Available via license: CC BY-NC
Content may be subject to copyright.
https://doi.org/10.1177/2042018817741852
https://doi.org/10.1177/2042018817741852
Ther Adv Endocrinol
Metab
1 –14
DOI: 10.1177/
2042018817741852
© The Author(s), 2017.
Reprints and permissions:
http://www.sagepub.co.uk/
journalsPermissions.nav
Therapeutic Advances in Endocrinology and Metabolism
journals.sagepub.com/home/tae 1
Introduction
Nonalcoholic fatty liver disease (NAFLD) is the
most common cause of elevated liver enzymes in
adults in the United States. The term NAFLD is
used to encompass a wide range of liver damage
from simple steatosis to nonalcoholic steatohepatitis
(NASH), advanced fibrosis, and cirrhosis.1,2
Although the pathogenesis of NAFLD and the role
of insulin resistance on its progression has not been
fully defined, the ‘multiple-hit’ hypothesis was
recently developed to explain the role of multiple
insults on the liver which may induce NAFLD.3
This hypothesis replaces the previously defined
‘two-hit’ hypothesis, where a sedentary lifestyle,
obesity, and insulin resistance lead to hepatic steato-
sis (first hit), which promotes liver inflammation
and cellular injury (second hit).4,5 The multiple-
hit hypothesis defines insulin resistance as a key
Treating nonalcoholic fatty liver disease
in patients with type 2 diabetes mellitus: a
review of efficacy and safety
Elizabeth P. Mills, K. Paige D. Brown, Jennifer D. Smith, Phillip W. Vang and Katie Trotta
Abstract
Objective: To review current literature for the efficacy and safety of treatment for nonalcoholic
fatty liver disease (NAFLD) in patients with type 2 diabetes mellitus (T2DM).
Data sources: A PubMed literature search from January 1990 to June 2017 was conducted
using the search terms nonalcoholic fatty liver disease, diabetes mellitus, type 2, therapy,
treatment, treat, therapeutics, nonalcoholic fatty liver, nonalcoholic hepatosteatosis, NASH,
NAFLD, metformin, and statin. Bibliographies of chosen articles were reviewed.
Study selection and data extraction: Relevant articles on metformin, thiazolidinediones (TZD),
glucagon-like peptide-1 receptor agonists (GLP-1 RA), and statins for the treatment of NAFLD
which included patients with T2DM were reviewed. A total of 23 relevant studies were found
and included randomized controlled, observational, and open-label designs, as well as three
meta-analyses.
Data synthesis: Metformin combined with weight loss provides a modest improvement in
steatosis and no improvement in fibrosis in patients with NAFLD and T2DM. TZDs showed
positive results on fibrosis and resolution of NASH but at least half of patients studied were
nonresponders. GLP-1 RAs also showed favorable results on reductions in transaminases
and steatosis and improvements in insulin sensitivity and weight loss but lack efficacy data
for resolution of NASH or improvement in fibrosis scores. Statins showed favorable results
on reductions in transaminases but mixed results for improvement in steatosis and fibrosis
scores.
Conclusion: All reviewed treatment options are safe for management of NAFLD in patients
with T2DM but long-term histological improvements are minimal. TZDs are efficacious for
resolution of NASH and improvements in fibrosis but long-term use is required to maintain
these results.
Keywords:
diabetes, glucagon-like peptide 1 receptor agonists, nonalcoholic fatty liver disease,
statin, steatohepatitis, thiazolidinediones, treatment, type 2
Received: 14 September 2017; accepted in revised form: 23 October 2017.
Correspondence to:
Elizabeth P. Mills
Campbell University
College of Pharmacy
& Health Sciences,
Pharmacy Practice, PO
Box 1090, Buies Creek, NC
27506, USA
mills@campbell.edu
Jennifer D. Smith
William Jennings Bryan
Dorn VA Medical Center,
Columbia, SC 29209, USA
Katie Trotta
K. Paige D. Brown
Phillip W. Vang
Campbell University
College of Pharmacy
& Health Sciences,
Pharmacy Practice, Buies
Creek, NC, USA
741852TAE0010.1177/2042018817741852Therapeutic Advances in Endocrinology and MetabolismEP Mills, K Trotta
review-article2017
Review
Therapeutic Advances in Endocrinology and Metabolism 00(0)
2 journals.sagepub.com/home/tae
contributor to the development of NAFLD because
of its impact on increases in de novo lipogenesis and
dysfunction in the release of free fatty acids (FFAs)
and triglycerides from the liver.3,5 These risk factors
are also associated with the development of type 2
diabetes (T2DM), explaining the high rate of these
diseases occurring concomitantly. Studies estimate
the prevalence of hepatic steatosis in patients with
T2DM to be 30–50%.6 The prognosis for patients
with concomitant NAFLD and T2DM is worsened
due to increased risk for life-threatening sequela
such as cardiovascular disease and hepatocellular
carcinoma, highlighting the need for improved
treatment options.
The American Association for the Study of Liver
Diseases, the American College of Gastroenterology,
and the American Gastroenterological Association
published joint practice guidelines in 2012 which
recommend lifestyle interventions (hypocaloric diet
and increased physical activity) and a body weight
reduction of 3–5% to achieve improvement of stea-
tosis; however, up to 10% weight loss is needed to
demonstrate improvements in necroinflammation.7
While there are no drugs approved for the treat-
ment of NASH by the US Food and Drug
Administration (FDA), guidelines recommend
vitamin E as a first-line treatment in individuals
without diabetes. The guidelines also recommend
pioglitazone, but warn most clinical studies were
conducted in patients without diabetes. At the time
of guideline publication, there was not enough evi-
dence to support a recommendation for the use of
metformin or statins as a treatment for NASH, but
the use of statins for dyslipidemia in patients with
NASH is encouraged as they appear safe.
Clinicians often question the safety of common
drug treatments for patients with T2DM and
NASH. In recent years, numerous trials have
been conducted utilizing insulin sensitizers and
statins to treat NASH, which included patients
with T2DM in the study design. The objective of
this literature review is to evaluate the safety and
efficacy of medications for the treatment of
NASH in patients with T2DM.
Methods
A review of published studies using PubMed was
conducted to identify reports pertaining to the
safety and efficacy of pharmacologic treatments of
NAFLD commonly used in patients with T2DM.
One author conducted the search and assessed eli-
gibility, and all authors contributed to the review
of data, drafting, and editing of the manuscript.
MeSH terms used in various combinations
included non-alcoholic fatty liver disease, diabetes
mellitus, type 2, therapy, treatment, treat, thera-
peutics, nonalcoholic fatty liver, nonalcoholic
hepatosteatosis, NASH, NAFLD, metformin,
and statin. PubMed search filters were applied for
published dates between 1 January 1990 and 30
June 2017, English language, adults (age ⩾19
years), clinical trial, meta-analysis, or observa-
tional studies. Other articles of interest were
obtained from bibliographies of included articles.
Results
A total of 397 abstracts were initially reviewed for
possible inclusion. Only 23 articles met inclusion
criteria based on relevancy to the study popula-
tion and outcomes relevant to safety and efficacy
of treatment.
Metformin
Metformin has several mechanisms by which it
helps to reduce blood glucose and improve insu-
lin sensitivity, including decreasing gluconeogen-
esis in the liver, increasing glucose uptake in the
periphery, and increasing fatty acid oxidation, all
leading to a decrease in cellular insulin produc-
tion.8,9 Metformin promotes weight loss, is inex-
pensive, and has long-term data showing safety
and tolerability, making it a viable option in the
treatment of NAFLD.
A meta-analysis completed by Li and colleagues
analyzed data from nine studies involving 417
participants on the use of metformin dosed at
0.5–3 g/day for NAFLD.4 The primary outcome
was histological response to therapy, including
steatosis, inflammation, hepatocellular balloon-
ing, and fibrosis. No significant difference was
seen in any of the variables between metformin
and diet and exercise alone; this was also true for
a subgroup analysis of patients with diabetes ver-
sus those without. A limitation of this meta-analy-
sis is that only five of the nine studies could be
assessed for the primary outcome due to provi-
sion of insufficient data in the individual studies.
A study completed by Nair and colleagues
included patients with NAFLD who took met-
formin (20 mg/kg/day in three divided doses) for
48 weeks and compared pre- and post-treatment
liver biopsies.8 Ten of 15 patients completed both
biopsies; of these only three saw a reduction in
steatosis at the study’s conclusion.
EP Mills, K Trotta et al.
journals.sagepub.com/home/tae 3
Although the data on improvements in histologic
outcomes with metformin in NAFLD have been
poor, improvements in metabolic markers have
been seen in two meta-analyses and several
smaller studies (Table 1). An important benefit of
metformin therapy is its contribution to weight
loss, possibly through its impact on insulin sensi-
tivity and the gastrointestinal adverse effects asso-
ciated with its use.9 Improvements in body mass
index (BMI) from baseline were seen in the meta-
analysis by Li and colleagues with a weight loss of
–0.82 kg/m2 (p < 0.04).4 Significant improve-
ments in the homeostatic model assessment for
insulin resistance (HOMA-IR) scores were seen
in the meta-analysis by Li and colleagues (p =
0.04)4 and Mazza and colleagues (p = 0.003)9,
indicating improvements in insulin sensitivity. A
study completed by Loomba and colleagues
found that patients with a lower baseline BMI
responded more significantly to metformin ther-
apy than those with a baseline BMI of at least 40
in terms of weight loss, reductions in HOMA-IR,
and histological improvements.5 The authors
concluded that there is a positive correlation
between weight loss and improvements in hepa-
tocellular injury and inflammation. Mild to mod-
erate increases in alanine aminotransferase (ALT)
is the most common laboratory finding in
NAFLD, and improvements in ALT and aspar-
tate aminotransferase (AST) have been seen with
metformin treatment in almost all patients in the
reviewed studies.4,8–10 These decreases do tend to
be more significant after the first three months of
treatment, at which time AST/ALT levels gener-
ally plateau.8
Another contributing factor to obesity and
NAFLD is leptin, a hormone produced by adi-
pose tissue to indicate fullness during mealtime as
well as to regulate the collection of lipids to the
adipose sites.3 Patients with obesity can experi-
ence leptin resistance, and high serum levels of
leptin were seen in patients with NAFLD.3,11
Increases in leptin can impact proinflammatory
responses and fibrogenesis as well as modify the
effects of insulin on hepatic fat metabolism and
increase insulin resistance.3,11 A small study (n =
34) looked at the effects of metformin in patients
with NAFLD in regards to decreases in serum
leptin in comparison to a lifestyle modification
intervention.11 Leptin was significantly reduced
in both groups from baseline to month 6 (p =
0.039 in the lifestyle group and p = 0.047 in the
metformin group), but there was no difference
between the groups. Leptin levels were reduced
in correlation with amount of weight loss in both
groups, further emphasizing the importance of a
focus on weight management in patients with
NAFLD.
Thiazolidinediones
It could be hypothesized that thiazolidinediones
(TZDs) are well suited for use in the treatment of
NASH due to their powerful insulin-sensitizing
properties. TZDs bind to peroxisome prolifera-
tor-activated receptor γ receptors and improve
insulin sensitivity in the liver, skeletal muscle, and
adipose tissue.12,13 In addition, they increase
plasma adiponectin levels14 and decrease proin-
flammatory cytokines,15 all of which are primary
processes involved in NASH. Several studies
examined the use of rosiglitazone and pioglita-
zone in patients with impaired glucose tolerance
or T2DM and biopsy proven NASH (Table 2).
Efficacy. Two trials16,17 examined rosiglitazone
use in patients with biopsy-confirmed NASH. In
one trial,16 rosiglitazone was studied in an open-
label design in 30 subjects, half of those with
either T2DM or impaired glucose tolerance, to
determine if rosiglitazone would improve insulin
sensitivity, improve hepatic steatosis, and reduce
serum liver aminotransferases. At the end of 48
weeks, serum levels of ALT and AST decreased
significantly from baseline. Significant improve-
ments in histologic markers, specifically steatosis
(p = 0.004) and ballooning (p = 0.003), were
observed. While significant changes in the charac-
teristic and pattern of fibrosis were seen, no sig-
nificant difference in the global fibrosis score was
observed (p = 0.583). Importantly, approximately
half of the participants were nonresponders. For
those who did respond to rosiglitazone, transami-
nases returned to baseline values 6 months after
stopping treatment. The second study, a random-
ized controlled trial completed by Ratziu and col-
leagues, compared rosiglitazone with placebo in
63 subjects, 20 with T2DM.17 At the end of 12
months, significantly more patients in the rosigli-
tazone group achieved over 30% reduction in ste-
atosis (47% versus 16%, p = 0.014) and
normalized serum ALT levels (38% versus 7%, p
= 0.005) compared with placebo. Additional clin-
ical improvements in HOMA-IR, fasting insulin
level, and adiponectin level were noted. Similar to
the first trial, there were no significant improve-
ments seen in fibrosis, and 50% of subjects were
nonresponders. Nonresponders had higher
γ-glutamyltransferase levels, higher instances of
Therapeutic Advances in Endocrinology and Metabolism 00(0)
4 journals.sagepub.com/home/tae
Table 1. Efficacy and safety trials of metformin for NAFLD in patients with T2DM.
Study and
trial
design
Treatment Comparator Number of
Participants
Duration
(months
unless
otherwise
stated)
ALT
mean
change
from
baseline
AST
mean
change
from
baseline
Imaging or
histologic
changes
Other significant
measurements
Li et al.4
Meta-
analysis
MET Diet/diet +
exercise/
various PBO
417
(68 with DM
or IGT)
6 or 12 –8.12 U/
liter, p =
0.03
–4.52 U/
liter, p =
0.04
No changes
seen in
steatosis,
inflammation,
ballooning, or
fibrosis
HOMA-IR
changes were
statistically
significant in
patients with
NAFLD but not
NASH
Loomba
et al.5
MET 2000
mg/day
26
(15 with DM)
48 weeks ↓ 7 U/
liter
↓ 4 U/
liter
Presence of
NASH
Pre: 26/26
Post: 18/26
NASH activity
index
Pre: 8.2 (1.5)
Post: 5.9 (2.2)
(p < 0.001)
Average weight ∆
–6 kg (range +1.3
to −18.9 kg)
Strong positive
correlation
between weight
loss and changes
in serum
aminotransferase
levels; higher
baseline BMI
↓response
HOMA-IR ∆ –3.4
(p < 0.04)
Nair et al.8
Open label
MET 20
mg/kg/day
(max 2 g)
15
(1 with DM)
1 year ↑ 6 IU/
liter
↓ 6 IU/
liter
20% of patients
showed
improvement
in degree of
steatosis at 1
year
BMI ∆ –1.7% (p <
0.05)
HOMA-IR ∆ –0.09
(p < 0.05)
Haukeland
et al.10
RCT
MET 2500
mg/day
PBO 48
(12 with DM)
6 MET: 22
U/liter,
p = 0.025
PBO: 15
U/liter,
p = 0.025
MET: 8 U/
liter,
p = 0.036
PBO: no
change
Treatment was
associated
with a slight
reduction of
liver steatosis
in both groups
Age, baseline
HOMA-IR, ∆ in
body weight:
independently
associated with
change in liver
steatosis
MET caused
weight loss (–4.3
± 4.3 kg) (p ⩽
0.001)
MET group had
a significant
change in leptin
levels (p ⩽ 0.001)
MET significantly
lowered LDL
levels, mean ∆
–27 mg/dl (p <
0.001); no change
in PBO group
Nar and
Gedik11
MET 1700
mg/day
plus
diet and
exercise
Diet and
exercise
34
(all with DM)
6 MET: 16
U/liter,
p = 0.015
Lifestyle:
7 U/liter,
p = 0.047
No
change
(either
group)
Liver
echogenicity
decreased
significantly in
both groups
MET group
decreased LDL ∆
–23 (p = 0.002)
MET group
increased HDL ∆
+4 (p = 0.035)
ALT, alanine aminotransferase; AST, aspartate aminotransferase, BMI, body mass index; DM, diabetes mellitus; HDL, high-density lipoprotein;
HOMA-IR, homeostatic model assessment insulin resistance; LDL, low-density lipoprotein; MET, metformin; NAFLD, nonalcoholic fatty liver
disease; NASH, nonalcoholic steatohepatits; PBO, placebo; RCT, randomized controlled trial.
EP Mills, K Trotta et al.
journals.sagepub.com/home/tae 5
Table 2. Efficacy and safety trials of thiazolidinediones for NAFLD in patients with T2DM.
Study and trial
design
Treatment Comparator Number of
participants
Duration
(months
unless
otherwise
stated)
ALT mean
change from
baseline
AST mean
change from
baseline
γ-GT mean
change from
baseline
Imaging or histologic
changes
Other significant measurements
Neuschwander-
Tetri et al.16
Open label
ROS 4 mg twice
daily
None 30
(8 with diabetes,
7 with IGT)
12 –54%, p < 0.001 –43.3%, p =
0.003
–62.5%, p <
0.001
Steatosis
14 improved, p = 0.004
1 worsened
Ballooning
11 improved, p = 0.003
Fibrosis
8 improved, p = 0.583
3 worsened
Mean Δ BMI
6.5% (range –5% to 18%), p < 0.001
Mean Δ HOMA-IR
–3.5, p < 0.001
Ratziu et al.17
RCT
ROS 4 mg/day
titrated to 8
mg/day after 1
month
PBO 63
(20 with DM)
12 ROS: ALT
normalized in 12
subjects (38%)
PBO: ALT
normalized in 2
subjects (7%)
p = 0.005
>30% improvement in
steatosis
ROS 47%
PBO 16%
p = 0.014
Ballooning
NS
Fibrosis
ROS NS
PBO NS
Inflammation
ROS NS
PBO NS
Mean Δ body weight kg (SD)
ROS: +1.5 (5.2)
PBO: –1.0 (3.5)
p = 0.03
Mean Δ HOMA-IR
ROS –1.4
PBO 0.61
p < 0.001
Mean Δ fasting insulin level (µUI/ml)
ROS –4.8 (9.2)
PBO 2.1 (14.8)
p < 0.002
Mean Δ adiponectin level (µg/mL)
ROS 1.91 (1.39)
PBO 0.94 (0.23)
p = 0.04
Torres et al.18
Randomized open
label
ROS 4 mg twice
daily
ROS 4 mg plus
MET 500 mg
twice daily
ROS 4 mg twice
daily plus LOS 50
mg/day
108
(18 with DM)
12 ROS –44%
ROS + MET
–47%
ROS + LOS
–52%
p < 0.001
overall
ROS –32%
ROS + MET
–38%
ROS + LOS
–36%
p < 0.001
overall
Steatosis improved
ROS 27%
ROS + MET 30%
ROS + LOS 26%
p < 0.001 overall
Hepatocellular inflammation
ROS 36%
ROS + MET 25%
ROS + LOS 21%
p < 0.001 overall
Fibrosis improved
ROS 50%
ROS + MET 50%
ROS + LOS 21%
p < 0.001 overall
Participants with diabetes had
significant improvements in NAS
compared with those without
diabetes, p = 0.046
Mostly due to steatosis, p = 0.006
Overall, no added benefit was seen
above ROS alone compared with
combination groups with respect to:
Steatosis, p = 0.905
Hepatocellular inflammation, p =
0.46
Fibrosis, p = 0.302
NAS, p = 0.671
Resolution of NASH (% of
participants)
ROS 46%
ROS + MET 36%
ROS + LOS 29%
(Continued)
Therapeutic Advances in Endocrinology and Metabolism 00(0)
6 journals.sagepub.com/home/tae
Study and trial
design
Treatment Comparator Number of
participants
Duration
(months
unless
otherwise
stated)
ALT mean
change from
baseline
AST mean
change from
baseline
γ-GT mean
change from
baseline
Imaging or histologic
changes
Other significant measurements
Omer et al.19
Open label,
randomized
MET 1700 mg/
day
ROS 4 mg/day
and MET 1700
mg/day + ROS 4
mg/day
64
(all with DM or
IGT)
12 MET –26%, NS
ROS –56%, p <
0.0001
MET+ROS
–31%, p = 0.017
p values pre–
post in-group
comparison
MET –28%, NS
ROS –25%, p =
0.005
MET+ROS
–30%, p = 0.01
MET –49%,
NS
ROS +30%,
NS
MET+ROS
–56%, p =
0.008
Only 35/64 subjects had
follow-up biopsies
NAFLD score improvement
(n)
MET (10): +0.7, p = 0.726
ROS (13): –2.6, p = 0.012
MET+ROS(12): –3.9, p =
0.026
Mean Δ BMI:
MET –3.2, p = 0.002
ROS –0.3, NS
MET+ROS –1.3, p = 0.006
Mean Δ fasting plasma insulin
MET –23%, NS
ROS –33%, p = 0.005
MET+ROS –33%, p ⩽ 0.005
Mean Δ HOMA-IR
MET –18%, NS
ROS –38%, p = 0.003
MET+ROS –28%, NS
Gastaldelli et al.20
RCT
PIO 45 mg/day
+ hypocaloric
diet
PBO +
hypocaloric diet
47 with IGT or
DM; 20 healthy
controls
6Necroinflammation improved
PIO 44%
PBO 12%
p ⩽ 0.001
Improvements in FFA metabolism
PIO versus PBO 20% difference,
p = 0.01
Adipo-IR decreased by ~47% in PIO
group compared with baseline
p = 0.03
Strong correlations were found in
improvement in Adipo-IR and the
following for the PIO group only:
Steatosis decreased ~50% (r =
0.29, p = 0.049)
Belfort et al.21
RCT
PIO 45 mg/
day plus
hypocaloric
diet
Placebo plus
hypocaloric diet
48; 10 healthy
controls
6 PIO –58%
PBO –34%
p < 0.001
PIO –40%
PBO –21%
p = 0.04
Steatosis improved
PIO 65%
PBO 38%
p = 0.003
Ballooning improved
PIO 54%
PBO 24%
p = 0.02
Fibrosis improved
PIO 46%
PBO 33%
p = 0.08
Mean Δ body fat % (body weight)
PIO 1.5% (+2.5±0.5 kg)
PBO –4% (–3.2 ± 0.5 kg)
p = 0.005
Mean Δ fasting plasma insulin
PIO –34%
PBO no change
p ⩽ 0.001
Mean Δ FFA levels
PIO –17%
PBO no change
p = 0.044
Table 2. (Continued)
EP Mills, K Trotta et al.
journals.sagepub.com/home/tae 7
Table 2. (Continued)
Study and trial
design
Treatment Comparator Number of
participants
Duration
(months
unless
otherwise
stated)
ALT mean
change from
baseline
AST mean
change from
baseline
γ-GT mean
change from
baseline
Imaging or histologic
changes
Other significant measurements
Cusi et al.22
RCT
PIO 45 mg/day
+ hypocaloric
diet
PBO +
hypocaloric diet
101
(all with DM or
IGT)
36
(18 blinded
18 open
label)
NAS improved by ⩾ 2 points
PIO 58%
PBO 17%
treatment difference of
41 percentage points (CI
23–59); p < 0.001
Resolution of NASH
PIO 51%
PBO 19%
treatment difference of 32
percentage points (CI 13–51); p <
0.001
Mean Δ fibrosis score
PIO –0.5
PBO 0
Treatment difference −0.5 (CI −0.9
to 0); p = 0.039
ALT, alanine aminotransferase; AST, aspartate aminotransferase, BMI, body mass index; CI, confidence interval; DM, diabetes mellitus; FFA, free fatty acid; γ-GT, γ glutamyl transferase;
HOMA-IR, homeostatic model assessment insulin resistance; IGT, impaired glucose tolerance; LOS, losartan; MET, metformin; NAS, NAFLD Activity Score; NASH, nonalcoholic
steatohepatits; NS, not significant; PBO, placebo; PIO, pioglitazone; RCT, randomized controlled trial; ROS, rosiglitazone; SD, standard deviation; IGT impaired glucose tolerance.
diabetes, lower adiponectin levels, and lower
amounts of steatosis. Four months after treat-
ment, serum transaminases returned to baseline
levels.
Combination treatments with rosiglitazone and
metformin have been investigated. Torres and
colleagues compared rosiglitazone with met-
formin or with the combinations rosiglitazone
plus metformin or rosiglitazone plus losartan and
reported significant within-group improvements
in steatosis, necroinflammation, ballooning and
fibrosis (p < 0.001 for all); there was no signifi-
cant between-group difference suggesting no
benefit to adding metformin or losartan to rosigli-
tazone in NASH treatment.18
When data from subjects with diabetes were ana-
lyzed separately, NAFLD Activity Score (NAS)
significantly improved in patients with diabetes
compared with those without, mostly due to
improvement in steatosis (p = 0.006). Omer and
colleagues compared metformin with rosiglita-
zone or the combination of rosiglitazone plus
metformin for the treatment of NASH in 64 sub-
jects with T2DM or impaired glucose tolerance.19
Significant within-group differences were
observed for reductions in serum ALT and AST
levels and NAS scores for the rosiglitazone and
rosiglitazone plus metformin groups but not for
the metformin group. HOMA-IR reduced signifi-
cantly in the rosiglitazone group (p < 0.05) only.
No significant change in fibrosis was noted in any
treatment group.
In a randomized placebo-controlled trial by
Gastaldelli and colleagues, histologic and meta-
bolic effects of pioglitazone were compared with
placebo for the treatment of NASH in patients
with T2DM or impaired glucose tolerance.20
Patients in both groups maintained a calorie-
restricted diet by reducing their intake by 500
kcal/day. Changes in glucose and lipid metabo-
lism, as well as adipose tissue insulin resistance
(Adipo-IR) were reported. At baseline, in com-
parison to a control group without NASH,
patients with NASH were found to have signifi-
cantly lower plasma adiponectin levels, two to
three times the concentration of plasma insulin
levels, and significantly higher plasma FFA con-
centrations. These metabolic differences are
indicative of systemic and adipose tissue insulin
resistance and were found to be true for obese as
well as lean subjects with NASH. At the end of six
months, pioglitazone significantly reduced FFA
Therapeutic Advances in Endocrinology and Metabolism 00(0)
8 journals.sagepub.com/home/tae
concentrations and reduced Adipo-IR by around
47% (p = 0.03). According to Belfort and col-
leagues, pioglitazone, compared with placebo,
significantly reduced serum transaminases,
decreased fasting plasma insulin, and decreased
FFA levels.21 Additionally, pioglitazone treat-
ment resulted in significantly greater improve-
ments in steatosis (65% versus 38%, p = 0.003),
ballooning (54% versus 24%, p = 0.02), and com-
bined mean necroinflammation score (44% versus
12%, p = 0.001) over placebo. Consistent with
other trials assessing TZDs, there was no signifi-
cant difference in reduction of fibrosis for piogl-
itazone over placebo (46% versus 33%, p = 0.08).
A recently published trial by Cusi and colleagues
describes the results from a three-year study of
efficacy and safety of pioglitazone in participants
with prediabetes or T2DM and biopsy-proven
NASH.22 After the initial 18 months, more
patients randomized to pioglitazone achieved the
primary endpoint of at least two points’ improve-
ment in the NAS than placebo (58% versus 17%,
p < 0.00). In addition, pioglitazone was more
effective than placebo at NASH resolution (51%
versus 19%, p < 0.001) and mean change in fibro-
sis scores (–0.5 versus 0, p = 0.039). Despite this,
progression of any fibrosis continued in both
groups but was significantly lower in the pioglita-
zone group compared with placebo (12% versus
28%, p = 0.039). Histologic and metabolic
improvements were maintained for the entire
study period of three years.
Safety. While it has been observed that long-term
treatment with a TZD is necessary to sustain clin-
ical improvements, concern exists over the safety
of prolonged use. No serious adverse events
reported were related to TZD treatment.16–22 In
the three-year study, no osteoporosis, osteopo-
rotic bone fractures, or bladder cancer was
detected.22 The most notable adverse events
reported were reduction in hemoglobin, median
decrease 0.7 g/dl (0.1–3.1 g/dl),16 lower limb
edema,17,21,22 and weight gain (1.5–6.4 kg).16,17,21
Glucagon-like peptide-1 receptor agonists
Agents in the glucagon-like peptide-1 receptor ago-
nist (GLP-1 RA) class improve glycemic control in
individuals with T2DM through multiple mecha-
nisms, including glucose-dependent insulin secre-
tion, decreased glucagon secretion, slowed gastric
emptying, and enhanced satiety.23 Historically,
GLP-1 receptors have been identified in the
pancreas, kidney, lung, gastric mucosa, heart,
hypothalamus,24 and most recently in the liver.25 In
murine models, GLP-1 RAs were shown to improve
transaminase levels, reduce oxidative stress, and
reduce hepatic steatosis, making them viable
options for the treatment of NASH (Table 3).26–28
Efficacy. Two trials assessed the metabolic and
hepatic effects of exenatide immediate release.
Fan and colleagues compared exenatide with
metformin in participants with T2DM and
NAFLD.27 At baseline, 52% of participants had
abnormal liver function, with 46 participants hav-
ing an ALT over 2.5 times the upper limit of nor-
mal (ULN). The second study, conducted by
Shao and colleagues, compared exenatide plus
insulin glargine U-100 with intensive insulin
treatment with insulin glargine U-100 plus insu-
lin aspart.28 Included patients had hepatic injury
biomarkers between 2.5 and 5 times the ULN;
‘normal’ for each was defined as ALT or AST up
to 40 U/liter and γ glutamyl transferase (γ-GT) up
to 50 U/liter. Exenatide was initiated at 5 μg twice
daily for the first 4 weeks to minimize gastrointes-
tinal effects and titrated to 10 μg twice daily for
the remaining 8 weeks in both studies.
All arms of both studies showed improvement in
hepatic markers. Fan and colleagues found exena-
tide to be superior to metformin in improving
ALT, AST, and γ-GT.27 Additionally, C-reactive
protein (CRP) was significantly decreased and
adiponectin was significantly increased in the
exenatide arm, suggesting improved oxidative
stress. Mean reductions in body weight and BMI
were statistically significant for the exenatide
group compared with the metformin group.
Lastly, both exenatide and metformin improved
insulin resistance similarly, measured by
HOMA-IR. In the study by Shao and colleagues
comparing exenatide plus glargine with glargine
plus aspart, body weight and waist circumference
were significantly decreased in the exenatide arm,
but increased in the intensive insulin arm.28 The
post-treatment mean for ALT, AST, and γ-GT
levels was statistically lower in the exenatide arm
compared with the insulin only arm. Exenatide in
combination with glargine was also superior to
insulin alone in the reversal rate of fatty liver dis-
ease, which was 93.3% and 66.7% (p < 0.01),
respectively.
The efficacy of liraglutide in patients with T2DM
and NASH was assessed in two separate trials by
the same primary investigator. In the first published
EP Mills, K Trotta et al.
journals.sagepub.com/home/tae 9
Table 3. Efficacy and safety trials of glucagon-like peptide 1 receptor agonists for NAFLD in patients with T2DM.
Study Treatment Comparator Number of
participants
Duration ALT mean
change from
baseline
(SD)
AST mean
change
from
baseline
(SD)
γ-GT mean
change
from
baseline
(SD)
Imaging or
histologic
changes
Other significant
measurements
Armstrong
et al.
(LEAD)25
LIR 0.6, 1.2,
and 1.8 mg
PBO
GLI 4 mg/day
149
(all with DM)
26 weeks LSAR
improvement with
LIR versus PBO:
mean difference
+0.10 (95% CI
–0.01 to 0.20;
p = 0.07)
Armstrong
et al.
(LEAN)26
LIR 1.8 mg PBO 45
(17 with DM)
48 weeks LIR: 26.6
(34.4); PBO:
10.2 (35.8);
p = 0.16
LIR: 15.8
(21.8); PBO:
8.6 (28.3);
p = 0.29
LIR: 33.7
(42.5); PBP:
7.2 (28.3);
p = 0.01
Mean ∆ BMI (SD): LIR,
1.8 (1.67); PBO, 0.3 (1.7);
p = 0.005
Mean ∆ HOMA-IR (SD):
LIR, 1.8 (3.7); PBO, 0.70
(9.49); p = 0.23
Fan et al.27 EXE 5 μg
twice daily
for 30 days,
increased to
10 μg twice
daily
MET 500 mg
twice daily,
adjusted up
to 2 g/day
117
(all with DM)
12 weeks EXE: 27.32
(15.96);
MET: 12.85
(11.38); p =
0.002
EXE: 7.89
(7.87); MET:
5.11 (6.98);
p = 0.048
EXE: 26.48
(17.34);
MET: 10.26
(14.11); p =
0.000
Mean ∆ CRP (SD): EXE,
0.89 (0.59); MET, 0.61
(0.54); p = 0.018
Mean ∆ adiponectin
(SD): EXE, 1.86 (2.22);
MET, 0.76 (1.3); p =
0.001
Mean ∆ BMI (SD): EXE,
1.31 (0.98); MET, 0.69
(0.94); p = 0.000
Mean ∆ HOMA-IR (SD):
EXE, 0.57 (0.36); MET
0.56 (0.49); p = 0.367
Shao
et al.28
iGLAR daily
(+) EXE 5 μg
twice daily
for 30 days,
increased to
10 μg twice
daily
iGLAR daily
(+) iASP
three times
daily
60
(all with DM)
12 weeks EXE: 42.51
(13.12); INS
only: 67.37
(15.78); p <
0.001
EXE: 32.28
(8.71); INS
only: 42.90
(10.0); p <
0.001
EXE: 34.37
(10.05); INS
only: 43.36
(3.60); p <
0.001
Reversal rate
of fatty liver
(regression from
greater to lower
degree of fatty
liver): EXE, 93.3%;
INS only, 66.7%;
p < 0.01
Mean ∆ BMI (SD):
ALT, alanine aminotransferase; AST, aspartate aminotransferase, BMI, body mass index; CI, confidence interval; CRP, C-reactive protein; EXE, exenatide; INS, insulin; iASP, insulin
aspart; iGLAR, insulin glargine; γ-GT, γ glutamyl transferase; GLI, glimepiride; HOMA-IR, homeostatic model assessment insulin resistance; LEAN, liraglutide efficacy and action in
NASH; LIR, liraglutide; LSAR, liver-to-spleen attenuation ratio; MET, metformin; NS, not significant; PBO, placebo; SD, standard deviation; DM diabetes mellitus.
Therapeutic Advances in Endocrinology and Metabolism 00(0)
10 journals.sagepub.com/home/tae
study, the investigators performed a meta-analysis
of the LEAD (liraglutide efficacy and action in dia-
betes) program.25 For the purposes of this article,
only the LEAD-2 substudy is reviewed as it is the
only trial with confirmed presence of fatty liver dis-
ease. In the LEAD-2 substudy, hepatic steatosis
was measured by computer tomography (CT)
imaging at randomization and conclusion of the
study and confirmed in 64.4% of individuals at
baseline. A liver-to-spleen attenuation ratio (LSAR)
of less than 1.0 defined hepatic steatosis and an
improvement in steatosis was an increase in the
LSAR. Participants were given metformin in com-
bination with liraglutide 0.6, 1.2, or 1.8 mg/day or
active placebo (glimepiride 4 mg/day or placebo). A
dose-dependent increase in LSAR was seen with
liraglutide 1.8 mg, but it was nonsignificant. No
significant differences in LSAR were seen between
the lower doses of liraglutide and placebo.
The second study on liraglutide by Armstrong
and colleagues, the LEAN (liraglutide efficacy
and action in NASH) study, is a more robust
assessment of liraglutide in participants with
biopsy-confirmed NASH.26 The study enrolled
52 participants, but only nine participants (35%)
in the liraglutide arm and eight participants (31%)
in the placebo arm had a diagnosis of T2DM.
Liraglutide was titrated over 14 days to 1.8 mg
per day and participants were allowed to remain
on previous treatment with metformin, sulfonylu-
rea, or a combination. Three participants (38%)
with T2DM in the treatment group achieved the
primary outcome of resolution of NASH with no
worsening of fibrosis whereas none of the partici-
pants with T2DM in the placebo arm were able
to achieve this outcome. Progression of fibrosis
was observed in two participants (9%) in the lira-
glutide group and eight participants (36%) in the
placebo group. Compared with placebo, the rela-
tive risk for participants with T2DM taking lira-
glutide achieving resolution of NASH without
worsening fibrosis was 4.7 [95% confidence inter-
val (CI) 0.3–75.0; p = 0.20]. Participants in the
liraglutide arm did have statistically significant
decreases in body weight, BMI, and γ-GT levels.
Interestingly, this study included participants
with stage 3 fibrosis and cirrhosis; study investiga-
tors observed that participants with more
advanced disease had positive treatment effects
from liraglutide, but not as pronounced as partici-
pants with mild to moderate disease.
Safety. Despite the slow titration of exenatide over
4 weeks, gastrointestinal side effects were listed in
both exenatide trials as the predominant side
effect in the treatment arm, but did not contribute
to study withdrawl.27,28 Adverse events were simi-
lar between liraglutide and placebo in the LEAN
study, with the exception of gastrointestinal disor-
ders, which were more common in the liraglutide-
treated arm.27 The information provided on safety
from the LEAD-2 substudy is underwhelming
and not delineated between the main LEAD pro-
gram analysis and LEAD-2 substudy; available
safety data indicate that gastrointestinal side
effects and hepatobiliary serious adverse events
were comparable for liraglutide 1.2, liraglutide
1.8, and placebo for participants with normal and
abnormal ALT levels at baseline.25
Antihyperlipidemics
In addition to the benefits HMG-CoA (3-hydroxy-
3-methyl-glutaryl-coenzyme A) reductase inhibi-
tors (statins) have on lipids, they improve insulin
sensitivity, decrease production of advanced gly-
cation endproducts (AGEs), and display anti-
inflammatory effects, all of which may be helpful
in treating the steatosis and inflammation associ-
ated with NASH.29,30 Several studies evaluate the
use of lipid medications as treatment options for
NASH, including atorvastatin, simvastatin, rosu-
vastatin, pitavastatin, ezetimibe/simvastatin com-
bination, and ursodeoxycholic acid (UDCA), a
bile acid used to reduce cholesterol absorption
(Table 4).30
Efficacy. In 2003, Kiyici and colleagues completed
a prospective study comparing UDCA with atorv-
astatin in the treatment of NASH.30 In a small
study of 44 patients, both groups saw significant
lowering in ALT and γ glutamyl transferase
(GGT), an enzyme used as a diagnostic marker
for liver disease (p < 0.02). The atorvastatin group
at baseline had higher cholesterol levels; after the
study period, a decrease in serum cholesterol was
seen in the atorvastatin group as well as a statisti-
cally significant normalization of transaminases
post treatment (p = 0.021). Imaging studies found
that liver densities did increase in the atorvastatin
group. There was no change in BMI, serum glu-
cose, or triglyceride levels in either group.
The PITCH study, a 2012 prospective rand-
omized open-label trial by Han and colleagues
compared pitavastatin (2–4 mg per day) with
atorvastatin (10–20 mg/day).31 Over 12 weeks,
the 135 study participants showed a statistically
significant lowering (p < 0.05) in serum GGT
EP Mills, K Trotta et al.
journals.sagepub.com/home/tae 11
Table 4. Efficacy and safety trials of HMG CoA reductase inhibitors for NAFLD in patients with T2DM.
Study Treatment Comparator Number of
participants
Duration
(months)
ALT mean change
from baseline
AST mean
change from
baseline
γ-GT mean
change from
baseline
Imaging or histologic
changes
Other significant
measurements
Kimura
et al.29
ATO 10 mg 43
(31 with IGT or
DM)
12 –33.5 U/liter; p <
0.001
–15.8 U/liter; p <
0.001
–25.3 U/liter; p <
0.001
Liver density increased;
p < 0.001
Necroinflammatory
grade improved; p < 0.05
NAS
Improved: 68%
Unchanged: 27%
Worsened: 5%
∆ BMI, FBG: unchanged
Fibrosis stage
Improved: 9%
Unchanged: 59%
Worsened: 32%
Kiyici
et al.30
UDCA 13–15 mg/
kg/day in
normolipidemic
patients
ATO 10
mg/day in
hyperlipidemic
patients
44
(10 with DM)
6 UDCA: –19 U/liter;
p = 0.002
ATO: –37 U/liter;
p = 0.0001
UDCA versus ATO:
NS
UDCA: NS
ATO: –13 U/liter;
p = 0.004
UDCA versus
ATO: p = 0.033
UDCA: –15.6 U/
liter (p = 0.016)
ATO: –27 U/liter
(p = 0.014)
UDCA versus
ATO: NS
UDCA: NS
ATO: significant increase
in liver density (improved
steatosis); p = 0.0001
UDCA versus ATO: NS
change in steatosis
BMI, serum glucose and TG
level changes: NS for both
groups
Han
et al.31
PIT 2–4 mg/day ATO 10–20 mg/
day
189
(53 with DM)
12 weeks PIT: –5 U/liter,
p = 0.047
ATO, NS
PIT, NS
ATO, NS
PIT: –10.9 U/liter,
p = 0.034
ATO: –11.1 U/
liter, p = 0.040
Hepatic steatosis
improvement:
PIT, p = 0.008
ATO, NS
Equal reduction in LDL, p <
0.0001
Increase in ALT: 17 patients; 2
severe (>3 × ULN)
Hyogo
et al.32
ATO 10 mg/day 31
(22 with IGT or
DM)
24 –53.5 U/liter,
p < 0.001
–25.3 U/liter, p <
0.001
–36 U/liter, p <
0.001
Steatosis grade and NAS:
improved, p < 0.001
Fibrosis: worsened in 4
patients
Transaminases normalized in
74.2% of patients
∆ BMI, FBG, HbA1c, NS
∆ HS CRP: decrease, p < 0.05
∆ HOMA-IR, NS
Abel
et al.33
SIM 20 mg/day EZE/SIM 10/10
mg/day
45
(all with DM)
6 SIM: –37 7 U/liter,
p < 0.0001
EZE/SIM: –31 U/
liter, p < 0.0001
SIM versus EZE/SIM,
p < 0.0112
SIM: –35.6 U/
liter, p < 0.0001
EZE/SIM: –27.1
U/liter, p <
0.0001
SIM versus EZE/
SIM, p < 0.0001
NR ∆ TC, HDL, TG: NS in either
group or between groups
∆ LDL: SIM versus EZE/SIM,
p = 0.0063
∆ CK U/L: NS in either group
Nelson
et al.34
SIM 40 mg/day PBO 16
(7 with DM)
12 SIM, NS
PBO, NS
SIM versus PBO, NS
SIM, NS
PBO, NS
SIM versus PBO,
NS
NR ∆ hepatic steatosis,
necroinflammatory
activity, fibrosis stage:
NS in either group or
between groups
∆ TC, LDL, TG: NS in either
group or between groups
ALT, alanine aminotransferase; AST, aspartate aminotransferase; ATO, atorvastatin; BMI, body mass index; CK, creatine kinase; CRP, C-reactive protein; DM, diabetes mellitus; EZE/
SIM, ezetimibe/simvastatin; FBG, fasting blood glucose; γ-GT, γ glutamyl transferase; HbA1c, glycosylated hemoglobin; HDL, high-density lipoprotein; HOMA-IR, homeostatic model
assessment insulin resistance; HS CRP, high-sensitivity C-reactive protein; IGT, impaired glucose tolerance; LDL, low-density lipoprotein; NAS, NAFLD Activity Score; NR, not reported;
NS, not significant; PBO, placebo; PIT, pitavastatin; TG, triglyceride; UDCA, ursodeoxycholic acid; ULN, upper limit of normal; SIM simvastatin TC total cholesterol.
Therapeutic Advances in Endocrinology and Metabolism 00(0)
12 journals.sagepub.com/home/tae
concentrations and LDL cholesterol (p < 0.0001)
from baseline in both treatment groups. Only the
pitavastatin group had significantly reduced ALT,
which was the primary endpoint. Tomography
revealed that both groups reduced hepatic steato-
sis severity in patients with overt fatty liver before
randomization.
Two additional studies investigated the use of
atorvastatin 10 mg per day in combination with
standard weight loss counselling.29,32 A controlled
trial by Hyogo and colleagues followed patients
over 24 months, twice the length of patients fol-
lowed in the open-label trial by Kimura and col-
leagues. Hyogo and colleagues found a mean
change in both ALT and AST from baseline
whereas Kimura also found a change in γ-GTP.
No changes in BMI or serum glucose were found
in either study; however, Hyogo and colleagues
also evaluated for changes in adiponectin, tumor
necrosis factor α, leptin, and long chain fatty
acids. Overall, no statistically significant changes
were found in these values. Both studies showed
improvement in NAFLD score and liver steatosis
grade. Kimura and colleagues additionally
reviewed the effects of atorvastatin on AGEs, as
they are commonly increased in patients with
NASH. Atorvastatin was found to decrease AGEs
significantly.29,32
A randomized, double-blind, placebo-controlled
trial by Nelson and colleagues investigated the
use of simvastatin 40 mg versus placebo in the
treatment of NASH.34 Over 12 months, 16
patients were not found to have a statistically sig-
nificant improvement in ALT or AST from base-
line, hepatic steatosis, necroinflammatory activity,
or fibrosis stage for either the treatment or pla-
cebo group. An additional retrospective trial by
Abel and colleagues compared simvastatin 20 mg
with ezetimibe/simvastatin 10 mg/10 mg over six
months.33 Both groups resulted in a statistically
significant change in ALT and AST from baseline
(p < 0.0001 for all groups), and simvastatin mon-
otherapy decreased both ALT and AST signifi-
cantly more than combination therapy (p <
0.0112 and p < 0.0001, respectively). Overall,
there was no difference between the two groups in
regards to cholesterol decrease, triglyceride
reduction, and HDL elevation.
Safety. Adverse effects of statin therapies used in
the aforementioned trials ranged from elevations
in ALT to a progression of fibrosis. Of the atorv-
astatin studies, there was no report of an elevation
in transaminases.29,30,32 However, the PITCH
study reported an elevation in ALT in both the
pitavastatin and atorvastatin treatment groups,
with one study participant from each treatment
group being excluded from analysis as a result of
severely elevated ALT.31 Progression in fibrosis
staging was found in atorvastatin treatment
groups in two of the studies that utilized atorvas-
tatin.29,32 The studies involving simvastatin did
not reveal any adverse effects of the therapy.33,34
Discussion
Based on the reviewed studies, metformin, TZDs,
GLP-1 RAs, and statins all appear to be safe
options for the treatment of NAFLD/NASH in
patients with concomitant T2DM, but efficacy
data surrounding each vary. Metformin shows lit-
tle positive impact on histological markers associ-
ated with NAFLD. The benefit from metformin
treatment can be attributed to improvement in
weight and metabolic profile. Data from reviewed
studies on metformin reiterated that weight man-
agement in patients with NAFLD has the most
benefit on steatosis.7
The mechanism of TZDs to improve insulin
sensitivity in the liver, muscle, and adipose tis-
sue has shown effectiveness in reversing NASH
in up to half of treated patients, but may need to
be continued indefinitely to avoid return to base-
line.16,17 TZDs provide significant histologic and
metabolic improvements overall, but did not
provide significant differences in fibrosis com-
pared with placebo. While it is not fully under-
stood why some patients do not respond to TZD
treatment, one trial found a higher rate of nonre-
sponders in those with T2DM.17 Conversely,
another trial reported NAS scores significantly
improved in those with diabetes versus those
without.18
Improvements in liver disease in patients on
GLP-1 RAs were detected through reduced
hepatic enzymes and liver histology via biopsy or
imaging in patients with NAFLD and T2DM.
These improvements may be attributed to the
beneficial effects GLP-1 RAs have on liver
inflammation, insulin resistance, and body
weight. Based on the available data, GLP-1 treat-
ment appears to be of benefit in individuals with
mild to moderate NAFLD and T2DM, and may
offer some advantages in advanced disease (e.g.
cirrhosis), albeit a lessened effect for a very costly
medication.
EP Mills, K Trotta et al.
journals.sagepub.com/home/tae 13
Atorvastatin is the most commonly studied statin
in the treatment of NAFLD/NASH. Overall,
atorvastatin appears to be moderately beneficial
in decreasing transaminases, the severity of
hepatic steatosis, and the NAS score. One study
also noted the importance of reducing AGEs and
the ability of atorvastatin to decrease those levels
in patients with dyslipidemia. However, the over-
all ability of AGEs to be utilized as an indicative
biomarker for NASH warrants further investiga-
tion.29 Certain trials included patients only with
dyslipidemia; therefore, the role of statins in the
treatment of NASH in patients with normal lipids
needs to be further investigated.30
Along with diet, exercise, and glycemic control,
the discussed medications may be a viable option
for the treatment of NAFLD. In addition to their
insulin-sensitizing benefits, they may improve the
prognosis of NAFLD in patients with T2DM by
decreasing the risk of serious consequences, such
as cardiovascular disease and hepatocellular car-
cinoma.35 Conflicting trial results, small cohorts,
and short study durations emphasize the need for
continued studies on the most viable and effica-
cious pharmacologic treatment options for
patients with NAFLD and T2DM.
Funding
This research received no specific grant from any
funding agency in the public, commercial, or not-
for-profit sectors.
Conflict of interest statement
The authors declare that there is no conflict of
interest.
References
1. Angulo P. Nonalcoholic fatty liver disease. N Engl
J Med 2002; 346: 1221–1231.
2. Farrell GC and Larter CZ. Nonalcoholic
fatty liver disease: from steatosis to cirrhosis.
Hepatology 2006; 43(2 Suppl. 1): S99–S112.
3. Buzzetti E, Pinzani M and Tsochatzis EA. The
multiple-hit pathogenesis of non-alcoholic fatty
liver disease (NAFLD). Metabolism 2016; 65:
1038–1048.
4. Li Y, Liu L, Wang B, etal. Metformin in non-
alcoholic fatty liver disease: a systematic review
and meta-analysis. Biomed Rep 2013; 1: 57–64.
5. Loomba R, Lutchman G, Kleiner DE, etal.
Clinical trial: pilot study of metformin for the
treatment of non-alcoholic steatohepatitis.
Aliment Pharmacol Ther 2009; 29: 172–182.
6. Bellentani S, Scaglioni F, Marino M, etal.
Epidemiology of non-alcoholic fatty liver disease.
Dig Dis 2010; 28: 155–161.
7. Chalasani N, Younossi Z, Lavine JE, etal. The
diagnosis and management of non-alcoholic fatty
liver disease: practice Guideline by the American
Association for the Study of Liver Diseases,
American College of Gastroenterology, and the
American Gastroenterological Association.
Am J Gastroenterol 2012; 107: 811–826.
8. Nair S, Diehl AM, Wiseman M, etal. Metformin
in the treatment of non-alcoholic steatohepatitis:
a pilot open label trial. Aliment Pharmacol Ther
2004; 20: 23–28.
9. Mazza A, Fruci B, Garinis GA, etal. The role of
metformin in the management of NAFLD. Exp
Diabetes Res 2012; 2012: 716404.
10. Haukeland JW, Konopski Z, Eggesbo HB, etal.
Metformin in patients with non-alcoholic fatty
liver disease: a randomized, controlled trial.
Scand J Gastroenterol 2009; 44: 853–860.
11. Nar A and Gedik O. The effect of metformin
on leptin in obese patients with type 2 diabetes
mellitus and nonalcoholic fatty liver disease. Acta
Diabetol 2009; 46: 113–118.
12. Yki-Jarvinen H. Thiazolidinediones. N Engl J
Med 2004; 351: 1106–1118.
13. Shulman GI. Cellular mechanisms of insulin
resistance. J Clin Invest 2000; 106: 171–176.
14. Riera-Guardia N and Rothenbacher D. The
effect of thiazolidinediones on adiponectin serum
level: a meta-analysis. Diabetes Obes Metab 2008;
10: 367–375.
15. Cunard R, Ricote M, DiCampli D, etal.
Regulation of cytokine expression by ligands of
peroxisome proliferator activated receptors.
J Immunol 2002; 168: 2795–2802.
16. Neuschwander-Tetri BA, Brunt EM, Wehmeier
KR, etal. Improved nonalcoholic steatohepatitis
after 48 weeks of treatment with the PPAR-
gamma ligand rosiglitazone. Hepatology 2003; 38:
1008–1017.
17. Ratziu V, Giral P, Jacqueminet S, etal.
Rosiglitazone for nonalcoholic steatohepatitis:
one-year results of the randomized placebo-
controlled fatty liver improvement with
rosiglitazone therapy (FLIRT) trial.
Gastroenterology 2008; 135: 100–110.
18. Torres DM, Jones FJ, Shaw JC, etal.
Rosiglitazone versus rosiglitazone and metformin
versus rosiglitazone and losartan in the treatment
Therapeutic Advances in Endocrinology and Metabolism 00(0)
14 journals.sagepub.com/home/tae
of nonalcoholic steatohepatitis in humans: a
12-month randomized, prospective, open-label
trial. Hepatology 2011; 54: 1631–1639.
19. Omer Z, Cetinkalp S, Akyildiz M, etal. Efficacy
of insulin-sensitizing agents in nonalcoholic fatty
liver disease. Eur J Gastroenterol Hepatol 2010; 22:
18–23.
20. Gastaldelli A, Harrison SA, Belfort-Aguilar R,
etal. Importance of changes in adipose tissue
insulin resistance to histological response during
thiazolidinedione treatment of patients with
nonalcoholic steatohepatitis. Hepatology 2009; 50:
1087–1093.
21. Belfort R, Harrison SA, Brown K, etal. A
placebo-controlled trial of pioglitazone in subjects
with nonalcoholic steatohepatitis. N Engl J Med
2006; 355: 2297–2307.
22. Cusi K, Orsak B, Bril F, etal. Long-term
pioglitazone treatment for patients with
nonalcoholic steatohepatitis and prediabetes or
type 2 diabetes mellitus: a randomized trial. Ann
Intern Med 2016; 165: 305–315.
23. Uccellatore A, Genovese S, Dicembrini I, etal.
Comparison review of short-acting and long-
acting glucagon-like peptide-1 receptor agonists.
Diabetes Ther 2015; 6: 239–256.
24. Pyke C, Heller RS, Kirk RK, etal. GLP-1
receptor localization in monkey and human
tissue: novel distribution revealed with extensively
validated monoclonal antibody. Endocrinology
2014; 155: 1280–1290.
25. Armstrong MJ, Houlihan DD, Rowe IA, etal.
Safety and efficacy of liraglutide in patients
with type 2 diabetes and elevated liver enzymes:
individual patient data meta-analysis of the
LEAD program. Aliment Pharmacol Ther 2013;
37: 234–242.
26. Armstrong MJ, Gaunt P, Aithal GP, etal.
Liraglutide safety and efficacy in patients
with non-alcoholic steatohepatitis (LEAN): a
multicentre, double-blind, randomised, placebo-
controlled phase 2 study. Lancet 2016; 387:
679–690.
27. Fan H, Pan Q, Xu Y, etal. Exenatide improves
type 2 diabetes concomitant with non-alcoholic
fatty liver disease. Arq Bras Endocrinol Metabol
2013; 57: 702–708.
28. Shao N, Kuang HY, Hao M, etal. Benefits of
exenatide on obesity and non-alcoholic fatty liver
disease with elevated liver enzymes in patients
with type 2 diabetes. Diabetes Metab Res Rev
2014; 30: 521–529.
29. Kimura Y, Hyogo H, Yamagishi S, etal.
Atorvastatin decreases serum levels of
advanced glycation endproducts (AGEs) in
nonalcoholic steatohepatitis (NASH) patients
with dyslipidemia: clinical usefulness of AGEs
as a biomarker for the attenuation of NASH. J
Gastroenterol 2010; 45: 750–757.
30. Kiyici M, Gulten M, Gurel S, etal.
Ursodeoxycholic acid and atorvastatin in the
treatment of nonalcoholic steatohepatitis. Can J
Gastroenterol 2003; 17: 713–718.
31. Han KH, Rha SW, Kang HJ, etal. Evaluation
of short-term safety and efficacy of HMG-CoA
reductase inhibitors in hypercholesterolemic
patients with elevated serum alanine transaminase
concentrations: PITCH study (PITavastatin
versus atorvastatin to evaluate the effect on
patients with hypercholesterolemia and mild to
moderate hepatic damage). J Clin Lipidol 2012; 6:
340–351.
32. Hyogo H, Tazuma S, Arihiro K, etal. Efficacy
of atorvastatin for the treatment of nonalcoholic
steatohepatitis with dyslipidemia. Metabolism
2008; 57: 1711–1718.
33. Abel T, Feher J, Dinya E, etal. Safety and
efficacy of combined ezetimibe/simvastatin
treatment and simvastatin monotherapy in
patients with non-alcoholic fatty liver disease.
Med Sci Monit 2009; 15: MS6–MS11.
34. Nelson A, Torres DM, Morgan AE, etal. A
pilot study using simvastatin in the treatment
of nonalcoholic steatohepatitis: a randomized
placebo-controlled trial. J Clin Gastroenterol 2009;
43: 990–994.
35. Baffy G. Hepatocellular carcinoma in non-
alcoholic fatty liver disease: epidemiology,
pathogenesis, and prevention. Journal of Clinical
and Translational Hepatology 2013; 1: 131–137.
Visit SAGE journals online
journals.sagepub.com/
home/tae
SAGE journals
