Liver: Original Paper
Dig Dis
DOI: 10.1159/000491428
The Pattern of Elevated Liver Function Tests in
Nonalcoholic Fatty Liver Disease Predicts Fibrosis
Stage and Metabolic-Associated Comorbidities
Dor Shirin a Noam Peleg a, b Orly Sneh-Arbib b Michal Cohen-Naftaly b
Marius Braun b, c Tzipora Shochat d Assaf Issachar a, b Amir Shlomai a–c
a Department
of Medicine D, Rabin Medical Center, Beilinson Hospital, Petah-Tikva, Israel; b The Liver Institute, Rabin
Medical Center, Beilinson hospital, Petah-Tikva, Israel; c The Sackler Faculty of Medicine, Tel-Aviv University, Tel-Aviv,
Israel; d Biostatistics Core, Rabin Medical Center, Petah-Tikva, Israel
Keywords
Cholestasis · Metabolic Syndrome · Liver steatosis
Abstract
Background: Patients with nonalcoholic fatty liver disease
(NAFLD) and with abnormal liver function tests (LFTs) most
commonly present with elevated hepatocellular enzymes (H
pattern), but a subset of patients is found to have elevated
cholestatic enzymes (C pattern) or a mixed (M) pattern. Aims
and Methods: To determine whether the epidemiologic
background and comorbidities, as well as the degree of liver
fibrosis, differ between NAFLD patients with different pat-
terns of elevated LFTs by retrospectively analyzing data of
106 patients with a biopsy-proven diagnosis of NAFLD. The
pattern of elevated LFTs was determined by adopting the
“R-Ratio” formula commonly used for drug-induced liver in-
jury. Results: Advanced fibrosis (F >2) was found in 15 out of
48 (31.3%) patients with a C pattern of elevated LFTs as com-
pared to 2 out of 44 (4.5%) in M patients and 2 out of 11
(18.2%) in H patients (p = 0.004). Group C patients are older
and also had a higher prevalence of diabetes, a higher mean
hemoglobin A1c, and a higher prevalence of hypertension,
as well as a trend for a higher prevalence of hypertriglyceri-
demia. Conclusions: Using a simple formula incorporating
© 2018 S. Karger AG, Basel
E-Mail karger@karger.com
www.karger.com/ddi
Received: February 21, 2018
Accepted: June 21, 2018
Published online: July 17, 2018
routine LFTs can help to categorize NAFLD patients as low or
high risk for advanced fibrosis stage and metabolic-associat-
ed comorbidities. © 2018 S. Karger AG, Basel
Introduction
Nonalcoholic fatty liver disease (NAFLD) is becoming
the most common liver disease worldwide with an esti-
mated prevalence of 20–30% among the general popula-
tion in many western countries [1, 2]. NAFLD is associ-
ated with the metabolic syndrome and particularly with
hypertension (HTN), insulin resistance, and type 2 dia-
betes mellitus [3]. About 10–20% of patients with NAFLD
have non-alcoholic steatohepatitis (NASH), the more ag-
gressive form of liver steatosis, histologically defined as
the presence of steatosis, inflammation, and hepatocyte
ballooning [4, 5].
Screening for NAFLD is recommended in asymptom-
atic patients with elevated liver functions tests (LFTs) since
this is the most common etiology in those patients [6]. Pa-
tients with NAFLD presenting with elevated LFTs typical-
ly present with elevated aminotransferase levels, with a
predominance of alanine transaminase (ALT) that rarely
128.111.121.42 - 7/18/2018 5:20:23 PM
Univ. of California Santa Barbara
Amir Shlomai, MD, PhD
Head, Department of Medicine D and the Laboratory of Liver Research
The Institute of Liver Disease and Felsenstein Medical Research Center
Rabin Medical Center, Beilinson Hospital, IL–4941492 Petah-Tikva (Israel)
Downloaded by:
E-Mail shlomaiamir @ gmail.com
exceed levels of 300 IU/L. Alkaline phosphatase (ALP) and
gamma glutamyltranspeptidase (GGT), on the other hand,
are less frequently elevated in patients with NAFLD [7].
Previous studies have concluded that elevation in cho-
lestatic enzymes among NAFLD patients is a predictor
for a histologically more advanced disease [8]. Another
study has found that NAFLD patients presented with iso-
lated ALP elevation were more likely to be older and of
female gender as compared to patients with a typical el-
evation in aminotransferase levels [9]. Yet, former studies
did not demonstrate significant differences between the 2
groups regarding the prevalence of metabolic syndrome-
related diseases, such as type 2 diabetes, HTN, or hyper-
triglyceridemia (Hyper TG), which often determine the
long-term prognosis of those patients.
Therefore, whether differences in the pattern of elevat-
ed LFTs are associated with co-morbidities that can neg-
atively affect the natural history of the disease in those
patients has not been thoroughly studied.
Based on our clinical practice and on previous studies,
we hypothesized that NAFLD patients with a more “cho-
lestatic pattern” of elevated liver enzymes at presentation
are more often inflicted with comorbidities related to the
metabolic syndrome and possibly merit more intensive
follow-up and screening programs. This study aims to
further investigate this hypothesis by retrospectively ana-
lyzing epidemiological and clinical data of biopsy-proven
NAFLD patients.
Materials and Methods
Patients and Study Design
This retrospective study was approved by the local institution-
al review board according to the local regulations. All documented
liver biopsy results performed from August 2005 to December
2012 as well as patients’ demographic, clinical, and laboratory data
obtained from their electronic records were reviewed. After an ex-
tensive review of the patients’ medical records, only those with
unequivocal diagnosis of NAFLD were included in the final analy-
sis.
Patients with evidence of viral hepatitis, Wilson’s disease, ma-
lignancy (including liver metastases), primary biliary cholangitis,
autoimmune hepatitis, or primary sclerosing cholangitis were ex-
cluded. Additionally, biopsies from liver transplanted patients
were excluded. In cases in which laboratory data were not available
at the time of liver biopsy, previously obtained data, up to 3 months
prior to liver biopsy (6 months was allowed for hemoglobin A1c)
were used. Components of the metabolic syndrome were recorded
including central obesity (waist circumference > 102 cm for men
and > 88 cm for women), diabetes (previously diagnosed type 2
diabetes or hemoglobin A1c >6.5%), Hyper TG (previously diag-
nosed or the presence of triglycerides > 150 mg/dL), and HTN
(previously diagnosed HTN). Body mass index (BMI) was calcu-
2 Dig Dis
DOI: 10.1159/000491428
lated using the formula: weight (in kilograms)/height (in meters)2.
The grade and stage of liver disease severity were assessed accord-
ing to the scoring system proposed by the NASH Clinical Research
Network [10]. Reviewing biopsy results, the presence of NAFLD
was determined when the pathologic report indicated the presence
of steatosis only, while biopsies that were reported as steatohepa-
titis or the presence of inflammation were diagnosed as NASH.
Fibrosis (0–4) was scored separately using the METAVIR score.
Patients were divided into 4 groups based on the pattern of el-
evated liver enzymes as follows: predominantly cholestatic pattern
(C pattern), predominantly hepatocellular pattern (H pattern),
mixed (M) pattern, and normal liver enzymes (N pattern). The pat-
tern of elevated LFTs was calculated using the “R-Ratio” formula:
R = (ALT/ALTULN)/(ALP/ALPULN) [11, 12]. C pattern group in-
cluded patients with an R-ratio of less than 2, whereas the H pattern
group included patients with an R-ratio of more than 5. The M pat-
tern group consisted of patients with an R-ratio between 2 and 5.
The upper limit of normal for ALT and ALP levels is 45 and 120
IU/L, respectively, as set by the local clinical laboratory in which
samples were analyzed. Demographic, comorbidities, biochemical,
and histological data were then analyzed to determine whether sig-
nificant differences existed between the groups. Following the main
analysis, a second, exploratory analysis was carried out, using upper
limits of normal ALT levels of 31 IU/L for men and 22 IU/L for
women, as suggested by the latest ACG guidelines [12].
Statistical Analysis
Statistical analysis was performed using the SAS 9.4 software.
The following variables were included in the univariate analysis:
demographics (age, gender), BMI, platelet count, albumin, ALT,
AST (aspartate transaminase), ALP, GGT, blood total cholesterol,
blood triglycerides level, hemoglobin A1c. All continuous variables
were analyzed after logarithmic transformation for normality of
distribution. Categorical variables were compared by X2 or the
Fisher exact tests, whereas continuous variables were compared
with the Anova test. Correlation was evaluated by the Spearman
correlation coefficient. A 2-sided p value of less than 0.05 was con-
sidered statistically significant. Logistic regression was made to
identify independent factors associated with the study endpoint.
Variables with missing values in more than 20% of the patients were
not included in the regression analysis. To assess the potential of a
bias, we added age and gender to the logistic regression analysis.
Results
Patients’ Characteristics
A total of 166 patients with biopsy-proven NAFLD
were screened. Fifty-three patients were excluded, 20 of
them due to insufficient clinical and/or laboratory data
and the rest because of comorbid conditions, such as pri-
or liver transplantation, chronic infection with HBV or
HCV, and prior treatment with hepatotoxic drugs.
The study population was divided into 4 groups ac-
cording to the pattern of LFTs: the group with predomi-
nantly elevated cholestatic enzymes (Group C, R ratio less
than 2) consisted of 49 (43.4%) patients, the group with
128.111.121.42 - 7/18/2018 5:20:23 PM
Univ. of California Santa Barbara
Shirin et al.
Downloaded by:

166 patients with biopsy-proven NAFLD screened
53 patients excluded
(insufficient data,
comorbid conditions)
113 patients
7 patients excluded
(normal LFTs)
106 patients
Categorize pattern of
elevated LFTs
R <2 2< R <5 R >5
Cholestatic (C) Mixed (M) Hepatocellular (H)
a
Fig. 1. a A diagram summarizing the study’s outline as described in detail in the text. b The distribution of the
study population (106 patients) according to the pattern of elevated LFTs. The 7 patients with normal LFTs were
excluded from analysis. NAFLD, nonalcoholic fatty liver disease; LFT, liver function test.
predominantly elevated hepatocellular enzymes (group
H, R ratio greater than 5) consisted of 11 (9.7%) patients,
the group with a mixed pattern of elevated LFTs (group
M, R ratio between 2 and 5) consisted of 46 (40.7%) pa-
tients, and the group with those whose both cholestatic
and hepatocellular enzymes are within the normal range
(group N) consisted of 7 (6.2%) patients. Due to its small
size, group N was excluded from final analysis (Fig. 1).
Of the 106 patients included in the final analysis, 46.2%
were females. The patients’ mean age was 49.4 years and
their mean BMI was 30.7, although only 45 (42.5%) pa-
tients had their BMI calculated at the time of presenta-
tion. Overall, 26.4% of patients had diabetes, 34.9% had
HTN, and 52.8% had Hyper TG (defined as triglycerides
>150 mg/dL) at the time of presentation.
Nineteen patients (17.9%) had NASH according to liv-
er biopsy, similar to the proportion of NASH among pa-
tients with NAFLD observed in other studies [13]. As ex-
pected, a higher proportion of patients with NASH had
diabetes as compared to patients with simple steatosis.
NASH patients also tended to be older, with higher BMI
and a higher proportion of HTN as compared to patients
with simple steatosis, although the difference in those pa-
rameters did not reach statistical significance (Table 1).
These observations are consistent with other studies [14].
As expected, patients with NASH had lower albumin and
thrombocytes levels and the percentage of patients with
LFTs Predict CoMorbidities in NAFLD
Color version available online
%) )
6.2 l (M
11 (9ellular (H
7 ( rma
.7%)
No
toc
Hepa
Mixed (M)
46 (40.7%)
Cholestatic (C)
49 (43.4%)
b
significant fibrosis (F >2) was higher as compared to pa-
tients with simple steatosis. Interestingly, the prevalence
of hyper TG was lower among NASH patients as was the
mean level of blood triglycerides, probably reflecting the
more severe liver disease in this group of patients.
Components of the Metabolic Syndrome in Relation to
the Pattern of LFTs
Basic epidemiological and laboratory characteristics of
the patients, according to their pattern of elevated LFTs,
are presented in Table 2. Patients in group C were signifi-
cantly older (mean age of 55.8 years as compared to 44.9
and 43.7 years in groups H and M respectively; p < 0.0001)
with a higher proportion of female patients (55.1% in
group C as compared to 36.3% in group H and 39.1% in
group M). Of note, patients in group C had significantly
lower serum albumin levels, as well as a trend for lower
platelet counts and higher INR level (Table 2).
Figure 2 summarizes the prevalence of major comor-
bidities, associated with NAFLD and the metabolic syn-
drome, among the 3 groups of patients. Among group C
patients, 31.3% had advanced fibrosis (F >2) in liver bi-
opsy compared to only 18.2 and 4.5% in groups H and M
respectively (p = 0.004; Fig. 2). This difference remained
significant after adjustment to age and gender.
Significant differences were also noted in the preva-
lence of diabetes between the 3 groups with elevated
128.111.121.42 - 7/18/2018 5:20:23 PM
Univ. of California Santa Barbara
Dig Dis 3
DOI: 10.1159/000491428
Downloaded by:
Table 1. Basic demographic and laboratory parameters of the study population according to their classification as steatosis or NASH

NASH (n = 19) Steatosis (n = 87) p value

Age, years, mean ± SD 52.8±10.4 48.7±13.7 ns

Gender, female, % (n) 52.6 (10) 44.8 (39) ns
BMI, mean ± SD 32.3±8.3 30.1±4.2 ns
DM, % (n) 52.6 (10) 20.7 (18) 0.008
HA1c, mean ± SD 7.57±2.5 6.52±1.2 ns
HTN, % (n) 42.1 (8) 34.1 (29)* ns
Hyper TG, % (n) 47.3 (9) 54 (47) ns
Triglycerides, mg/dL, mean ± SD 162.3±100.4 185.3±110.4 ns
GGT, U/L, mean ± SD 256.9±359 209.2±423.3 ns
ALP, U/L, mean ± SD 136.5±145.8 105.1±50.4 ns
AST, U/L, mean ± SD 74.2±52.4 60.3±47.4 ns
ALT, U/L, mean ± SD 88.9±59.9 82.2±51.7 ns
Albumin, g/dL, mean ± SD 3.92±0.8 4.44±0.3 0.018
Total bilirubin, mg/dL, mean ± SD 0.97±0.6 0.76±0.4 ns
Thrombocytes, K/micl, mean ± SD 176,895±78,947 240,758±71,003 <0.001
INR, mean ± SD 1.10±0.2 1.05±0.1 ns
Fibrosis score >2, % (n) 33.3 (6)** 15.3 (13)* 0.09

* Data was available for 85 patients, ** data was available for 18 patients.
BMI, body mass index; DM, diabetes mellitus; HTN, hypertension; Hyper TG, hypertriglyceridemia; GGT, gamma glutamyltranspep-
tidase; ALT, alanine transaminase; NASH, nonalcoholic steatohepatitis; ALP, alkaline phosphatase; INR, international normalized ratio;
ns, non-significant.

Table 2. Basic demographic and laboratory parameters of the study population according to the pattern of elevated LFTs

Hepatocellular (n = 11) Cholestatic (n = 49) Mixed (n = 46) p value

Age, years, mean ± SD 44.9±9.4 55.8±10.7 43.7±13.5 <0.0001

Gender, female, % (n) 36.3 (4) 55.1 (27) 39.1 (18) ns
HA1c, mean ± SD 6.03±0.7 6.85±1.5 6.82±2 ns
Triglycerides, mg/dL, mean ± SD 130.4±42.1 181.6±104.7 195.2±121.6 ns
ALP, U/L , mean ± SD 68.5±23.4 138.3±99 91.5±38.3 0.001
GGT, U/L, mean ± SD 82.9±52.9 256.6±306.5 208.5±544.6 ns
AST, U/L, mean ± SD 107.7±75.4 49.7±27.7 66±51.8 0.0009
ALT, U/L, mean ± SD 171.4±60.4 50.7±26.7 97.3±41.7 <0.0001
Albumin, g/dL, mean ± SD 4.63±0.2 4.14±0.6 4.5±0.3 <0.0004
Total bilirubin, mg/dL, mean ± SD 1.15±0.6 0.79±0.5 0.71±0.4 0.0182
Thrombocytes, K/micl, mean ± SD 241,909±55,542 216,347±90,379 240,109±61,640 ns
INR, mean ± SD 1.04±0.1 1.10±0.2 1.02±0.1 0.061

LFTs, liver function tests; GGT, gamma glutamyltranspeptidase; ALT, alanine transaminase; ALP, alkaline phosphatase.

LFTs; among group C, 40.8% of patients had diabetes as
compared to only 9.1% in group H and 15.2% in group M
(p = 0.007). Similar results were noted regarding the prev-
alence of HTN; 50% of patients from group C had HTN
as compared to only 18.2 and 24.4% among H and M
groups respectively (p = 0.016). When adjusted to age and
gender, however, these differences were not statistically
significant. A similar trend that has not reached statistical
significance was noted regarding the prevalence of hyper
TG; 61.2% of patients from group C had hyper TG as
compared to only 36.4% in group H and 47.8% in the M
group.
Subgroup Analyses of Simple Steatosis versus NASH
Patients According to their Pattern of LFTs
We next analyzed only patients with simple steatosis
based on the pattern of their elevated LFTs (Table 3).
Among patients with simple steatosis, the prevalence of
128.111.121.42 - 7/18/2018 5:20:23 PM
Univ. of California Santa Barbara

4 Dig Dis Shirin et al.

DOI: 10.1159/000491428
Downloaded by:
 Color version available online
80 Cholestatic Hepatocellular Mixed

**

Percentage of patients
60 *
***

40

Fig. 2. The prevalence of major comorbidi- 20

ties associated with the metabolic syn-
drome as a function of the patients’ pattern
of elevated LFTs. * p = 0.007, ** p = 0.016, 0
*** p = 0.004. DM, diabetes mellitus; HTN, DM HTN Hyper TG Fibrosis >2
hypertension; Hyper TG, hypertriglyceri- Comorbidity
demia.

Table 3. Basic demographic data and co-morbidities among patients with simple steatosis according to the pat-
tern of elevated LFTs

Hepatocellular (n = 9) Cholestatic (n = 40) Mixed (n = 38) p value

Gender, female, % (n) 22.2 (2) 57.5 (23) 36.8 (14) 0.066
Age, years, mean ± SD 55.1±11.1 44.7±10.2 42.9±14.2 0.0001
DM, % (n) 0 (0) 35 (14) 10.5 (4) 0.008
HTN, % (n) 22.2 (2) 48.7 (19)* 21.6 (8)** 0.03
Hyper TG, % (n) 44.4 (4) 60 (24) 50 (19) ns

* Data available for 39 patients, ** data available for 37 patients.

LFTs, liver function tests; DM, diabetes mellitus; HTN, hypertension; Hyper TG, hypertriglyceridemia.

diabetes as well as HTN was significantly higher in group
C as compared to groups M and H (35, 10.5 and 0% in
groups C, M and H, respectively for DM [p = 0.008] and
48.7, 21.6 and 22.2% in groups C, M and H, respectively,
for HTN [p = 0.03]). A trend for a higher prevalence of
hyper TG was noted in group C as well (60% vs. 50% and
44.4% in groups M and H respectively) (Table 3). Because
of its small size, it was impossible to show statistical sig-
nificance in NASH patients regarding those parameters,
although a similar trend was documented in NASH pa-
tients, as well.
An Exploratory Data Analysis Using Lower Cutoffs for
Abnormal LFTs
In an exploratory data analysis, using lower values
for upper limit of normal ALT (ALT levels of 31 IU/L
for men and 22 IU/L for women) [12], the relative siz-
es of the groups changed considerably. Group H con-
sisted of 28 (25.9%) patients, group C consisted of 24
(22.2%) patients, and group M consisted of 56 (51.9%)
patients (online suppl. Table 1, see www.karger.com/
doi/10.1159/000491428). These changes resulted from
the increment in the average R-ratio, which was 2.4
in the primary analysis and 4.08 in the exploratory anal-
ysis.
Similar to the primary analysis, more patients in group
C had diabetes (54.2, 25 and 7.1% respectively; p = 0.001)
and HTN (52.2, 38.2, and 14.3%, p = 0.014) as compared
to patients in groups M and H. When adjusted to age and
gender, the difference in the prevalence of diabetes re-
mained statistically significant, while the difference in
HTN did not. Dyslipidemia and severe fibrosis (F > 2)
were more common in patients with elevated cholestatic
enzymes (groups C and M) as compared to patients in
group H. The prevalence of severe fibrosis was 19, 25, and
13.4% in groups C, M, and H, respectively, while the prev-
alence of dyslipidemia was 54.2, 58.9, and 38.9% in groups
C, M, and H respectively. The difference between the
groups regarding the latter 2 parameters did not reach
statistical significance.
128.111.121.42 - 7/18/2018 5:20:23 PM
Univ. of California Santa Barbara

LFTs Predict CoMorbidities in NAFLD Dig Dis 5

DOI: 10.1159/000491428
Downloaded by:
 Discussion
In this study, we investigated the prevalence of ad-
vanced fibrosis and of comorbidities associated with the
metabolic syndrome among NAFLD patients with dis-
tinct patterns of elevated LFTs. Our results strongly sug-
gest that patients with a dominant elevation of choles-
tatic liver enzymes (C group) are more likely to have ad-
vanced fibrosis. This correlation remained significant
after adjusting for age and gender. These patients also
have a higher prevalence of metabolic-associated comor-
bidities as compared to NAFLD patients with either ele-
vated hepatocellular enzymes or a mixed pattern of ele-
vated LFTs. However, after adjustment to age and gender,
only age remained significantly associated with different
patterns of liver enzymes. Therefore, this study suggests
that patients with NAFLD accompanied by elevated cho-
lestatic liver enzymes represent a higher risk population
that is older and represent with more advanced liver dis-
ease. These patients are more likely to suffer from various
components of the metabolic syndrome, although this
could very well result from the older average age of this
group of patients.
NAFLD is tightly associated with the metabolic syn-
drome [7, 15]. In accordance with this, liver fat content,
as measured by H-MRS, is much higher in patients with
the metabolic syndrome as compared to patients with-
out it [16, 17]. NAFLD is not universally accompanied
by elevated LFTs, but it is now recognized as one of the
most common etiologies for this laboratory finding.
Moreover, the finding of elevated liver enzymes by itself,
among patients with or without established NAFLD, has
been repeatedly shown to correlate with the presence of
various components of the metabolic syndrome. For ex-
ample, a retrospective study on 157,308 healthy indi-
viduals has demonstrated a strong association between
higher ALT level and the development of DM in the
forthcoming years [18]. Another study on 1,309 healthy,
nondiabetic individuals has shown that increased GGT
and ALT were biomarkers of both systemic and hepatic
insulin resistance [19]. Likewise, a large meta-analysis
has shown that hepatic fat content, as well as elevated
ALT or GGT (with GGT more than ALT) are strong
predictors for diabetes [20] and that elevated GGT is
strongly associated with cardiovascular risk, indepen-
dent of alcohol intake [21, 22]. Moreover, a recent trial
showed that a correlation exists between liver enzymes
(AST, ALT, and ALP) and the degree of coronary steno-
sis diameter among patients with acute myocardial in-
farction [23]. Yet, it is still not clear whether elevated
6 Dig Dis
DOI: 10.1159/000491428
LFTs and especially their pattern (i.e., cholestatic versus
hepatocellular) reflects the severity of liver disease as
well as the risk for complications associated with the
metabolic syndrome in patients with established
NAFLD. Our study addresses this question by demon-
strating that NAFLD patients with a cholestatic pattern
of liver enzymes are more prone to diabetes, HTN and
hyper TG as compared to patients with either hepatocel-
lular or a mixed pattern of elevated LFTs. Moreover, the
observation that patients with a mixed pattern of elevat-
ed LFTs tend to have an intermediate risk for metabolic-
associated comorbidities supports the link between the
cholestatic pattern of elevated LFTs and comorbidities
associated with the metabolic syndrome. Yet, since the
correlation between the pattern of liver enzymes and
metabolic comorbidities was not significant after adjust-
ing to gender and age, we cannot rule out that a choles-
tatic pattern of elevated LFTs is a consequence of more
advanced age, which by itself is a risk factor for addi-
tional comorbidities.
The actual normal ALT value is an area of ongoing
controversy, and differences in the normal range are not-
ed between studies, in the result of different manufactur-
er’s recommendations, and in the distinct populations
tested in order to set the normal value. Numerous studies
have suggested that a lower threshold for normal ALT
could be better associated with liver or non-liver related
morbidities [16, 18, 24]. In our study, we set the normal
range according to manufacturer’s recommendations
and the individual R ratios were calculated accordingly.
In an effort to validate our results more thoroughly, we
also ran a second analysis after setting the upper normal
ALT values according to the updated recommendations
of the latest ACG guidelines [12]. Our results suggest that
even upon lowering the bar for abnormal ALT levels, a
similar trend for a higher rate of comorbidities and in-
creased severity of liver disease in patients with choles-
tatic liver enzymes still exists.
A larger cohort and probably a prospective study are
needed in order to establish a direct association between
components of the metabolic syndrome and the pattern
of elevated liver enzymes, although such an association is
suggested by the substantial distinction between the
groups regarding the presence of advanced fibrosis.
A previous study has shown that patients with NAFLD
who have biochemical cholestasis tend to have a more se-
vere histological liver impairment [8]. However, although
our study showed a clear correlation between a choles-
tatic pattern of elevated LFTs and a more advanced liver
fibrosis, it failed to show a correlation with the existence
128.111.121.42 - 7/18/2018 5:20:23 PM
Univ. of California Santa Barbara
Shirin et al.
Downloaded by:
of NASH per liver biopsy. Indeed, patients with NASH in
our cohort tended to have higher levels of both choles-
tatic as well as hepatocellular liver enzymes as compared
to NAFLD patients with simple steatosis, but this differ-
ence has not reached a statistical significance for either
pattern of elevated LFTs (Table 1). We speculate that this
could be a result of a relatively small sample size and spe-
cifically the small number of NASH patients in our study.
The mechanism by which a predominantly elevated
cholestatic liver enzymes is linked to a more progressive
course of NAFLD, as suggested in our study, is still to be
determined. Given the emerging role of bile acids as sig-
naling molecules that act at both hepatic and extrahepat-
ic tissues to regulate lipid and carbohydrate metabolic
pathways [25], one can speculate that liver cholestasis has
a direct role in worsening the hepatic and extra-hepatic
metabolic dysregulation in patients with NAFLD. In this
context, it will be interesting to investigate whether FXR
agonists, such as obeticholic acid, currently evaluated for
its therapeutic efficiency in NAFLD [26, 27], may have a
better benefit in a subset of NAFLD patients whose cho-
lestatic, rather than hepatocellular enzymes, are elevated.
Our study has several limitations that result from its
retrospective nature and from its limited sample size. The
first limitation is that while the vast majority of patients
with NAFLD present with elevated hepatocellular en-
zymes [28, 29], only 10% of patients in our first analysis
and 22.2% in the exploratory analysis presented with pure
elevated hepatocellular enzymes. This phenomenon is
probably derived from a selection bias introduced by the
more liberal biopsies performed in patients with suspect-
ed NAFLD but in whom cholestasis, rather than elevated
hepatocellular enzymes, predominates. In accordance
with this, a very low number of patients in our study
(6.7%) presented with enzymes in the normal range, as
opposed to 15–30% of patients reported in other studies.
In a recent study examining 106 patients with steatosis
(60 of whom had NAFLD), only 43% had elevated ALT
[30]. This is most probably the consequence of the afore-
mentioned selection bias, as patients with normal LFTs
seldom represent a diagnostic dilemma and therefore
much less frequently undergo liver biopsies.
Second, the study is based on patients’ data from only
one center specialized in liver diseases. Nevertheless, this
center is the largest institute in the country, treating a
large population of patients with diverse demographic,
socioeconomic, and clinical features and therefore large-
ly represents a heterogenic population. Third, due to a
selection bias, patients who have a greater chance for un-
dergoing liver biopsy are over represented, raising a con-
LFTs Predict CoMorbidities in NAFLD
cern that the study population might not accurately rep-
resent the overall population of NAFLD patients. Finally,
the R-ratio that was used to differentiate the subgroups
according to liver enzymes pattern was originally in DILI
patients but not in NAFLD patients. However, we believe
that the advantage of using a standardized parameter,
which enables objective comparison between variable
laboratory upper limit values outweighs this limitation.
In summary, our findings have implications for the
diagnostic evaluation, the follow-up and the treatment of
NAFLD patients. We suggest that although elevation in
cholestatic LFTs is not the rule in patients with NAFLD,
its presence may represent a variant of NAFLD that is
probably more common in older individuals and that is
implicated in more severe liver disease and metabolic-
related comorbidities. However, we cannot rule out that
the increased prevalence of metabolic-associated comor-
bidities is related to the more advanced age of this group
of patients, rather than to the pattern of elevated LFTs.
Further studies should investigate whether the pattern of
elevated LFTs changes with aging and is a marker for ac-
celeration of liver disease and the appearance of comor-
bidities. In addition, the question whether NAFLD pa-
tients with predominantly elevated cholestatic LFTs also
have a worse prognosis is still open and should be ad-
dressed.
Ethical Approval
All procedures performed in studies involving human partici-
pants were in accordance with the ethical standards of the institu-
tional and/or national research committee and with the 1964 Hel-
sinki declaration and its later amendments or comparable ethical
standards.
Disclosure Statement
The authors declare that they have no conflicts of interest to
disclose.
Author Contribution
D.S. collected and analyzed data and participated in writing the
manuscript. N.P. participated in collecting and analyzing data.
O.S.-A. participated in collecting and analyzing data. M.C.-N. par-
ticipated in collecting and analyzing data. M.B. participated in col-
lecting and analyzing data. T.S. analyzed the data and performed
the statistical analysis for this study. A.I. participated in collecting
and analyzing data. A.S. devised the idea for the project, directed
all data collection and analysis, and wrote the paper.
128.111.121.42 - 7/18/2018 5:20:23 PM
Univ. of California Santa Barbara
Dig Dis 7
DOI: 10.1159/000491428
Downloaded by:
 References
  1 Vernon G, Baranova A, Younossi ZM: Sys-
tematic review: the epidemiology and natural
history of non-alcoholic fatty liver disease and
non-alcoholic steatohepatitis in adults. Ali-
ment Pharmacol Ther 2011;34:274–285.
  2 Younossi ZM, Stepanova M, Afendy M, Fang
Y, Younossi Y, Mir H, et al: Changes in the
prevalence of the most common causes of
chronic liver diseases in the United States
from 1988 to 2008. Clin Gastroenterol Hepato
2011;9:524–530.e1.
  3 Alberti KG, Zimmet P, Shaw J: Metabolic syn-
drome – a new world-wide definition. A con-
sensus statement from the international dia-
betes federation. Diabet Med 2006; 23: 469–
480.
  4 Farrell GC, Larter CZ: Nonalcoholic fatty liv-
er disease: From steatosis to cirrhosis. Hepa-
tology 2006;43(2suppl):S99–S112.
  5 Hashimoto E, Taniai M, Tokushige K: Char-
acteristics and diagnosis of NAFLD/NASH. J
Gastroenterol Hepatol 2013;28:64–70.
 6 EASL-EASD-EASO clinical practice guide-
lines for the management of non-alcoholic
fatty liver disease. J Hepatol 2016; 64: 1388–
1402.
  7 Paschos P, Paletas K: Non alcoholic fatty liver
disease and metabolic syndrome. Hippokratia
2009;13:9–19.
  8 Sorrentino P, Tarantino G, Perrella A, Mi-
cheli P, Perrella O, Conca P: A clinical-mor-
phological study on cholestatic presentation
of nonalcoholic fatty liver disease. Dig Dis Sci
2005;50:1130–1135.
  9 Pantsari MW, Harrison SA: Nonalcoholic fat-
ty liver disease presenting with an isolated el-
evated alkaline phosphatase. J Clin Gastroen-
terol 2006;40:633–635.
10 Kleiner DE, Brunt EM, Van Natta M, Behling
C, Contos MJ, Cummings OW, et al: Design
and validation of a histological scoring system
for nonalcoholic fatty liver disease. Hepatol-
ogy 2005;41:1313–1321.
11 DeLeve LD, Kaplowitz N: Mechanisms of
drug-induced liver disease. Gastroenterol
Clin North Am 1995;24:787–810.
12 Kwo PY, Cohen SM, Lim JK: ACG Clinical
Guideline: Evaluation of Abnormal Liver
8 Dig Dis
DOI: 10.1159/000491428
Chemistries. Am J Gastroenterol 2017; 112:
18–35.
13 Bellentani S, Scaglioni F, Marino M, Bedogni
G: Epidemiology of non-alcoholic fatty liver
disease. Dig Dis 2010;28:155–161.
14 Lonardo A, Ballestri S, Marchesini G, Angulo
P, Loria P: Nonalcoholic fatty liver disease: a
precursor of the metabolic syndrome. Dig
Liver Dis 2015;47:181–190.
15 Yki-Järvinen H: Non-alcoholic fatty liver dis-
ease as a cause and a consequence of meta-
bolic syndrome. Lancet Diabetes Endocrinol
2014;2:901–910.
16 Kotronen A, Juurinen L, Hakkarainen A,
Westerbacka J, Cornér A, Bergholm R, et al:
Liver fat is increased in type 2 diabetic pa-
tients and underestimated by serum alanine
aminotransferase compared with equally
obese nondiabetic subjects. Diabetes Care
2007;31:165–169.
17 Kotronen A, Westerbacka J, Bergholm R, Pi-
etilainen KH, Yki-Jarvinen H: Liver fat in the
metabolic syndrome. J Clin Endocrinol
Metab 2007;92:3490–3497.
18 Shlomai A, Kariv R, Leshno M, Beth-or
A,  Sheinberg B, Halpern Z: Large-scale
population analysis reveals an extremely
­ tes and nonalcoholic fatty liver disease. Gas-
low  threshold for "non-healthy" alanine
aminotransferase that predicts diabetes mel-
litus. J Gastroenterol Hepatol 2010;25:1687–
1691.
19 Bonnet F, Ducluzeau PH, Gastaldelli A,
Laville M, Anderwald CH, Konrad T, et al:
Liver enzymes are associated with hepatic in-
sulin resistance, insulin secretion, and gluca-
gon concentration in healthy men and wom-
en. Diabetes 2011;60:1660–1667.
20 Fraser A, Harris R, Sattar N, Ebrahim S, Dav-
ey Smith G, Lawlor DA: Alanine aminotrans-
ferase, gamma-glutamyltransferase, and inci-
dent diabetes: the British women’s heart and
health study and meta-analysis. Diabetes Care
2009;32:741–750.
21 Fraser A, Harris R, Sattar N, Ebrahim S, Smith
GD, Lawlor DA: Gamma-glutamyltransferase
is associated with incident vascular events in-
dependently of alcohol intake: analysis of the
British women’s heart and health study and
meta-analysis. Arterioscler Thromb Vasc Biol
2007;27:2729–2735.
22 Wannamethee SG, Lennon L, Shaper AG: The
value of gamma-glutamyltransferase in car-
diovascular risk prediction in men without
diagnosed cardiovascular disease or diabetes.
Atherosclerosis 2008;201:168–175.
23 Baars T, Neumann U, Jinawy M, Hendricks S,
Sowa JP, Kälsch J, et al: In acute myocardial
infarction liver parameters are associated
with stenosis diameter. Medicine 2016;
95:e2807.
24 Kälsch J, Bechmann LP, Heider D, Best J,
Manka P, Kälsch H, et al: Normal liver en-
zymes are correlated with severity of metabol-
ic syndrome in a large population based co-
hort. Sci Rep 2015;5:13058.
25 Arab JP, Karpen SJ, Dawson PA, Arrese M,
Trauner M: Bile acids and nonalcoholic fatty
liver disease: Molecular insights and thera-
peutic perspectives. Hepatology 2017;65:350–
362.
26 Mudaliar S, Henry RR, Sanyal AJ, Morrow L,
Marschall HU, Kipnes M, et al: Efficacy and
safety of the farnesoid X receptor agonist
obeticholic acid in patients with type 2 diabe-
troenterology 2013;145:574–582.e571.
27 Neuschwander-Tetri BA, Loomba R, Sanyal
AJ, Lavine JE, Van Natta ML, Abdelmalek
MF, et al: Farnesoid X nuclear receptor ligand
obeticholic acid for non-cirrhotic, non-alco-
holic steatohepatitis (FLINT): a multicentre,
randomised, placebo-controlled trial. Lancet
2015;385:956–965.
28 Ekstedt M, Franzén LE, Mathiesen UL,
Thorelius L, Holmqvist M, Bodemar G, et al:
Long-term follow-up of patients with NAFLD
and elevated liver enzymes. Hepatology 2006;
44:865–873.
29 Sanyal AJ: AGA technical review on nonalco-
holic fatty liver disease. Gastroenterology
2002;123:1705–1725.
30 Kälsch J, Keskin H, Schütte A, Baars T, Baba
HA, Bechmann LP, et al: Patients with ultra-
sound diagnosis of hepatic steatosis are at
high metabolic risk. Z Gastroenterol 2016;54:
1312–1319.
128.111.121.42 - 7/18/2018 5:20:23 PM
Univ. of California Santa Barbara
Shirin et al.
Downloaded by: