepatitis features, due to ballooning and lobular in- 23 mortality than NASH in NAFLD patients. flammation predicted fibrosis progression. 24 Consequently, these Interestingly, despite the lack of a definite NASH diagnosis, patients would more likely develop complications and should be most of the patients included in our study showed some inflam- cautiously monitored. In this scenario, we observed that both matory degree in the liver biopsy, irrespective of the fibrosis stage. patients with NASH and indeterminate NASH showed similar | 14783231, 2021, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/liv.14898 by Readcube (Labtiva Inc.), Wiley Online Library on [14/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License AMPUERO et al 2085 progression rates to cirrhosis, decompensated cirrhosis and mor- should individualize the most appropriate criteria for diagnosing and tality. These findings would support the concept of burn-­out in including patients in clinical trials according to the integration of clin- NASH-­related cirrhosis. 25 ical and histological features due to the NAS score limitation for such Indeterminate NASH, irrespective of fibrosis stage, belongs to the purpose. subpopulation of non-­NASH, similar to NAFL, according to the Case Definitions Working Group.5 Consequently, some scenarios could not C O N FL I C T O F I N T E R E S T contemplate this entity as the main etiology of liver disease. We ob- None. served a similar clinical phenotype between NASH and indeterminate NASH patients, including the presence of metabolic factors (MAFLD AU T H O R C O N T R I B U T I O N S was shown by more than 95% for the two groups) and liver tests (eg, al- Guarantor of the article: JA and MRG. Study design: JA. Drafting the bumin and bilirubin). An integral diagnosis, including histological crite- manuscript: JA and MRG. Statistical analyses and interpretation: JA ria and metabolic risk factors, could avoid the diagnosis of cryptogenic and MRG. Data acquisition and critical review of the manuscript: All liver disease in a significant number of individuals.26 Most importantly, authors. All authors approved the final version of the article, includ- it could help to decrease screening failures in clinical trials and avoid ing the authorship list. wasting the opportunity of useful therapy in patients at risk of pro- gression.27 Besides, considering the clinical phenotype to define the ORCID origin of liver cirrhosis in some circumstances would have two signifi- Javier Ampuero https://orcid.org/0000-0002-8332-2122 cant benefits. First, knowing the etiology of cirrhosis is mandatory for Rocío Gallego-­Durán https://orcid.org/0000-0002-9452-1661 personalized counselling and management. For instance, in the liver Águeda González-­Rodríguez https://orcid.org/0000-0001-6428-6210 transplant, immunosuppressive approach, monitoring of extrahepatic Germán Soriano https://orcid.org/0000-0002-9267-6811 events and the likelihood of recurrence could change depending on Raúl J. Andrade https://orcid.org/0000-0002-1565-0757 the etiology. Second, patients suffer from anxiety and frustration and Jesús M. Banales https://orcid.org/0000-0002-5224-2373 have many future uncertainties when they are diagnosed with a lead- Salvador Benlloch https://orcid.org/0000-0002-0794-6937 ing cause of mortality, such as cirrhosis, but they have no explanations about how the disease appeared and how they have to proceed.28,29 REFERENCES For instance, primary biliary cirrhosis has been renamed to cholangitis 1. Diehl AM, Day C. Cause, pathogenesis, and treatment of nonalco- owing to these reasons.30 holic steatohepatitis. N Engl J Med. 2017;377:2063-­2072. 2. Burt AD, Lackner C, Tiniakos DG. Diagnosis and assessment of The study has some limitations usually observed when real-­ NAFLD: definitions and histopathological classification. Semin Liver world data are evaluated. Ideally, the biopsies of all patients should Dis. 2015;35:207-­220. be assessed by one single pathologist to avoid introducing some 3. Eslam M, Sanyal AJ, George J. an international consensus panel. interobserver variability. Besides, the sampling variability of the bi- MAFLD: a consensus-­driven proposed nomenclature for metabolic associated fatty liver disease. Gastroenterology. 2020;158(7):1999-­ opsies might explain part of the findings. Despite these concerns, 2014.e1. https://doi.org/10.1053/j.gastro.2019.11.312 the results were consistent and robust after analyzing the data cen- 4. Ampuero J, Aller R, Gallego-­Durán R, et al. The effects of met- tre by centre (in fact, almost half of the cohort was reviewed by a abolic status on non-­alcoholic fatty liver disease-­related out- single central pathologist -­MJPM-­). Probably, the fact of having a comes, beyond the presence of obesity. Aliment Pharmacol Ther. 2018;48:1260-­1270. very selected group of expert hepato-­pathologists who review all 5. Siddiqui MS, Harrison SA, Abdelmalek MF, et al. Case definitions for the biopsies after an initial discussion to homogenizes the interpre- inclusion and analysis of endpoints in clinical trials for nonalcoholic tation of histologic criteria19 and that patients whose samples are steatohepatitis through the lens of regulatory science. Hepatology. not suitable for diagnosis are excluded from Hepamet Registry could 2018;67:2001-­2012. mitigate the worries. Also, a mean follow-­up time of 4.7 years could 6. Davison BA, Harrison SA, Cotter G, et al. Suboptimal reliability of liver biopsy evaluation has implications for randomized clinical not be enough to observe a high number of prognostic outcomes (eg, trials. J Hepatol. 2020;73(6):1322-­1332. https://doi.org/10.1016/j. hepatocellular carcinoma). However, we performed an exploratory jhep.2020.06.025 analysis in which we found statistically significant differences for 7. Kleiner DE, Brunt EM, Van Natta M, et al. Design and validation cirrhosis progression, probably because there were fast progressors, of a histological scoring system for nonalcoholic fatty liver disease. Hepatology. 2005;41:1313-­1321. including the use of annual incidence rate that considers for subjects 8. Loomba R, Lawitz E, Mantry PS, et al. The ASK1 inhibitor selon- the actual time at risk and does not require to complete the total sertib in patients with nonalcoholic steatohepatitis: A randomized, follow-­up time. phase 2 trial. Hepatology. 2018;67:549-­559. In summary, steatohepatitis features seem to be absent in half 9. Pai RK, Kleiner DE, Hart J, et al. Standardising the interpretation of liver biopsies in non-­alcoholic fatty liver disease clinical trials. of the patients with significant fibrosis and cirrhosis, which could Aliment Pharmacol Ther. 2019;50:1100-­1111. increase the screening failure rates and preclude to receive new 10. Dulai PS, Singh S, Patel J, et al. Increased risk of mortality by fibro- therapeutic options in a subset of patients at high risk of progres- sis stage in nonalcoholic fatty liver disease: Systematic review and sion and extrahepatic and/or liver-­related outcomes. Future studies meta-­analysis. Hepatology. 2017;65:1557-­1565. | 14783231, 2021, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/liv.14898 by Readcube (Labtiva Inc.), Wiley Online Library on [14/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 2086 AMPUERO et al 11. Rinella ME, Tacke F, Sanyal AJ, Anstee QM. Report on the AASLD/ fatty liver disease. J Hepatol. 2020;72(5):828-­838. https://doi. EASL joint workshop on clinical trial endpoints in NAFLD. J Hepatol. org/10.1016/j.jhep.2019.12.008 2019;71:823-­833. 23. Younossi ZM, Stepanova M, Rafiq N, et al. Nonalcoholic steatofi- 12. Bedossa P. Diagnosis of non-­alcoholic fatty liver disease/non-­ brosis independently predicts mortality in nonalcoholic fatty liver alcoholic steatohepatitis: why liver biopsy is essential. Liver disease. Hepatol Commun. 2017;1:421-­428. International. 2018;38:64-­66. 24. Pais R, Charlotte F, Fedchuk L, et al. A systematic review of fol- 13. Shah AG, Lydecker A, Murray K, et al. Comparison of noninvasive low-­up biopsies reveals disease progression in patients with non-­ markers of fibrosis in patients with nonalcoholic fatty liver disease. alcoholic fatty liver. J Hepatol. 2013;59:550-­556. Clin Gastroenterol Hepatol. 2009;7:1104-­1112. 25. Yoshioka Y, Hashimoto E, Yatsuji S, et al. Nonalcoholic steatohep- 14. Castera L, Friedrich-­Rust M, Loomba R. Noninvasive assessment atitis: cirrhosis, hepatocellular carcinoma, and burnt-­out NASH. J of liver disease in patients with nonalcoholic fatty liver disease. Gastroenterol. 2004;39:1215-­1218. Gastroenterology. 2019;156(5):1264-­1281.e4. 26. Younossi Z, Stepanova M, Sanyal AJ, et al. The conundrum of cryp- 15. Ampuero J, Pais R, Aller R, et al. Development and validation of togenic cirrhosis: adverse outcomes without treatment options. J hepamet fibrosis scoring system a simple, noninvasive test to iden- Hepatol. 2018;69:1365-­1370. tify patients with nonalcoholic fatty liver disease with advanced 27. Asrani SK, Devarbhavi H, Eaton J, Kamath PS. Burden of liver dis- fibrosis. Clin Gastroenterol Hepatol. 2020;18(1):216-­225.e5. https:// eases in the world. J Hepatol. 2019;70:151-­171. doi.org/10.1016/j.cgh.2019.05.051 28. Shiha G, Korenjak M, Eskridge W, et al. Redefining fatty liver disease: 16. Angulo P, Hui JM, Marchesini G, et al. The NAFLD fibrosis score: an international patient perspective. Lancet Gastroenterol Hepatol. a noninvasive system that identifies liver fibrosis in patients with 2021;6(1):73-­79. https://doi.org/10.1016/S2468​-­1253(20)30294​-­6 NAFLD. Hepatology. 2007;45:846-­854. 29. Kimbell B, Boyd K, Kendall M, Iredale J, Murray SA. Managing un- 17. Grundy SM, Cleeman JI, Daniels SR, et al. Diagnosis and manage- certainty in advanced liver disease: a qualitative, multiperspective, ment of the metabolic syndrome: an American Heart Association/ serial interview study. BMJ Open. 2015;5:e009241. National Heart, Lung, and blood institute scientific statement. 30. Beuers U, Gershwin ME, Gish RG, et al. Changing nomencla- Circulation. 2005;112:2735-­2752. ture for PBC: from “Cirrhosis” to “Cholangitis”. Gastroenterology. 18. Barritt AS, Gitlin N, Klein S, et al. Design and rationale for a real-­ 2015;149:1627-­1629. world observational cohort of patients with nonalcoholic fatty liver disease: The TARGET-­NASH study. Contemp Clin Trials. 2017;61:33-­38. S U P P O R T I N G I N FO R M AT I O N 19. Hardy T, Wonders K, Younes R, et al. The European NAFLD Registry: a real-­world longitudinal cohort study of nonalcoholic Additional supporting information may be found online in the fatty liver disease. Contemp Clin Trials. 2020;98:106175. Supporting Information section. 20. Brunt EM, Kleiner DE, Wilson LA, Belt P, Neuschwander-­Tetri BA. Nonalcoholic fatty liver disease (NAFLD) activity score and the his- topathologic diagnosis in NAFLD: Distinct clinicopathologic mean- How to cite this article: Ampuero J, Aller R, Gallego-­Durán R, ings. Hepatology. 2011;53:810-­820. et al; HEPAmet Registry. Definite and indeterminate 21. Pelusi S, Cespiati A, Rametta R, et al. Prevalence and risk factors of nonalcoholic steatohepatitis share similar clinical features and significant fibrosis in patients with nonalcoholic fatty liver with- prognosis: A longitudinal study of 1893 biopsy-­proven out steatohepatitis. Clin Gastroenterol Hepatol. 2019;17(11):2310-­​ nonalcoholic fatty liver disease subjects. Liver Int. 2319.e6. 22. Nascimbeni F, Bedossa P, Fedchuk L, et al. Clinical validation of the 2021;41:2076–­2086. https://doi.org/10.1111/liv.14898 FLIP algorithm and the SAF score in patients with non-­alcoholic