Journal of the Formosan Medical Association (2021) 120, 68e77

Available online at www.sciencedirect.com

ScienceDirect

journal homepage: www.jfma-online.com

Review Article

Non-alcoholic fatty liver disease: A review

with clinical and pathological correlation
Yen-Ying Chen a,b, Matthew M. Yeh c,d,*

a
Department of Pathology and Laboratory Medicine, Taipei Veterans General Hospital, Taipei, Taiwan
b
School of Medicine, National Yang-Ming University, Taipei, Taiwan
c
Department of Pathology, University of Washington School of Medicine, Seattle, United States
d
Department of Medicine, University of Washington School of Medicine, Seattle, United States

Received 8 January 2020; received in revised form 4 May 2020; accepted 2 July 2020

KEYWORDS Non-alcoholic fatty liver disease (NAFLD) is the most common chronic liver disease in North
Fatty liver; America and Europe, with increasing prevalence in other regions of the world. Its spectrum
Nonalcoholic fatty encompass steatosis, non-alcoholic steatohepatitis (NASH), fibrosis and cirrhosis. It is consid-
liver disease; ered as the manifestation of metabolic syndrome in liver, and its development and progression
Nonalcoholic is influenced by complex interaction of environmental and genetic factors. In this review we
steatohepatitis discuss the histopathological features, differential diagnoses, and the commonly used grading
and staging systems of NAFLD. NAFLD associated with other diseases, histological changes after
therapeutic intervention and recurrence or occurrence of NAFLD after liver transplantation are
also addressed.
Copyright ª 2020, Formosan Medical Association. Published by Elsevier Taiwan LLC. This is an
open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-
nc-nd/4.0/).

Introduction syndrome, but it could affect non-obese patients as well.

The disease spectrum ranges from bland steatosis with or
Non-alcoholic fatty liver disease (NAFLD) is the leading without inflammation (non-alcoholic fatty liver, NAFL) to
cause of chronic liver disease in western countries. As the steatosis plus inflammation and hepatocellular injury, i.e.,
name implies, it develops without significant alcohol non-alcoholic steatohepatitis (NASH), fibrosis and
intake, defined as less than approximately 30 g/day for cirrhosis. In this article we review the pathological fea-
women and 40 g/day for men by the American Association tures, issues to be considered in diagnosis, grading and
for the Study of Liver Disease (AASLD).1 NAFLD is staging system, and recent advances focusing on the pa-
commonly associated with obesity, diabetes and metabolic thology of NAFLD.

* Corresponding author. Department of Pathology, University of Washington School of Medicine, Seattle, WA 98195, United States.
E-mail address: [email protected] (M.M. Yeh).

https://doi.org/10.1016/j.jfma.2020.07.006
0929-6646/Copyright ª 2020, Formosan Medical Association. Published by Elsevier Taiwan LLC. This is an open access article under the CC
BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Non-alcoholic fatty liver disease 69

Epidemiology hepatocellular carcinoma (HCC) was reported to be 0.44

and 5.29 per 1000 person-years in NAFLD and NASH pa-
The prevalence of NAFLD varies among different pop- tients, respectively.2 The adjusted hazard ratio for NAFLD
ulations and is influenced by the methods of diagnosis, liver-specific mortality was 2.60. Fibrosis stage is the most
which include serum biochemistry, imaging and histologic important prognostic factor.13
examination. Liver biopsy is impractical for screening and
the studies using it may be subjected to selection bias.
Studies using blood tests tend to underestimate the prev-
Genetic modifiers
alence of NAFLD compared to those utilizing imaging mo-
dalities. In one meta-analysis, the global prevalence of Numerous genes have been implicated in the development
NAFLD diagnosed by imaging was reported to be 25.24%. and progression of NAFLD. Among them the single nucleo-
The highest prevalence was in Middle East (37.19%) and tide polymorphism (SNP) causing a substitution of isoleu-
South America (30.45), followed by Asia (27.17%), North cine to methionine at position 148 in patatin-like
America (24.13%), Europe (23.17%) and Africa (13.48%).2 phospholipase domain-containing 3 (PNPLA3) is most
Among the NAFLD patients who underwent biopsy, the robustly validated to be associated with NAFLD. PNPLA3 has
prevalence of NASH was 59.1%. Overall, the prevalence of hydrolytic activity towards triacylglycerols, diacylglycerol
NAFLD in western and eastern countries range 20e40% and and monoacylglycerol, and the I148M substitution causes a
7.9e43.3%, respectively.3,4 The prevalence in AsiaePacific loss of function in the enzyme.14 This genetic variant is
region is associated with economic prosperity. A larger associated with higher liver lipid content, greater NASH
proportion of the affected individuals were non-obese activity, and increased risk of liver fibrosis and develop-
compared to North America. In Taiwan the prevalence in ment of hepatocellular carcinoma.15,16 Other widely
general population ranges from 11.5 to 44.5%.5,6 The investigated genetic variants include transmembrane 6 su-
prevalence increases with age.6 Other risk factors associ- perfamily member 2 (TM6SF2) rs58542926 C > T, membrane
ated with NAFLD include male gender, obesity, diabetes, bound O-acyltransferase domain-containing 7 (MBOAT7)
hyperlipidemia, hypertension, elevated alanine amino- rs641738 C > T, and glucokinase regulator (GCKR) P446L.
transferase (ALT) and hyperuricemia.5,6 They are also associated with increased NAFLD severity and
In pediatric population the prevalence of NAFLD is also risk of fibrosis.17e20
on the rise. The prevalence of NAFLD in children has
increased more than 2-fold in the United States over the Pathology of non-alcoholic fatty liver disease
past 2 decades.7 Studies showed that the prevalence of
NAFLD in children ranges from 7.6% in overall population to
34.2% in obese children.8 Hispanic adolescents have a 4-fold Steatosis vs. steatohepatitis
increased risk of hepatic steatosis compared with non-His-
panic.9 White and Asian children have higher prevalence Fatty liver is defined as >5% of steatosis in the liver pa-
than African American children.10 renchyma histologically. The fatty change in NAFLD is pri-
marily macrovesicular, characterized by a single large
droplet occupying the cytoplasm and displacing the nucleus
Natural history to the periphery. In adults the steatosis in NAFLD typically
starts in zone 3, sparing the periportal areas. Mild lobular
There is great heterogeneity in the disease severity and and portal inflammation is not uncommonly seen in asso-
outcome of NAFLD, and only a minority of the patients ciation with steatosis in NAFLD, but this is not equivalent to
progress to cirrhosis or liver-related death. In one system- steatohepatitis (Fig. 1). Hepatocyte ballooning, character-
atic review of paired biopsy studies including 150 NAFL ized by swelling of hepatocyte with rarefied cytoplasm, is
patients and 261 NASH patients, 33.6% had fibrosis pro- required in addition to steatosis and inflammation to
gression over 2145.5 person-years of follow-up, while 43.1% establish the diagnosis of steatohepatitis (Fig. 2). Mallory-
had stable and 22.3% had improved fibrosis stage. In NAFL Denk bodies (MDBs) may be seen in the ballooned hepato-
patients 39.1% had progression and 8.3% had improvement cytes, but are not necessary for diagnosis. Inflammation is
in fibrosis; in NASH patients 34.5% had fibrosis progression more commonly seen in the lobules than in the portal
and 26.7% had improvement.11 About 20% of the progression areas, which is usually mild to moderate and composed of
group had rapid fibrosis progression. In another paired bi- mixed inflammatory infiltrates, predominantly of lympho-
opsy study including 446 patients, NAFL resolved in 12.8% cytes and Kupffer cells. Apoptotic (acidophil) bodies are
and progressed to steatohepatitis in 41.9% of patients over commonly seen, and are associated with inflammation,
a mean interval of 4.9 years. Steatohepatitis resolved in ballooning, and NAFLD activity.21 Satellitosis (neutrophils
about one-fourth of the patients. Fibrosis progression and surrounding ballooned hepatocytes) and microgranuloma
regression was noted in 30% and 34% of patients, respec- can sometimes be observed. In adult NAFLD, fibrosis also
tively. Lower baseline insulin level, baseline aspartate typically begins in zone 3 with the characteristic peri-
aminotransferase (AST), changes in alanine aminotrans- cellular, “chicken wire” pattern (Fig. 3). In cases where
ferase (ALT) and decreasing NAFLD activity score (NAS) NASH has “burnt out” such as in cirrhosis or ballooned cells
were associated with fibrosis regression. Metabolic syn- are indefinite, this fibrosis pattern may be a clue to suggest
drome, baseline NAFLD activity score (NAS), smaller an underlying metabolic etiology.22 Portal/periportal,
reduction in NAS, baseline AST and changes in AST are bridging fibrosis and eventually cirrhosis can subsequently
associated with fibrosis progression.12 Annual incidence of occur as the disease progresses.
70 Y.-Y. Chen, M.M. Yeh

Figure 1 This liver biopsy shows area surrounding central Figure 3 Masson trichrome stain shows perivenular/peri-
vein region with hepatocytes containing large fat droplets, cellular (chicken wire) fibrosis in the lobule, typical fibrosis
occupying the cytoplasm and displacing the nucleus to the pattern in adult NASH.
periphery (short arrows) and few foci of inflammatory cells
(long arrows), composed predominantly of lymphocytes and
Kupffer cells. There are no ballooned hepatocytes, hence a
diagnosis of steatohepatitis cannot be established (Hematox- Iron deposition
ylin and eosin stain). CV: Central vein.
Iron deposition have been reported in 10e55% of NAFLD
patients.25,26 It is usually mild and can occur in hepato-
Megamitochondria cytes, reticuloendothelial cells or both. There have been
conflicting results on the relationship between iron accu-
Megamitochondria are eosinophilic round or crystal shaped mulation pattern and fibrosis. Nelson et al. showed that
intracytoplasmic structures. They are sometimes observed patients with reticuloendothelial iron deposition were more
in NASH, usually within hepatocytes with microvesicular likely to have advanced fibrosis compared to those with
steatosis. This finding is not specific to NASH and may hepatocellular iron deposition.27 On the contrary, Valenti
reflect a response to mitochondrial/oxidation injury in he- et al. demonstrated that iron accumulation in hepatocytes
patocytes.23 The distribution of megamitochondria is equal but not non-parenchymal siderosis was associated with
in all zones and shows no difference between low and high fibrosis.28 The impact of iron deposition, genetic mutation
fibrosis stage.24 of HFE and insulin resistance on fibrosis in NAFLD are still
under investigation.26e28

Pediatric NAFLD

In a study analyzing 100 liver biopsies from children with

NAFLD, two different histologic forms of steatohepatitis
have been described in pediatric patients. Type 1 was
described in 17% of patients and showed similar features as
adult NASH, while type 2 was described in 51% of patients
and showed more severe steatosis, less hepatocyte
ballooning, more portal inflammation and zone 1 fibrosis
pattern (Fig. 4) but not pericentral sinusoidal fibrosis seen
in adults (zone 3 pattern, Fig. 3).29 Patients with type 2
NASH are younger, more obese, and more of male gender
and Asian, Native American, and Hispanic ethnicity
compared to those with type 1 NASH. However, subsequent
Figure 2 This liver biopsy shows steatosis, lobular inflam- studies have shown that overlapping features can be
matory foci, and ballooning of hepatocytes (short arrows), with observed in a large proportion (51e82%) of the pediatric
enlarged and rarefied cytoplasm, hence fulfilling the diagnosis patients.30,31 This finding indicates that although zone 1
of steatohepatitis. Mallory-Denk bodies (MDBs, long arrows) pattern is more commonly seen in pediatric NAFLD, zone 3
showing ropy and eosinophilic materials are also present in few injury is still present to some extent in many of the cases,
ballooned hepatocytes, but MDBs are not required for diagnosis suggesting the zonal pattern may evolve or transition as
of steatohepatitis (Hematoxylin and eosin stain). patients grow from children through adolescents to adults.
Non-alcoholic fatty liver disease 71

portal inflammation can be observed in NAFLD at times. As

mentioned earlier, it is commonly seen in pediatric NAFLD.
In adult patients, increased portal inflammation can be
seen in post-treatment changes, or associated with more
advanced disease in NAFLD.34 It should also raise the
consideration of other concomitant chronic liver disease,
such as hepatitic B and C, autoimmune hepatitis, and pri-
mary biliary cholangitis, etc.

Differential diagnosis

Alcoholic liver disease (ALD)

Correlation with clinical history is essential to distinguish

ALD vs NAFLD as they share many histologic features and
Figure 4 Pediatric NASH shows steatosis surrounding portal can be almost indistinguishable histologically. However,
tracts and zone 1 fibrosis pattern (Masson trichrome stain). CV: steatohepatitis in ALD tends to be more severe than that in
Central vein. PT: portal tract. NAFLD, and some features including sclerosing hyaline ne-
crosis, outflow vein occlusion, lymphocytic phlebitis, alco-
holic foamy degeneration and acute cholestasis were
Special stains described in ALD but not in NAFLD.35 On the other hand,
marked steatosis, periportal glycogenated nuclei and lipo-
Identification of ballooned hepatocytes can be difficult granulomas are more commonly seen in NAFLD.36 Steatosis
sometimes with inter-observer variability.22 Immunohisto- can be absent in ALD, but it is required for the diagnosis of
chemical stains have been proposed to better demonstrate NAFLD.37 One study has utilized immunostains for insulin
the ballooning degeneration, including CK8/18, ubiquitin receptor (IR) and protein tyrosine phosphatase 1B (PTP1B)
and p62. The CK8/18 staining seen in the cytoplasm of to distinguish between ASH and NASH. PTP1B is a negative
normal hepatocytes is markedly reduced or lost in bal- regulator of IR, and increased PTP1B expression and
looned hepatocyte.32 MDBs, often seen in ballooned hepa- decreased IR expression are seen in NASH patients
tocytes, can be stained with the CK8/18, ubiquitin and p62 compared to ASH.38 Whether these immunohistochemical
immunohistochemical stains. Since MDBs are not always stains are clinically useful awaits further validation.
present in ballooned hepatocytes, the latter two stains may
only help in detecting small, poorly formed MDBs, without
further increasing the detection of ballooned Hepatitis C
hepatocytes.32
Steatosis is present in 40e86% of hepatitis C virus (HCV)
infected patients, with the highest prevalence in patients
Diagnostic pitfalls with HCV genotype 3 infection.39 Steatosis in HCV genotype
3 infected patients is associated with viral load and is
Microvesicular steatosis therefore considered to be secondary to viral effect,40
while in non-genotype 3 patients it is likely related to
Microvesicular steatosis is characterized by multiple tiny host factors. Histologically, the perivenular fat distribution,
fat droplets in the cytoplasm without displacing the nu- ballooned hepatocytes and pericellular fibrosis character-
cleus. It is occasionally seen in coexistence with macro- istic of NAFLD are usually not specific to hepatitis C, and
vesicular steatosis in NAFLD, and its presence is associated the inflammation in hepatitic C is portal-accentuated.
with more advanced disease histology.33 Microvesicular
steatosis can sometimes be confused with ballooning
Autoimmune hepatitis
degeneration. Megamitochondria in the former and
MalloryeDenk bodies in the latter, though not always pre-
sent, may be helpful to make the distinction.22 Diffuse Elevated serum autoantibodies, when detected, can lead to
microvesicular steatosis is not a feature of NAFLD and concerns for autoimmune hepatitis. However, autoanti-
should raise the consideration of conditions associated with bodies such as antinuclear antibody (ANA) and smooth
liver failure secondary to mitochondria dysfunction, such as muscle antibody (SMA) can be present in about 20% of
acute liver failure of pregnancy, Reye syndrome, drug and NAFLD patients in the absence of autoimmune hepatitis.41
toxin injury, among others. In two studies including 71 and 225 NAFLD patients,
respectively, the presence of autoantibodies is associated
with higher inflammatory activity,42,43 whereas the pres-
Portal-accentuated inflammation ence of autoantibodies is not associated with more
advanced histologic features in a study including 864 NAFLD
Inflammation in adult NAFLD is typically more prominent in patients.41 The differences may be attributed to different
the lobules and mild in portal areas, however, increased diagnostic criteria or cohorts from different regions.
72 Y.-Y. Chen, M.M. Yeh

Steatotic lesions specimens.22 The presence of unpaired arteries is often

used as a clue for hepatocellular neoplasm, but it can
Some fat-containing hepatocellular lesions can be confused mimic the centrizonal arteries in NASH, a potential diag-
with NAFLD histologically, including focal nodular hyper- nostic pitfall pathologists need to be aware of. Reticulin
plasia (FNH), hepatocellular adenoma and steatohepatitic stain is also of limited use in this setting because reticulin
variant of HCC (SH-HCC). Correlation with radiological and loss can be seen in benign steatosis.53 Absence of portal
clinical information is of great importance. tracts, and the presence of sinusoidal capillarization and
Variable degree of steatosis can be seen in FNH,44 a positive glypican-3 immunohistochemical stain may help in
benign hepatocellular lesion frequently seen in younger distinguishing from steatohepatitis.50
women, characterized by a well-circumscribed nodule
composed of a central scar containing inflammatory cells,
thick-walled arteries and proliferating ductules, and benign Grading and staging of NAFLD
hepatocytes within the nodular parenchyma. A compensa-
tory response of the hepatic parenchyma to vascular injury Histologic system
is thought to be the underlying etiology. A recent study has
also reported steatohepatitis-like changes with ballooned In 1999 Brunt et al. first proposed a grading and staging
hepatocytes and MDBs in addition to steatosis in FNHs.45 system for NASH based on the liver biopsies from 51 adult
The thick-walled vessels within the fibrous septa and the patients.54 A three-tiered grading system was built based
characteristic “map-like” staining pattern of glutamine on the compilation of steatosis, lobular and portal inflam-
synthetase typical for FNH can help distinguish it from mation, and ballooning degeneration (Table 1A). Fibrosis is
NAFLD. staged based on the location and extent. Stage 1 represents
Steatosis is common in hepatocellular adenoma, espe- zone 3 perisinusoidal/pericellular fibrosis; stage 2 is
cially in hepatocyte nuclear factor (HNF)1a-inactivated assigned when there is stage 1 plus periportal fibrosis;
hepatocellular adenoma. It occurs more commonly in young bridging fibrosis is designated as stage 3; and stage 4 is
women and are associated with maturity-onset diabetes of established where there is evidence of cirrhosis (Table 1B).
the young type 3 (MODY3).46 Pathologically it is character- In 2005, the NASH Clinical Research Network (CRN)
ized by prominent steatosis, and cytologic atypia and designed and validated a NAFLD Activity Score (NAS). The
inflammation are typically absent or mild. Loss of liver fatty scoring system derived from the Brunt system and NAS is
acid binding protein (L-FABP) staining in hepatocytes is the summation of the scores for steatosis (0e3), lobular
useful in establishing the diagnosis.47 inflammation (0e3), and ballooning (0e2) (Table 2A).55 The
Fatty change, as well as morphological features remi- staging of fibrosis is also similar to the Brunt system, but
niscent of steatohepatitis is commonly seen in HCC, known stage 1 is further subdivided to separate zone 3 peri-
as steatohepatitic variant of HCC (SH-HCC), occurring sinusoidal (1a), dense zone 3 perisinusoidal (1b) and portal-
frequently in patients with fatty liver disease, but can also only fibrosis (1c) to reflect the fibrosis pattern in pediatric
be seen in other chronic liver disease, such as hepatitis C. NASH (Table 2B). It needs to be emphasized that NAS re-
Its association with NAFLD and metabolic syndrome has flects the disease activity and should not be used as a
been reported in a few studies.48e50 However, it can also substitute of histological diagnosis for NASH. A validation
develop in patients without NAFLD and metabolic syn- study has also shown that NAS does not always correlate
drome, and novel chromosomal alterations were found in a with the diagnosis of steatohepatitis.56 The SAF score
subset of these cases.51 Another genetic study has shown (steatosis, activity, fibrosis) was proposed in 2012 by the
that SH-HCC has frequent IL-6/JAK/STAT activation but no European Fatty Liver Inhibition of Progression consortium.57
CTNNB1, TERT and TP53 pathway alterations.52 SH-HCCs In this scoring system, steatosis is dissociated from activity
share the pathological features of steatohepatitis, score, and disease activity is reflected only by lobular
including macrovesicular steatosis, pericellular fibrosis, inflammation and hepatocellular ballooning. A diagnostic
hepatocyte ballooning, Mallory-Denk bodies and inflam- algorithm for NAFLD and NASH is also developed based on
mation. They are often low-grade and may be under- the SAF scoring system. It is important to note the diagnosis
diagnosed as mere steatohepatitis, especially in biopsy of NASH needs to be based on the pattern of injury and not

Table 1A Steatohepatitis: Grading by Brunt system (modified from Brunt et al.54).

Grade Steatosis Ballooning Inflammation

1 (Mild) 1-2 (up to 66%) Minimal L: 1-2
P: None-mild
2 (Moderate) 2-3 (>33%, may be >66%) Present L: 2
P: Mild-moderate
3 (Severe) 2e3 Marked L: 3
P: Mild-moderate
Steatosis Grade 1:
33%; Grade 2: >33% < 66%; Grade 3: 66%.
L (lobular): 0, None; 1, <2 foci/20X objective; 2, 2e4 foci/20X objective; 3, >4 foci/20X objective.
P: portal.
Non-alcoholic fatty liver disease 73

Table 1B Steatohepatitis: Staging by Brunt system (modified from Brunt et al.54).

Stage Zone 3, peri-sinusoidal Portal-based Bridging Cirrhosis

1 Focal or extensive 0 0 0
2 Focal or extensive Focal or extensive 0 0
3 Bridging septa Bridging septa þ 0
4   Extensive þ

Table 2A NAFLD Activity Score (NAS) by NASH CRN modalities to detect steatosis is better than blood bio-
(modified from Kleiner et al.55). markers. MR-imaging, including magnetic resonance spec-
Score
troscopy (MRS) and MRI-Proton density fat fraction (MRI-
PDFF), is most accurate for quantification of steatosis.61
Steatosis However, it is expensive and the use in routine clinical
<5% 0 care is limited. Conventional ultrasound is not sensitive in
5e33% 1 detecting steatosis, but is widely available and has a high
34e66% 2 specificity for detection of moderate to severe hepatic
>66% 3 steatosis when positive. Controlled attenuation parameter
Lobular inflammation (foci per field using 20 objective) (CAP) can be used to detect steatosis, but is inferior to MR
0 0 imaging in grading of steatosis.61
<2 1 Currently no imaging modality is reliable for diagnosing
2-4 2 NASH. The presence of metabolic syndrome is a strong
>4 3 predictor for the presence of NASH in patients with NAFLD.1
Ballooning Cytokeratin 18 (CK18) is the most extensively studied
None 0 biomarker for NASH diagnosis. In one meta-analysis
Few 1 including 11 studies, the pooled sensitivity and specificity
Many 2 of diagnosing NASH is 66% and 82%, respectively.58 Combi-
nation of CK18 with other serum biomarkers is also
commonly used. In a small study, combination of CK18,
merely the lesion itself. In cases where the diagnosis is adiponectin and interleukin-6 was reported to have a
established, this scoring system is reproducible. sensitivity of 84.5% and specificity of 85.7% for diagnosing
NASH, with an AUROC of 0.903.59 Other biomarkers such as
CXCL10, tumor necrosis factor-a(TNF-a), interleukin-8 and
Non-invasive systems fibroblast growth factor 21 (FGF21) had also been shown to
have moderate accuracy for diagnosing NASH.59 Many
Liver biopsy is currently considered the gold standard of biomarker panels had also been developed. The NASH Test,
diagnosis of NASH, however, it is invasive and limited by for example, is an algorithm combining 13 parameters to
cost. Many efforts have therefore been made in search for predict the presence of NASH, and the reported AUROC for
non-invasive methods to diagnose and monitor the disease. NASH was 0.79.60 Currently the NASH biomarkers are still
For the evaluation of steatosis, several biomarker panels under investigation and none of them are being used in
have been developed, including fatty liver index (FLI), he- routine clinical practice.
patic steatosis index (HSI) and SteatoTest. These panels For the assessment of advanced fibrosis, several clinical
show modest accuracy to detect fatty liver, with area under scoring systems, serum biomarkers and imaging modalities
the receiver-operating characteristic curve (AUROC) have been investigated. In one study, NAFLD fibrosis score
ranging from 0.79 to 0.84.60 The performance of imaging (NFS, based on age, BMI, hyperglycemia, platelet count,
albumin, and AST/ALT ratio) and fibrosis-4 index (FIB-4,
Table 2B Stage system by NASH CRN (modified from based on platelet count, age, AST, and ALT) outperformed
Kleiner et al.55). other indices such as BARD score, AST to Platelet Ratio
Index (APRI) and AST/ALT ratio in predicting advanced
Extent of fibrosis Stage fibrosis.62 VCTE (fibroscan) is an elastographic method for
None 0 measuring liver stiffness which has standardized quality
Mild, zone 3 perisinusoidal 1a criteria for application. However the failure rate in obese
(requires trichrome stain to visualize) patients is high.61 Overall, magnetic resonance elastog-
Moderate, zone 3 perisinusoidal 1b raphy (MRE) is the best imaging modality to assess fibrosis,
(easily visualized by H&E stain) but its application in clinical practice is limited by the cost
Portal fibrosis only 1c and availability of MRI facilities.60
Zone 3 perisinusoidal and periportal 2 At present, none of these non-invasive modalities can
Bridging fibrosis 3 reliably differentiate NASH from simple steatosis, and their
Cirrhosis 4 ability to detect early fibrosis is limited.60 Therefore liver
biopsy is still essential for the diagnosis of NASH and
74 Y.-Y. Chen, M.M. Yeh

evaluating the degree of necroinflammatory activity, liver asymptomatic liver enzyme abnormalities to acute liver
cell injury and fibrosis. Liver biopsy can also provide in- failure or cirrhosis. Steatosis, glycogenated nuclei,
formation of concurrent liver disease. Current clinical ballooning of hepatocytes and Mallory-Denk bodies can be
guidelines recommend liver biopsy in NAFLD patients at seen in Wilson disease, which may resemble NAFLD. Other
increased risk of having NASH and/or advanced fibrosis, and histologic patterns include chronic hepatitis, acute hepa-
also in patients in whom concurrent liver disease cannot be titis and cirrhosis.74 Wilson disease should especially be
excluded clinically.1 considered in young patients with unexplained liver dis-
ease. The diagnostic tests include serum ceruloplasmin,
24 h urine copper excretion, tissue copper quantitation and
NAFLD in other clinical settings
genetic testing.

Drug-induced fatty liver disease

Regression after treatment
Many therapeutic agents can cause macrovesicular stea-
tosis in liver, including amiodarone, chemotherapeutic Current treatment of NAFLD include weight reduction,
agents (5-fluorouracil, irinotecan, cisplatin, asparaginase), lifestyle modification, medication and surgical interven-
glucocorticoids, methotrexate, tamoxifen, total parenteral tion.75,76 Clinical trials with lifestyle intervention, insulin
nutrition (TPN), and highly active antiretrovial therapy sensitizing agents, antioxidants and bariatric surgery have
(HAART) for HIV infection.63e65 Most of these agents can shown varying degree of improvement in steatosis, inflam-
also cause steatohepatitis or even lead to fibrosis, although mation, ballooning and fibrosis.76e80 It has also been noted
the causality could be difficult to establish given the high that there was a shift of lobular to portal inflammation
prevalence of pre-existing NAFLD, and the true drug- after treatment intervention.34 However, reversal of im-
induced steatohepatitis might be uncommon. The drug- provements have been reported after discontinuing medi-
associated fatty liver disease has similar histologic fea- cation treatment. In a follow-up study of NASH patients
tures as adult NAFLD, but the fat distribution may not be treated with pioglitazone, worsening of inflammation and
accentuated in perivenular areas.66 Some drugs including steatosis and recurrence of NASH were noted at repeat liver
methotrexate, tamoxifen and glucocorticoid can lead to biopsy at least 48 weeks after treatment discontinuation,
worsening of the underlying NAFLD.63 but there was no change in fibrosis.81 On the other hand, in
a prospective study on the long-term effect after bariatric
surgery, the improvement of NASH is sustained for 5 years
Inflammatory bowel disease (IBD)
but fibrosis is increased.80
The overall prevalence of NAFLD in IBD patients was re-
ported to be 27.5% in a meta-analysis based on 19 studies,67 Recurrent/de novo NAFLD after liver
and older age, metabolic risk factors, methotrexate use, transplantation
prior surgery and longer duration of IBD were identified to
be associated with NAFLD in IBD patients. Conversely, The recurrence or occurrence of NAFLD after orthotopic
another meta-analysis based on 7 studies found no associ- liver transplantation (OLT) has frequently been
ation between IBD medications, including methotrexate, reported.82e85 In some cases it may be difficult to deter-
and the incidence of NAFLD.68 If compared to non-IBD mine whether the NAFLD is recurrent or de novo, as the
NAFLD patients, IBD patients with NAFLD are younger and histological features can be burnt out and the etiology of
have fewer metabolic risk factors.69 Comorbid primary cirrhosis can be hard to ascertained in the native liver. In
sclerosing cholangitis has been reported to protect against patients transplanted for NAFLD or cryptogenic cirrhosis,
NAFLD in IBD patients.70 the incidence of recurrent steatosis can be up to 100%.82
The risk factors that had been reported for developing
Celiac disease (CD) NAFLD after transplantation in multivariate analysis include
obesity, hyperlipidemia, diabetes, hypertension, hepatitis
Patients with celiac disease have an increased risk for C, tacrolimus or sirolimus-based immunosuppression, pre-
NAFLD compared to the general population. In a Swedish LT alcoholic cirrhosis and liver graft steatosis.83,85 The
population-based study, celiac disease patients had a 2.8- recurrence or occurrence of NAFLD does not seem to affect
fold increased risk of NAFLD, with the highest risk estimates patient survival.84,85
in children.71 Another case-control study also confirmed an
increased risk of NAFLD in celiac disease patients.72 The Future perspectives
precise pathogenesis of NAFLD in celiac disease is still un-
clear. On the other hand, the prevalence of celiac disease As the epidemic of NAFLD continues to grow, almost every
in NAFLD patients is also significantly higher than the gen- aspect of NAFLD is currently under intense research. From
eral population.73 the pathologists’ perspective, the area making the most
progress in recent years may be the genetics of NAFLD.
Wilson disease Multiple gene loci as well as epigenetic factors associated
with NAFLD susceptibility and progression have been
Wilson disease is an inherited disorder of copper meta- discovered.86 The search for more genetic and epigenetic
bolism. The liver presentation can range from determinants is still ongoing, and these new discoveries can
Non-alcoholic fatty liver disease 75

lead to more clinical implications and translational re- 10. Vos MB, Abrams SH, Barlow SE, Caprio S, Daniels SR, Kohli R,
searches. Another area that is gaining attention is the use et al. NASPGHAN clinical practice guideline for the diagnosis
of artificial intelligence and digital pathology to aid in and treatment of nonalcoholic fatty liver disease in children:
pathological diagnosis. Digital image analysis has been used recommendations from the expert committee on NAFLD
(ECON) and the North American society of pediatric gastroen-
to quantify liver fibrosis and steatosis in NAFLD
terology, Hepatology and Nu. J Pediatr Gastroenterol Nutr
patients.87,88 2017;64(2):319e34.
The use of second harmonic generation/two-photon 11. Singh S, Allen AM, Wang Z, Prokop LJ, Murad MH, Loomba R.
excitation fluorescence microscopy (SHG/TPFE) can Fibrosis progression in nonalcoholic fatty liver vs nonalcoholic
further improve the evaluation of fibrosis by eliminating the steatohepatitis: a systematic review and meta-analysis of
staining variability that might be encountered in traditional paired-biopsy studies. Clin Gastroenterol Hepatol 2015;13(4):
histochemical stains, and studies have demonstrated that 643e54. e9.
distinguishing different fibrosis stages based on NASH CRN 12. Kleiner DE, Brunt EM, Wilson LA, Behling C, Guy C, Contos M,
system in NAFLD patient is feasible.89,90 Introduction of et al. Association of histologic disease activity with progression
these new technologies can improve the accuracy of diag- of nonalcoholic fatty liver disease. JAMA Netw Open 2019;
2(10). e1912565ee1912565.
nosis and avoid inter-observer variability. Their future
13. Angulo P, Kleiner DE, Dam-Larsen S, Adams LA, Bjornsson ES,
implication in routine diagnosis, clinical trials and assess- Charatcharoenwitthaya P, et al. Liver fibrosis, but no other
ment of progression/regression of fibrosis in NAFLD may be histologic features, is associated with long-term outcomes of
expected. patients with nonalcoholic fatty liver disease. Gastroenter-
ology 2015;149(2):389e97. e10.
14. Huang Y, Cohen JC, Hobbs HH. Expression and characterization
of a PNPLA3 protein isoform (I148M) associated with nonalco-
Declaration of Competing Interest holic fatty liver disease. J Biol Chem 2011;286(43):37085e93.
15. Sookoian S, Pirola CJ. Meta-analysis of the influence of I148M
The authors have no conflicts of interest relevant to this variant of patatin-like phospholipase domain containing 3 gene
article. (PNPLA3) on the susceptibility and histological severity of
nonalcoholic fatty liver disease. Hepatology 2011;53(6):
1883e94.
References 16. Singal AG, Manjunath H, Yopp AC, Beg MS, Marrero JA, Gopal P,
et al. The effect of PNPLA3 on fibrosis progression and devel-
1. Chalasani N, Younossi Z, Lavine JE, Charlton M, Cusi K, opment of hepatocellular carcinoma: a meta-analysis. Am J
Rinella M, et al. The diagnosis and management of nonalco- Gastroenterol 2014;109(3):325e34.
holic fatty liver disease: practice guidance from the American 17. Kozlitina J, Smagris E, Stender S, Nordestgaard BG, Zhou HH,
Association for the Study of Liver Diseases. Hepatology 2018; Tybjærg-Hansen A, et al. Exome-wide association study iden-
67(1):328e57. tifies a TM6SF2 variant that confers susceptibility to nonalco-
2. Younossi ZM, Koenig AB, Abdelatif D, Fazel Y, Henry L, holic fatty liver disease. Nat Genet 2014;46:352e6.
Wymer M. Global epidemiology of nonalcoholic fatty liver dis- 18. Liu YL, Reeves HL, Burt AD, Tiniakos D, McPherson S,
easedmeta-analytic assessment of prevalence, incidence, and Leathart JBS, et al. TM6SF2 rs58542926 influences hepatic
outcomes. Hepatology 2016;64(1):73e84. fibrosis progression in patients with non-alcoholic fatty liver
3. Sherif ZA, Saeed A, Ghavimi S, Nouraie SM, Laiyemo AO, disease. Nat Commun 2014;5:1e6.
Brim H, et al. Global epidemiology of nonalcoholic fatty liver 19. Mancina RM, Dongiovanni P, Petta S, Pingitore P, Meroni M,
disease and perspectives on US minority populations. Dig Dis Rametta R, et al. The MBOAT7-TMC4 variant rs641738 increases
Sci 2016;61(5):1214e25. risk of nonalcoholic fatty liver disease in individuals of Euro-
4. Wong MCS, Huang JLW, George J, Huang J, Leung C, Eslam M, pean descent. Gastroenterology 2016;150(5):1219e30. e6.
et al. The changing epidemiology of liver diseases in the 20. Petta S, Miele L, Bugianesi E, Cammà C, Rosso C, Boccia S,
AsiaePacific region. Nat Rev Gastroenterol Hepatol 2019; et al. Glucokinase regulatory protein gene polymorphism af-
16(1):57e73. fects liver fibrosis in non-alcoholic fatty liver disease. PloS One
5. Chen CH, Huang MH, Yang JC, Nien CK, Yang CC, Yeh YH, et al. 2014;9(2):1e7.
Prevalence and risk factors of nonalcoholic fatty liver disease 21. Yeh MM, Belt P, Brunt EM, Kowdley KV, Wilson LA, Ferrell L.
in an adult population of Taiwan: metabolic significance of Acidophil bodies in nonalcoholic steatohepatitis. Hum Pathol
nonalcoholic fatty liver disease in nonobese adults. J Clin 2016;52:28e37.
Gastroenterol 2006;40(8):745e52. 22. Koch LK, Yeh MM. Nonalcoholic fatty liver disease (NAFLD):
6. Huang JF, Tsai PC, Yeh ML, Huang CF, Huang CI, Hsieh MH, et al. diagnosis, pitfalls, and staging. Ann Diagn Pathol 2018;37:
Risk stratification of non-alcoholic fatty liver disease across 83e90.
body mass index in a community basis [Internet] J Formos Med 23. Caldwell SH, Chang CY, Nakamoto RK, Krugner-Higby L. Mito-
Assoc 2020;119(1):89e96. chondria in nonalcoholic fatty liver disease. Clin Liver Dis
7. Welsh JA, Karpen S, Vos MB. Increasing prevalence of nonal- 2004;8(3):595e617.
coholic fatty liver disease among United States adolescents, 24. Le TH, Caldwell SH, Redick JA, Sheppard BL, Davis CA,
1988e1994 to 2007e2010. J Pediatr 2013;162(3):496e500.e1. Arseneau KO, et al. The zonal distribution of mega-
8. Anderson EL, Howe LD, Jones HE, Higgins JPT, Lawlor DA, mitochondria with crystalline inclusions in nonalcoholic stea-
Fraser A. The prevalence of non-alcoholic fatty liver disease in tohepatitis. Hepatology 2004;39(5):1423e9.
children and adolescents: a systematic review and meta- 25. Chitturi S, Weltman M, Farrell GC, McDonald D, Liddle C,
analysis. PloS One 2015;10(10):e0140908. Samarasinghe D, et al. HFE mutations, hepatic iron, and
9. Rehm JL, Connor EL, Wolfgram PM, Eickhoff JC, Reeder SB, fibrosis: ethnic-specific association of NASH with C282Y but not
Allen DB. Predicting hepatic steatosis in a racially and ethni- with fibrotic severity. Hepatology 2002;36(1):142e9.
cally diverse cohort of adolescent girls. J Pediatr 2014;165(2): 26. Bugianesi E, Manzini P, D’Antico S, Vanni E, Longo F, Leone N,
319e325.e1. et al. Relative contribution of iron burden, HFE mutations, and
76 Y.-Y. Chen, M.M. Yeh

insulin resistance to fibrosis in nonalcoholic fatty liver. Hep- from steatohepatitic variant of hepatocellular carcinoma. Am
atology 2004;39(1):179e87. J Surg Pathol 2017;41(2):277e81.
27. Nelson JE, Wilson L, Brunt EM, Yeh MM, Kleiner DE, Unalp- 46. Zucman-Rossi J, Jeannot E, VanNhieu JT, Scoazec J-Y,
Arida A, et al. Relationship between the pattern of hepatic iron Guettier C, Rebouissou S, et al. Genotypeephenotype corre-
deposition and histological severity in nonalcoholic fatty liver lation in hepatocellular adenoma: new classification and
disease. Hepatology 2011;53(2):448e57. relationship with HCC. Hepatology 2006;43(3):515e24.
28. Valenti L, Fracanzani AL, Bugianesi E, Dongiovanni P, 47. Bioulac-Sage P, Rebouissou S, Thomas C, Blanc J-F, Saric J, Sa
Galmozzi E, Vanni E, et al. HFE genotype, parenchymal iron Cunha A, et al. Hepatocellular adenoma subtype classification
accumulation, and liver fibrosis in patients with nonalcoholic using molecular markers and immunohistochemistry. Hepatol-
fatty liver disease. Gastroenterology 2010;138(3):905e12. ogy 2007;46(3):740e8.
29. Schwimmer JB, Behling C, Newbury R, Deutsch R, Nievergelt C, 48. Salomao M, Yu WM, Brown Jr RS, Emond JC, Lefkowitch JH.
Schork NJ, et al. Histopathology of pediatric nonalcoholic fatty Steatohepatitic hepatocellular carcinoma (SH-HCC): a distinc-
liver disease. Hepatology 2005;42(3):641e9. tive histological variant of HCC in hepatitis C virus-related
30. Nobili V, Marcellini M, Devito R, Ciampalini P, Piemonte F, cirrhosis with associated NAFLD/NASH. Am J Surg Pathol
Comparcola D, et al. NAFLD in children: a prospective clinical- 2010;34(11):1630e6.
pathological study and effect of lifestyle advice. Hepatology 49. Salomao M, Remotti H, Vaughan R, Siegel AB, Lefkowitch JH,
2006;44(2):458e65. Moreira RK. The steatohepatitic variant of hepatocellular
31. Carter-Kent C, Yerian LM, Brunt EM, Angulo P, Kohli R, Ling SC, carcinoma and its association with underlying steatohepatitis.
et al. Nonalcoholic steatohepatitis in children: a multicenter Hum Pathol 2012;43(5):737e46.
clinicopathological study. Hepatology 2009;50(4):1113e20. 50. Jain D, Nayak NC, Kumaran V, Saigal S. Steatohepatitic hepa-
32. Lackner C, Gogg-Kamerer M, Zatloukal K, Stumptner C, tocellular carcinoma, a morphologic indicator of associated
Brunt EM, Denk H. Ballooned hepatocytes in steatohepatitis: metabolic risk factors: a study from India. Arch Pathol Lab Med
the value of keratin immunohistochemistry for diagnosis. J 2013;137(7):961e6.
Hepatol 2008;48(5):821e8. 51. Yeh MM, Liu Y, Torbenson M. Steatohepatitic variant of hepa-
33. Tandra S, Yeh MM, Brunt EM, Vuppalanchi R, Cummings OW, tocellular carcinoma in the absence of metabolic syndrome or
Ünalp-Arida A, et al. Presence and significance of micro- background steatosis: a clinical, pathological, and genetic
vesicular steatosis in nonalcoholic fatty liver disease. J Hep- study. Hum Pathol 2015;46(11):1769e75.
atol 2011;55(3):654e9. 52. Calderaro J, Couchy G, Imbeaud S, Amaddeo G, Letouzé E,
34. Brunt EM, Kleiner DE, Wilson LA, Unalp A, Behling CE, Lavine JE, Blanc J-F, et al. Histological subtypes of hepatocellular carci-
et al. Portal chronic inflammation in nonalcoholic fatty liver noma are related to gene mutations and molecular tumour
disease (NAFLD): a histologic marker of advanced NAFLDdcli- classification. J Hepatol 2017;67(4):727e38.
nicopathologic correlations from the nonalcoholic steatohepa- 53. Singhi AD, Jain D, Kakar S, Wu T-T, Yeh MM, Torbenson M.
titis clinical research network. Hepatology 2009;49(3):809e20. Reticulin loss in benign fatty liver: an important diagnostic
35. Brunt EM. Alcoholic and nonalcoholic steatohepatitis. Clin pitfall when considering a diagnosis of hepatocellular carci-
Liver Dis 2002;6(2):399e420. noma. Am J Surg Pathol 2012;36(5):710e5.
36. Nakano M, Fukusato T. Histological study on comparison be- 54. Brunt EM, Janney CG, DiBisceglie AM, Neuschwander-Tetri BA,
tween NASH and ALD. Hepatol Res 2005;33(2):110e5. Bacon BR. Nonalcoholic steatohepatitis: a proposal for grading
37. Yeh MM, Brunt EM. Pathology of nonalcoholic fatty liver dis- and staging the histological lesions. Am J Gastroenterol 1999;
ease. Am J Clin Pathol 2007;128(5):837e47. 94(9):2467e74.
38. Sanderson SO, Smyrk TC. The use of protein tyrosine phos- 55. Kleiner DE, Brunt EM, VanNatta M, Behling C, Contos MJ,
phatase 1B and insulin receptor immunostains to differentiate Cummings OW, et al. Design and validation of a histological
nonalcoholic from alcoholic steatohepatitis in liver biopsy scoring system for nonalcoholic fatty liver disease. Hepatology
specimens. Am J Clin Pathol 2005;123(4):503e9. 2005;41(6):1313e21.
39. Adinolfi LE, Rinaldi L, Guerrera B, Restivo L, Marrone A, 56. Brunt EM, Kleiner DE, Wilson LA, Belt P, Neuschwander-Tetri BA.
Giordano M, et al. NAFLD and NASH in HCV infection: preva- (CRN)NCRN. Nonalcoholic fatty liver disease (NAFLD) activity
lence and significance in hepatic and extrahepatic manifesta- score and the histopathologic diagnosis in NAFLD: distinct clin-
tions. Int J Mol Sci 2016;17(6):1e12. icopathologic meanings. Hepatology 2011;53(3):810e20.
40. Rubbia-Brandt L, Quadri R, Abid K, Giostra E, Malé P-J, 57. Bedossa P, Poitou C, Veyrie N, Bouillot J-L, Basdevant A,
Mentha G, et al. Hepatocyte steatosis is a cytopathic effect of Paradis V, et al. Histopathological algorithm and scoring system
hepatitis C virus genotype 3. J Hepatol 2000;33(1):106e15. for evaluation of liver lesions in morbidly obese patients.
41. Vuppalanchi R, Gould RJ, Wilson LA, Unalp-Arida A, Hepatology 2012;56(5):1751e9.
Cummings OW, Chalasani N, et al. Clinical significance of serum 58. Kwok R, Tse Y-K, Wong GL-H, Ha Y, Lee AU, Ngu MC, et al.
autoantibodies in patients with NAFLD: results from the Systematic review with meta-analysis: non-invasive assess-
nonalcoholic steatohepatitis clinical research network. Hep- ment of non-alcoholic fatty liver disease e the role of transient
atol Int 2012;6(1):379e85. elastography and plasma cytokeratin-18 fragments. Aliment
42. Niwa H, Sasaki M, Haratake J, Kasai T, Katayanagi K, Pharmacol Ther 2014;39(3):254e69.
Kurumaya H, et al. Clinicopathological significance of antinu- 59. Zhou JH, Cai JJ, She ZG, Li HL. Noninvasive evaluation of
clear antibodies in non-alcoholic steatohepatitis. Hepatol Res nonalcoholic fatty liver disease: current evidence and prac-
2007;37(11):923e31. tice. World J Gastroenterol 2019;25(11):1307e26.
43. Adams LA, Lindor KD, Angulo P. The prevalence of autoanti- 60. Wong VW-S, Adams LA, deLédinghen V, Wong GL-H, Sookoian S.
bodies and autoimmune hepatitis in patients with nonalcoholic Noninvasive biomarkers in NAFLD and NASH d current progress
fatty liver disease. Am J Gastroenterol 2004;99(7):1316e20. and future promise. Nat Rev Gastroenterol Hepatol 2018;
44. Nguyen BN, Fléjou J-F, Terris B, Belghiti J, Degott C. Focal 15(8):461e78.
nodular hyperplasia of the liver: a comprehensive pathologic 61. Loomba R. Role of imaging-based biomarkers in NAFLD: recent
study of 305 lesions and recognition of new histologic forms. advances in clinical application and future research directions.
Am J Surg Pathol 1999;23(12):1441e54. J Hepatol 2018;68(2):296e304.
45. Deniz K, Moreira RK, Yeh MM, Ferrell LD. Steatohepatitis-like 62. Imajo K, Kessoku T, Honda Y, Tomeno W, Ogawa Y, Mawatari H,
changes in focal nodular hyperplasia, A finding to distinguish et al. Magnetic resonance imaging more accurately classifies
Non-alcoholic fatty liver disease 77

steatosis and fibrosis in patients with nonalcoholic fatty liver 77. Thoma C, Day CP, Trenell MI. Lifestyle interventions for the
disease than transient elastography. Gastroenterology 2016; treatment of non-alcoholic fatty liver disease in adults: a
150(3):626e37. e7. systematic review. J Hepatol 2012;56(1):255e66.
63. Patel V, Sanyal AJ. Drug-induced steatohepatitis. Clin Liver Dis 78. Sanyal AJ, Chalasani N, Kowdley KV, McCullough A, Diehl AM,
2013;17(4). 533evii. Bass NM, et al. Pioglitazone, Vitamin E, or placebo for nonal-
64. Naini BV, Lassman CR. Total parenteral nutrition therapy and coholic steatohepatitis. N Engl J Med 2010;362(18):1675e85.
liver injury: a histopathologic study with clinical correlation. 79. Harrison SA, Torgerson S, Hayashi P, Ward J, Schenker S.
Hum Pathol 2012;43(6):826e33. Vitamin E and Vitamin C treatment improves fibrosis in patients
65. Vallet-Pichard A, Mallet V, Pol S. Nonalcoholic fatty liver dis- with nonalcoholic steatohepatitis. Am J Gastroenterol 2003;
ease and HIV infection. Semin Liver Dis 2012;32(2):158e66. 98(11):2485e90.
66. Kleiner DE, Brunt EM. Nonalcoholic fatty liver disease: patho- 80. Mathurin P, Hollebecque A, Arnalsteen L, Buob D, Leteurtre E,
logic patterns and biopsy evaluation in clinical research. Semin Caiazzo R, et al. Prospective study of the long-term effects of
Liver Dis 2012;32(1):3e13. bariatric surgery on liver injury in patients without advanced
67. Zou Z-Y, Shen B, Fan J-G. Systematic review with meta- disease. Gastroenterology 2009;137(2):532e40.
analysis: epidemiology of nonalcoholic fatty liver disease in 81. Lutchman G, Modi A, Kleiner DE, Promrat K, Heller T, Ghany M,
patients with inflammatory bowel disease. Inflamm Bowel Dis et al. The effects of discontinuing pioglitazone in patients with
2019;25(11):1764e72. nonalcoholic steatohepatitis. Hepatology 2007;46(2):424e9.
68. Lapumnuaypol K, Kanjanahattakij N, Pisarcik D, 82. Contos MJ, Cales W, Sterling RK, Luketic VA, Shiffman ML,
Thongprayoon C, Wijarnpreecha K, Cheungpasitporn W. Effects Mills AS, et al. Development of nonalcoholic fatty liver disease
of inflammatory bowel disease treatment on the risk of after orthotopic liver transplantation for cryptogenic cirrhosis.
nonalcoholic fatty liver disease: a meta-analysis. Eur J Gas- Liver Transplant 2001;7(4):363e73.
troenterol Hepatol 2018;30(8):854e60. 83. Dumortier J, Giostra E, Belbouab S, Morard I, Guillaud O,
69. Sartini A, Gitto S, Bianchini M, Verga MC, DiGirolamo M, Spahr L, et al. Non-alcoholic fatty liver disease in liver trans-
Bertani A, et al. Non-alcoholic fatty liver disease phenotypes in plant recipients: another story of “seed and soil”. Am J Gas-
patients with inflammatory bowel disease. Cell Death Dis 2018; troenterol 2010;105(3):613e20.
9(2):87. 84. Dureja P, Mellinger J, Agni R, Chang F, Avey G, Lucey M, et al.
70. Bosch DE, Yeh MM. Primary sclerosing cholangitis is protective NAFLD recurrence in liver transplant recipients. Trans-
against nonalcoholic fatty liver disease in inflammatory bowel plantation 2011;91(6):684e9.
disease. Hum Pathol 2017;69:55e62. 85. Galvin Z, Rajakumar R, Chen E, Adeyi O, Selzner M, Grant D,
71. Reilly NR, Lebwohl B, Hultcrantz R, Green PHR, Ludvigsson JF. et al. Predictors of de novo nonalcoholic fatty liver disease
Increased risk of non-alcoholic fatty liver disease after diag- after liver transplantation and associated fibrosis. Liver
nosis of celiac disease. J Hepatol 2015;62(6):1405e11. Transplant 2019;25(1):56e67.
72. Tovoli F, Negrini G, Farı̀ R, Guidetti E, Faggiano C, Napoli L, 86. Eslam M, Valenti L, Romeo S. Genetics and epigenetics of
et al. Increased risk of nonalcoholic fatty liver disease in pa- NAFLD and NASH: clinical impact [Internet] J Hepatol 2018 Feb
tients with coeliac disease on a gluten-free diet: beyond 1;68(2):268e79. https://doi.org/10.1016/j.jhep.2017.09.003.
traditional metabolic factors. Aliment Pharmacol Ther 2018; Available from:.
48(5):538e46. 87. Masseroli M, Caballero T, O’Valle F, Moral RMGD, Pérez-
73. Kamal S, Aldossari KK, Ghoraba D, Abdelhakam SM, Kamal AH, Milena A, Moral RGD. Automatic quantification of liver fibrosis:
Bedewi M, et al. Clinicopathological and immunological char- design and validation of a new image analysis method: com-
acteristics and outcome of concomitant coeliac disease and parison with semi-quantitative indexes of fibrosis. J Hepatol
non-alcoholic fatty liver disease in adults: a large prospective 2000;32(3):453e64.
longitudinal study. BMJ open Gastroenterol 2018;5(1). 88. Munsterman ID, vanErp M, Weijers G, Bronkhorst C,
e000150ee000150. deKorte CL, Drenth JPH, et al. A novel automatic digital al-
74. Guindi M. Wilson disease. Semin Diagn Pathol 2019;36(6): gorithm that accurately quantifies steatosis in NAFLD on his-
415e22. topathological whole-slide images. Cytom Part B Clin Cytom
75. Lindor KD, Kowdley KV, Heathcote EJ, Harrison ME, 2019;96(6):521e8.
Jorgensen R, Angulo P, et al. Ursodeoxycholic acid for treat- 89. Liu F, Zhao JM, Rao HY, Yu WM, Zhang W, Theise ND, et al.
ment of nonalcoholic steatohepatitis: results of a randomized Second harmonic generation reveals subtle fibrosis differences
trial. Hepatology 2004;39(3):770e8. in adult and pediatric nonalcoholic fatty liver disease. Am J
76. Neuschwander-Tetri BA, Loomba R, Sanyal AJ, Lavine JE, Clin Pathol 2017;148(6):502e12.
VanNatta ML, Abdelmalek MF, et al. Farnesoid X nuclear re- 90. Yu Y, Wang J, Ng CW, Ma Y, Mo S, Fong ELS, et al. Deep learning
ceptor ligand obeticholic acid for non-cirrhotic, non-alcoholic enables automated scoring of liver fibrosis stages. Sci Rep
steatohepatitis (FLINT): a multicentre, randomised, placebo- 2018;8(1):1e10.
controlled trial. Lancet 2015;385(9972):956e65.