Adiponectin as a target for the treatment of nonalcoholic steatohepatitis with thiazolidinediones: A systematic review

ME TAB O L IS M CL I N ICA L A N D EX P ER IM EN T AL 6 5 ( 2 0 16 ) 12 9 7 – 1 30 6

Available online at www.sciencedirect.com

Metabolism www.metabolismjournal.com

Adiponectin as a target for the treatment of nonalcoholic steatohepatitis with thiazolidinediones: A systematic review Stergios A. Polyzos a,⁎, Christos S. Mantzoros b a

Medical School, Aristotle University of Thessaloniki, Thessaloniki, Greece Division of Endocrinology, Diabetes and Metabolism, Department of Internal Medicine, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA, USA b

A R T I C LE I N FO

AB S T R A C T

Article history:

Thiazolidinediones (TZDs; pioglitazone and rosiglitazone) have provided promising results in

Received 19 May 2016

clinical trials for nonalcoholic steatohepatitis (NASH). The main purpose of this systematic

Accepted 23 May 2016

review was to summarize evidence on circulating adiponectin levels in relation to histological changes following TZD treatment in patients with histologically confirmed NASH. We performed a systematic search in PubMed, Scopus and Cochrane Library. We included four

Keywords:

studies, published between 2006 and 2012, providing data for 187 histologically confirmed NASH

Adiponectin

adult patients (105 on TZD and 82 controls) treated for 6–12 months. Significant increase in

Nonalcoholic steatohepatitis

adiponectin (80–178%) after TZD treatment was observed in all included studies. Improvement

Pioglitazone

in steatosis following treatment was observed in all studies. A trend towards improvement in

Rosiglitazone

lobular inflammation was observed in all studies after pioglitazone, but not after rosiglitazone.

Thiazolidinediones

Trends toward improvement in ballooning and fibrosis were observed in the two studies after pioglitazone using either the highest doses or the longest duration of therapy. Overall disease activity score was improved in all studies after pioglitazone, but not after rosiglitazone. Insulin resistance and liver function tests were also improved after treatment. Despite weight gain, circulating leptin was not increased after treatment. In conclusion, parallel increases in circulating adiponectin levels and histological improvement were observed in this systematic review. These results warrant further consideration of TZDs, but even more importantly point to a key role for novel potential treatments for NASH patients such as the newer selective peroxisome proliferator activated receptor-γ modulators, which increase adiponectin without significant weight gain. © 2016 Elsevier Inc. All rights reserved.

1.

Introduction

Nonalcoholic fatty liver disease (NAFLD) is a global health problem, affecting 6–45% of the general population, rising up to

70% in patients with type 2 diabetes mellitus (T2DM) and 90% in morbidly obese patients [1]. NAFLD starts as simple steatosis (SS) and can progress to nonalcoholic steatohepatitis (NASH), characterized by the addition of hepatic inflammation and

Abbreviations: ALT, Alanine transaminase; AST, Aspartate transaminase; BMI, Body mass index; ELISA, Enzyme-linked immunosorbent assay; HbA1c, Glycosylated hemoglobin; HOMA-IR, Homeostasis model of assessment insulin resistance; IR, Insulin resistance; NAFLD, Nonalcoholic fatty liver disease; NAS, NAFLD activity score; NASH, Nonalcoholic steatohepatitis; PPAR, Peroxisome proliferator activated receptor; RIA, Radioimmunoassay; SS, Simple steatosis; T2DM, Type 2 diabetes mellitus; TZD, Thiazolidinedione. ⁎ Corresponding author at: 13 Simou Lianidi, 551 34 Thessaloniki, Macedonia, Greece. Tel./fax: +30 2310424710. E-mail address: [email protected] (S.A. Polyzos). http://dx.doi.org/10.1016/j.metabol.2016.05.013 0026-0495/© 2016 Elsevier Inc. All rights reserved.

1298

ME TAB O L IS M CL I N ICA L A N D EX PE R IM EN T AL 6 5 ( 2 0 16 ) 12 9 7 – 13 0 6

fibrosis [2]. NAFLD has been linked to hepatic (i.e., cirrhosis and hepatocellular carcinoma) and extra-hepatic morbidity, including cardiovascular disease and malignancies, which all contribute to higher mortality observed in NASH patients [3]. Despite its high prevalence and the intensive research on the field, the treatment of NAFLD remains an unmet medical need, since no pharmacological treatment has been approved to-date for NAFLD [4]. Adipokines, hormones secreted by adipose tissue, seem to participate in the pathogenesis of NAFLD, thereby being potential pharmacologic targets [5]. In this regard, the properties of adiponectin render it an ideal target, since it decreases IR and has anti-apoptotic, anti-steatotic, anti-inflammatory and antifibrogenic effects, as previously reviewed [6]. In a meta-analysis of 27 observational studies with histologically confirmed NAFLD patients, we showed that circulating adiponectin levels decreased from controls to SS patients and even more to NASH patients [7]. Thus, we have proposed that increasing adiponectin levels are beneficial to NASH and/or hepatic fibrosis [8,9]. Although recombinant adiponectin has been shown to exert hepatoprotective effects in mice with NASH [10,11], it is rather difficult to produce functionally active recombinant adiponectin, due to the need for extensive posttranslational modifications [12]. Therefore, it is more feasible to pharmacologically upregulate endogenous adiponectin, for which thiazolidinediones (TZDs) are currently considered to be the agents of choice [6]. TZDs (pioglitazone and rosiglitazone) are approved treatment for T2DM, usually as a second-line orally administered drugs [13]. TZDs have also been used in several clinical trials for NASH, because they primarily target insulin resistance (IR), which is considered to play a crucial role in the pathogenesis of both T2DM and NAFLD [14]. In the most recent meta-analysis of randomized controlled trials in histologically confirmed NASH patients, TZDs have been shown to improve steatosis, lobular inflammation and ballooning, whereas their effect on fibrosis improvement did not marginally reach the level of statistical significance [15]. Two previous meta-analyses provided similar results in histological outcomes, although improvement in fibrosis was marginally significant in them [16,17]. The aim of this systematic review is to focus on the TZD – adiponectin interplay as a potential pharmacologic axis for the treatment of NASH; therefore, circulating adiponectin levels were summarized in parallel with histological changes following TZD treatment in patients with NASH. Secondary aims were the effect of TZD on body mass index (BMI), liver function tests, circulating leptin levels and IR.

2.

Patients and Methods

2.1.

Literature Search

This systematic review was conducted following an a priori established protocol according to the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) statement for randomized controlled trials [18] and the Metaanalysis Of Observational Studies in Epidemiology (MOOSE)

statement for observational studies [19]. We performed a computerized literature search using the PubMed, Scopus and Cochrane Library electronic databases. Search was not limited by language or publication time. Medical Subject Heading (MeSH) database was used as a terminological search filter. From the combination of terminological (MeSH terms) and methodological search filters, as proposed elsewhere in detail [7], the following query was formatted: “(NASH OR NAFLD OR (nonalcoholic fatty liver disease) OR (non-alcoholic fatty liver disease) OR (nonalcoholic steatohepatitis) OR (non-alcoholic steatohepatitis)) AND adiponectin AND (pioglitazone OR rosiglitazone OR thiazolidinediones)”. This query was used with minimal differences in format, according to the requirements of different databases. The literature search was extended by “hand searching” in the “related citations” links of all included articles in PubMed (first 60 articles per included article, after sorting according to the relevance), and the references of all included articles, as proposed elsewhere in detail [7].

2.2.

Inclusion and Exclusion Criteria

Human interventional studies of any design, providing circulating adiponectin levels for NASH patients before and after any TZD treatment were eligible for this systematic review. Studies were included in this systematic review, if: 1) they were original full-text publications; 2) they were randomized controlled trials or prospective cohort studies; 3) NASH was histologically confirmed at baseline; 4) any TZD treatment was administered in NASH patients of at least one arm (or the unique arm) of the study; 5) a repeat liver biopsy was performed at the end of the study; 6) circulating adiponectin levels were measured at both baseline and end of the study. Studies were excluded from this systematic review, if: 1) patients had other causes of liver disease (e.g., alcoholic fatty liver disease, viral or autoimmune hepatitis) or NAFLD coexisted with other liver disease(s); 2) there was overlap of patients, i.e., the same participants were included in more than one study; 3) additional data were considered to be necessary, but the corresponding authors did not respond; 4) retrieved articles were reviews, editorials, case series or case reports, letters to the editor, hypotheses, book chapters, studies on animals or cell lines; 5) they were unpublished data or abstracts from conferences.

2.3.

Data Extraction

All retrieved articles were saved in an EndNote file (Thompson Reuters, Philadelphia, PA), so as to facilitate their handling. Two reviewers (S.A.P. and C.S.M.), qualified in systematic review, reviewed independently the title and abstract of all initial search results and excluded studies that did not address the research question, based on our pre-specified inclusion and exclusion criteria. If multiple publications from the same cohort were found, only data from the largest comprehensive study were included. For articles published in a language other than English, a list of required items was requested from the corresponding authors. Subsequently, review of full-text articles was independently carried out by the aforementioned

Identification

ME TAB O L IS M CL I N ICA L A N D EX P ER IM EN T AL 6 5 ( 2 0 16 ) 12 9 7 – 1 30 6

143 Potentially relevant articles identified in initial database search 32 PubMed 98 Scopus 13 Cochrane Central

1299

2 Additional articles identified by hand searching

27 Duplicates removed

Screening

118 Potentially relevant articles evaluated 105 Excluded 77 Reviews 7 Congress abstracts 1 Book chapter 1 Hypothesis 17 Non-human studies (animal or cell line) 2 Human observational studies

Eligibility

13 Potentially relevant articles evaluated in detail

6 Excluded 4 Not histologically confirmed NASH 1 Repeat biopsy was not performed 1 Corresponding author did not provide necessary data

Included

7 Articles initially selected for the systematic review

3 Excluded 2 Patients’ overlap 1 Effect of pioglitazone discontinuation on NASH 4 Articles finally included in the systematic review

Fig. 1 – A flowchart presenting the literature search process, according to the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) statement.

reviewers. In case of disagreement between the two reviewers, it was resolved by consensus. We extracted the following variables from each study: 1) study's general characteristics (reference, name of the first author; year of publication; country where the study was carried out; design); 2) study's special characteristics (duration, histological definition of NASH, method of adiponectin measurement, frequency of NASH-related cirrhosis and T2DM); 3) patient characteristics (number, sex, age, BMI); 4) histological characteristics (steatosis, fibrosis, ballooning, lobular and portal inflammation); 5) laboratory measurements (adiponectin, leptin, aspartate aminotransferase [AST], alanine aminotransferase [ALT]); 6) evaluation of IR by homeostasis model of assessment insulin resistance (HOMA-IR) or other indices. When needed, corresponding (and/or first) author(s) of the eligible studies were contacted by e-mail to request additional data. Conflicts of data extraction were resolved by consensus after discussion and/or referral back to the corresponding (or first) author(s).

2.4.

Outcomes

The main outcome of this systematic review was circulating adiponectin levels together with specific hepatic histological changes (steatosis, inflammation and fibrosis) before and after any TZD treatment in NASH patients. BMI, AST, ALT, circulating leptin and IR, before and after TZD treatment were also evaluated as secondary outcomes.

3.

Results

3.1.

Literature Search

We initially retrieved 32 articles from PubMed, 98 from Scopus and 13 from Cochrane Library (last update January 1, 2016). Two additional articles were added from the references of the selected articles. A flowchart, which was prepared according to the PRISMA guidelines [18], summarizes the

1300

First author year, Study design origin [reference] a

Active medication; dose (mg/per day)

Duration Histological (months) definition

Cirrhosis T2DM Method of N (%) N (%) adiponectin

Additional information

measurement Belfort 2006, USA [29] Randomized, double blind, placebo-controlled trial

Pioglitazone; 30 (2 months) 45 (4 months)

Lutchman 2006, USA [30] Ratziu 2008, France [31]

Pioglitazone; 30 mg Rosiglitazone; 4 (1 month) 8 (11 months)

Open-label, uncontrolled, cohort study Randomized, double blind, placebo-controlled trial

Sharma 2012, India Open-label, randomized controlled trial; [32] two active arms

Pioglitazone; 30 vs. Pentoxifylline; 1200

6

NAS

0 (0%)

23 (49%)

RIA (plasma)

12

Unspecified

na

na

ELISA (serum)

12

NAS; Brunt (fibrosis)

2 (3%)

20 (32%)

RIA (serum)

Brunt

0 (0%)

na

ELISA (plasma)

6

Another group of healthy controls (n = 10) evaluated at baseline was not included. No control group. 1) Baseline data were available for 52 (83%) patients (24 rosiglitazone and 28 placebo) for adiponectin. 2) Paired measurements were available in 22 patients (12 rosiglitazone and 10 placebo) for adiponectin. No placebo group; active comparator (pentoxifylline).

Abbreviations: ELISA, enzyme-linked immunosorbent assay; IGT, impaired glucose tolerance; na, not available; NAS, nonalcoholic fatty liver disease activity score; RIA, radioimmunoassay; T2DM, type 2 diabetes mellitus. a References are presented in alphabetical order according to the first author.

ME TAB O L IS M CL I N ICA L A N D EX PE R IM EN T AL 6 5 ( 2 0 16 ) 12 9 7 – 13 0 6

Table 1 – Main characteristics of the studies included in the systematic review.

Table 2 – Main anthropometric and laboratory characteristics per group of the studies included in the systematic review. N Age (males) baseline (years)

BMI AST AST ALT BMI baseline end-point baseline end-point baseline (IU/l) (IU/l) (IU/l) (kg/m2) (kg/m2)

ALT end-point (IU/l)

IR b IR b baseline end-point

Belfort 2006 [29]

Placebo

21 (7)

51 ± 10

32.9 ± 4.4 32.7 ± 4.5

42 ± 16

33 ± 10

61 ± 33

40 ± 17 ⁎

na

Pioglitazone

26 (14)

51 ± 7

33.5 ± 4.9 34.6 ± 5.7 ⁎

47 ± 15

28 ± 7 ⁎

67 ± 26

28 ± 12 ⁎

na

Lutchman 2006 [30]

No control Pioglitazone

18 (7)

45.8 ± 10.6 32.4 ± 5.7 33.7 ± 6.3 ⁎

61 ± 36

34 ± 15

99 ± 71

40 ± 25

4.3 ± 3.0

Ratziu 2008 [31]

Placebo

31 (18)

54.1 ± 10.4 30.5 ± 4.4 na

61 ± 46

na

84 ± 38

na

4.2 (4.3) e,f 4.8 (na) e,f,d

Rosiglitazone

32 (19)

53.1 ± 11.5 31.5 ± 6.0 na

46 ± 26

na

69 ± 40

Pentoxifylline 30 (19)

37.3 ± 7.2

24.2 ± 2.7 24.2 ± 3.6

63 ± 20

28 ± 10 ⁎

87 ± 32

na 4.6 (4.7) e,f 3.2 (na) e,f,d,⁎ (decreased) 37 ± 20 ⁎ 2.4 ± 1.8 1.6 ± 1.2 ⁎

Pioglitazone

40.4 ± 9.9

25.7 ± 3.6 26.2 ± 10.2

67 ± 17

28 ± 9 ⁎

106 ± 28

Sharma 2012 [32]

29 (13)

34 ± 16 ⁎

3.3 ± 1.5

Leptin end-point (ng/ml)

Adiponectin Adiponectin baseline end-point (μg/ml) (μg/ml)

na na (unchanged) c na na (improved) c

na

7.5 ± 1.0 d

8.0 ± 1.1 d

na

6.0 ± 0.6 d

14.2 ± 1.1 d,⁎

2.6 ± 1.9 ⁎

29.6 ± 22.9

3.7 ± 1.9

10.3 ± 8.6 ⁎

na (unchanged) na (unchanged) 23.6 (3.4-52.3) g,⁎ 14.4 (14.2-155.1) g,⁎

5.5 ± 0.5 d

5.2 ± 0.6 d

5.7 ± 0.6 d

11.2 ± 2.1 d,⁎

2.1 ± 0.6 ⁎

Leptin baseline (ng/ml)

24.8 ± 16.9 14.1 (11.4) e 15.2 (14.6) e 4.5 (1.8-37.9) g 13.5 (6.3-64.6) g

4.4 ± 1.9

7.2 ± 4.7

4.0 ± 1.9

7.2 ± 2.3 ⁎

Data are presented as mean ± standard deviation (SD) for continuous variables and as absolute frequencies for categorical variables, unless otherwise stated. Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; HOMA-IR, homeostasis model of assessment–insulin resistance; IR, insulin resistance; na, not available. a References are presented in alphabetical order according to the first author; b IR is presented as HOMA-IR, unless otherwise stated; c Hepatic and adipose tissue insulin resistance; d Approximately calculated from data; e Median (interquartile range); f In nondiabetic patients only; g Median (range). ⁎ Statistically significant compared to baseline (p < 0.05).

ME TAB O L IS M CL I N ICA L A N D EX P ER IM EN T AL 6 5 ( 2 0 16 ) 12 9 7 – 1 30 6

Group First author year [reference] a

1301

1302

ME TAB O L IS M CL I N ICA L A N D EX PE R IM EN T AL 6 5 ( 2 0 16 ) 12 9 7 – 13 0 6

Table 3 – Changes in the main histological lesions per group of the studies included in the systematic review. First author year [reference] a

Group

Steatosis

Lobular inflammation

Portal inflammation

Ballooning

Fibrosis

Activity score b

Belfort 2006 [29]

Placebo Pioglitazone No control Pioglitazone Placebo Rosiglitazone Pentoxifylline Pioglitazone

Unchanged Improved

Improved Improved

na na

Unchanged Improved

Unchanged Improved

Unchanged Improved

Improved Unchanged Improved Improved Improved

Improved Unchanged Unchanged Unchanged Improved

Unchanged na na Unchanged Improved

Improved Unchanged Unchanged Unchanged Unchanged

Improved Unchanged Unchanged Unchanged Unchanged

Improved Unchanged Unchanged Unchanged Improved

Lutchman 2006 [30] Ratziu 2008 [31] Sharma 2012 [32]

Abbreviations: na, not available; NAS, nonalcoholic fatty liver disease activity score. a References are presented in alphabetical order according to the first author. b NAS or NASH activity index score or Brunt's grade.

identification, screening, eligibility and final selection of the studies included in the systematic review (Fig. 1). After a first screening, we excluded 105 studies and we further evaluated for eligibility 13 studies. We further excluded six studies (NASH was not histologically confirmed in four [20–23], paired liver biopsies were not performed in one [24] and necessary data were not provided by the corresponding author of one study [25]). The main characteristics of the studies excluded from the systematic review at the stage of eligibility are presented in Supplementary Table 1. We finally excluded two studies, because of patients' overlap [26,27] and one study, because it reported data on pioglitazone discontinuation [28]. Subsequently, we included four studies in this systematic review [29–32]. However, data from the two studies with patients' overlap [26,27] were complimentary to the initial study [29] and were added and discussed. Since the retrieved data did not facilitate the performance of a meta-analysis, we communicated with both the first and senior authors of all included studies, but none responded to our invitation to provide additional data or perform data transformations. This fact deterred us from conducting a meta-analysis.

3.2.

Characteristics of the Included Studies

The four included studies were published between 2006 and 2012, and reported data on 105 adult NASH patients on TZD and 82 adult NASH patients on a comparator (52 on placebo and 30 on pentoxifylline) [29–32]. The main characteristics of these studies are presented in Table 1. Lutchman et al. published the first relevant study in 2006 [29]. Two studies were carried out in North America [29,30], one in Europe [31] and one in Asia [32]. Two studies were randomized, double blind, placebo-controlled trials [29,30], comparing pioglitazone [29] or rosiglitazone [30] vs. placebo. Another study was an open-label, randomized controlled trial, comparing pioglitazone vs. pentoxifylline [32]. The fourth one was a prospective, open-label, uncontrolled, cohort study, evaluating the effect of pioglitazone [30]. Liver histology was classified according to Brunt et al. criteria [33] in one study [32], and NAFLD activity score (NAS) [34] in another [29]. Both

Brunt et al. and NAS criteria were used in another study [31], whereas the criteria are unknown for the last one [30]. Circulating total adiponectin levels were measured with radioimmunoassay (RIA) in two studies [29,31] and enzymelinked immunosorbent assay (ELISA) in the other two studies [30,32]. The main demographic and laboratory characteristics per group of each study included in the systematic review are presented in Table 2. We reported circulating adiponectin levels, age, sex, BMI, AST, ALT, leptin levels and IR, when available.

3.3.

Outcomes

3.3.1.

Main Outcomes

Circulating adiponectin levels per group of the included studies are presented in Table 2. Significant increase in adiponectin was observed in all included studies [29–32]. The magnitude of increase in adiponectin was 80–178%; more specifically, 80% in Sharma et al. (pioglitazone 30 mg × 6 months) [32], 97% in Ratziu et al. (rosiglitazone 4 mg × 1 month and then 8 mg × 11 months) [31], 137% in Belfort et al. (pioglitazone 30 mg × 2 months and then 45 mg × 4 months for a total of 6 months) [29] and 178% in Lutchman et al. (pioglitazone 30 mg × 12 months) [30]. Regarding the comparator arms, no statistically significant difference was observed in adiponectin levels in any study [29,31,32] (Table 2); more specifically, adiponectin increased by 64% after pentoxifylline treatment, whereas remained essentially unchanged in placebo groups (increase by 7% in Belfort et al. [29] and decrease by 6% in Ratziu et al. [31]). Changes in the main histological lesions per group of the included studies are presented in Table 3. Improvement in steatosis following TZD treatment was observed in all studies [29–32]. A trend towards improvement in lobular inflammation was observed in all studies after pioglitazone [29,30,32], but not after rosiglitazone [31]. Trends toward improvement in ballooning or fibrosis were observed in the two studies after pioglitazone using either the highest doses (pioglitazone 30 mg × 2 months and then 45 mg × 4 months for a total of 6 months) or 30 mg for the longest duration of therapy

ME TAB O L IS M CL I N ICA L A N D EX P ER IM EN T AL 6 5 ( 2 0 16 ) 12 9 7 – 1 30 6

(12 months) [29,30], whereas this trend did not appear in the remaining two studies with either rosiglitazone or pioglitazone 30 mg for 6 months [31,32]. Limited data on portal inflammation show a trend towards a rather neutral effect of TZDs (Table 3). Overall disease activity score was improved in all studies after pioglitazone [29,30,32], but not after rosiglitazone [31]. Association between changes in circulating adiponectin and histological lesions were evaluated in three studies [29–31]. An inverse relationship between the reduction of hepatic steatosis, necroinflammation, fibrosis and overall disease activity score, and the increase in circulating adiponectin was observed in Belfort et al. study [29], as reported in another study where a subanalysis is being reported [27]. Lutchman et al. also reported an inverse relationship between the reduction of hepatic steatosis and overall disease activity score, and the increase in circulating adiponectin [30]. Finally, Ratziu et al. reported an inverse relationship between the reduction of hepatic steatosis and the increase in circulating adiponectin [31].

3.3.2.

Secondary Outcomes

As expected, TZD treatment increased BMI in two studies [29,30] (Table 2) and weight in another one, for which endpoint BMI was not available [31]. Notably, despite weight increase, circulating leptin levels were not significantly affected (Table 2). TZD treatment also decreased IR in all studies (Table 2). HOMA-IR was used as an index of IR in three studies [30–32], whereas hepatic and adipose tissue IR was used in one study [29], as reported in another study where a sub-analysis is being reported [26]. Regarding AST and ALT levels, they were decreased after TZD treatment, although the difference for AST did not reach the level of statistical significance in some studies (Table 2).

3.3.3.

Adverse Effects

Both pioglitazone and rosiglitazone were reported to be well tolerated. One patient had myocardial infarction during treatment in one study [30], as reported in another study where a sub-analysis is being reported [35]. Belfort et al. also reported that coronary artery disease was diagnosed in two patients (one in placebo and one in pioglitazone group) [29]. No other serious adverse events were observed in the included studies, including bladder cancer or osteoporotic fracture. Weight gain (mean by 2.0–3.5 kg) was the main adverse effect. Swollen legs, muscular cramps, headache and fatigue were observed in a small number of patients. Only one patient discontinued treatment because of painful swollen legs (attributed to rosiglitazone) [31] and one because of dizziness after pioglitazone [30]. Another patient experienced autoimmune iritis, which, however, was regarded as unrelated to pioglitazone, and the patient did not discontinue treatment [30]. There was also a slight but significant increase in serum cholesterol and low-density lipoprotein cholesterol levels in rosiglitazone-treated patients [31].

4.

Discussion

In this systematic review, we summarized data on adiponectin and hepatic histological lesions following TZD treatment in patients with histologically confirmed NASH. Circulating adiponectin levels were substantially increased in all studies

1303

(by 80–178%) after TZD treatment (Table 2). This increase parallels the degree of improvement in hepatic histology (Table 3). Furthermore, there is evidence for a direct association between increase in adiponectin levels and histological improvement [29–31]. Although there was a tendency toward improvement in all histological lesions, this did not reach the level of statistical significance in all cases. In this regard, it seems that the most stable finding was the improvement in steatosis, which was evident in all included studies. Moreover, there was an inverse association between increase in adiponectin and improvement in steatosis in all studies providing relevant data (three of four). The findings regarding inflammation showed a tendency towards improvement. Regarding fibrosis, data should be more carefully considered, despite the marginally favorable results of two meta-analyses [16,17]. Although there is a trend towards improvement from baseline to the end of the study in two of the included studies [29,30], the difference did not remain significant in one of them, after subtracting the placebo effect [29]. Fibrosis is regarded as a hard prognostic endpoint of NASH, because it is independently associated with progression to cirrhosis, long-term overall mortality and liver transplantation [36]. However, the length of intervention in all studies (6–12 months) may be short to provide definite results on fibrosis. Nevertheless, an improvement in steatosis seems to be easier to achieve, but it may be a positive prognostic sign, thus predicting improvement in inflammation and fibrosis on a longer-term treatment, as we previously hypothesized [14]. Beyond histological improvement, the increase in adiponectin levels parallels also the improvement in IR and liver function tests (AST and ALT) (Table 2). Based on the main mechanism of action, TZDs target IR and this is their main pathophysiologic action in patients with T2DM [13]. Targeting IR may also be the main mechanism through which they exert a beneficial effect on NASH, since IR is a key pathogenetic factor of NASH [2]. However, TZDs, as peroxisome proliferator activated receptor (PPAR)-γ ligands, exert pleiotropic actions beyond IR, including anti-inflammatory, anti-atherosclerotic and anti-cancer ones, which may also play a role on NASH. Differences in the molecule of TZD may explain their differences in effectiveness, but also in adverse events. It should be highlighted that the first TZD introduced in the treatment of T2DM, troglitazone, was withdrawn owing to relatively rare (1/20,000 patients), but severe liver failure [37]. However, the next two TZDs, rosiglitazone and pioglitazone, were not associated with hepatic adverse events [37], and they also had therapeutic value for NASH. Contrary to other pharmacologic agents (e.g. orlistat, sibutramine, rimonabant), which may lead to an increase in adiponectin only when weight loss is achieved and due to reduced body weight [6], TZDs increase adiponectin directly and irrespective of any weight gain, usually observed with these drugs. Weight gain usually results in both decreased adiponectin levels and an unfavourable effect on NAFLD, but it seems that TZDs induce adiponectin production and secretion, which overcomes any effect of the drugs on body weight [6]. Notably, circulating leptin levels in NASH patients receiving TZDs do not seem to increase (Table 2). It has been shown that TZD-induced PPAR-γ activation reduces leptin expression in adipose tissue [38]. This direct effect may

1304

ME TAB O L IS M CL I N ICA L A N D EX PE R IM EN T AL 6 5 ( 2 0 16 ) 12 9 7 – 13 0 6

contradict the increase in leptin due to increased adipose tissue after TZD use, thereby TZDs have a neutral effect on leptin levels. This is considered to be of paramount importance in NASH, because experimental studies showed an adverse effect of leptin administration on hepatic inflammation and fibrosis, despite its favorable effect on hepatic steatosis [39]. Although this has not been shown in humans yet, leptin administration in normo- or hyper-leptinemic NASH patients is not warranted, as we have proposed elsewhere in detail [40,41]. We have also shown higher leptin levels in NASH than SS patients in a meta-analysis of observational studies [42]. In this regard, the use of TZDs for the treatment of NASH is further supported, since they increase adiponectin without affecting leptin, despite the weight gain. It is currently unknown how quickly TZDs increase circulating adiponectin in NASH patients. Adiponectin was measured at baseline and at the endpoint in all included studies, but there is no information about when adiponectin levels start to increase. However, evidence from in vitro studies in both 3T3-L1 and human adipocytes demonstrated that pioglitazone or rosiglitazone acutely (in hours) stimulate adiponectin secretion [43,44]. Phase I studies with INT131 (a selective PPAR-γ modulator) have shown that adiponectin increases in response to INT131 in both a dose and a time response manner [45]. Data on the effect of TZD discontinuation on adiponectin levels are very limited. Lutchman et al. followed-up 13 patients of the cohort included in this systematic review [30] for one more year after pioglitazone discontinuation [28]. Stopping pioglitazone was associated with subsequent decrease in adiponectin (from 9.7 ± 9.1 to 5.1 ± 4.5 μg/ml; by 90%) together with worsening of lobular inflammation and steatosis, albeit no change in fibrosis [28]. ALT and HOMA-IR were also increased 12 months after pioglitazone discontinuation [28]. Although it remains to be definitively shown, these proof-of-concept data indicate that long-term therapy with TZDs may be required to maintain both adiponectin levels' increase and histological improvement, as in T2DM patients, in whom discontinuation of TZD results in deterioration of glycemic control [14]. Currently, the use of rosiglitazone has been restricted because of an increase in myocardial infarction risk [46], and pioglitazone use has been suspended in some European countries because of a possible slight increase in bladder cancer risk after long-term use in T2DM patients [47]. Nevertheless, selective PPAR-γ modulators have been developed, including the aforementioned INT131 (formerly AMG131) [45]. INT131 is a potent non-thiazolidinedione selective PPAR-γ designed to exhibit a biological profile of strong efficacy, but minimal side effects compared to PPAR-γ full agonists [45]. In phase 2 trials, INT131 was well tolerated and improved, in a dosedependent manner, glycated hemoglobin (HbA1c) compared with placebo in T2DM patients not receiving pharmacotherapy (phase 2a) [48] or inadequately controlled with metformin and/or sulfonylureas (phase 2b) [49]. On the other hand, less adverse effects, including edema, fluid retention and weight gain were observed compared with rosiglitazone [48] or pioglitazone [49]. Regarding the effect of INT131 on adiponectin, a 15-day treatment with INT131 was reported to increase adiponectin

levels in a dose-dependent manner and more than pioglitazone or rosiglitazone in Zucker rats. Stimulation of adiponectin was also seen in healthy volunteers in a phase I trial and confirmed in T2DM patients in phase 2 trials [48,49]. Based on these observations, INT131 is a very promising candidate for clinical trials in NASH patients, since INT131 would have a beneficial effect on hepatic histology without the adverse effects observed with rosiglitazone or pioglitazone. This systematic review has certain limitations: 1) The number of included studies is small, as well as the total number of patients, owing largely to the need for paired liver biopsies. The relatively short duration of the studies may also not be sufficient to show histological improvement, especially for a hard endpoint like fibrosis. Despite the limited number of patients, however, the study's end points were reached, indicating that studies with 187 subjects (as was the total in all studied to date) or even better with more than 225 subjects (20% higher than the above number) for up to 12 months would most probably result in highly statistically significant results in all parameters to be studied. 2) Existing evidence in human NASH cannot prove a causal effect: TZD-adiponectinNASH improvement. However, experimental studies suggest that TZDs exert their beneficial effect on NASH through elevating adiponectin levels [50,51]. 3) A formal meta-analysis of primary data could not be performed, because additional necessary data were not provided. 4) Data on concomitant antidiabetic medications and statins are not presented in detail in the included studies. Although they might have affected liver histology [52], it is expected that concomitant medications were equally distributed between groups, at least in the three randomized studies. 5) By excluding unpublished studies or abstracts from conferences, a publication bias may have occurred. Excluding such studies, however, is essential to avoid introduction of lowquality data, since the quality of unpublished data and abstracts cannot be ascertained [53]. 6) Levels of γ-glutamyltransferase were not included, because data were available in only one of the included studies [31]. In conclusion, despite a 2.5–3 kg weight gain observed after pioglitazone or rosiglitazone, an increase in circulating adiponectin levels which parallels histological improvements were observed in this systematic review. Improvement in steatosis was a consistent and pronounced finding, whereas improvement in lobular inflammation, ballooning and fibrosis was less pronounced, at least partly owing to relatively small sample sizes and short duration of the included studies. Furthermore, liver function tests and IR were also improved after treatment, whereas the effect of TZDs on leptin levels was rather neutral. These results warrant further consideration of TZDs, but more importantly newer selective PPAR-γ modulators, as potential treatment for NASH patients.

Authors' Contribution Both authors have contributed to the design, conduct of the study/data collection, interpretation of data and writing of the manuscript.

ME TAB O L IS M CL I N ICA L A N D EX P ER IM EN T AL 6 5 ( 2 0 16 ) 12 9 7 – 1 30 6

Financial Disclosure This study was financially supported by InteKrin Therapeutics Inc. (Redwood City, CA). However, the funders had no role in the design of this systematic review, data collection, data analysis, data interpretation or the writing of the manuscript.

Acknowledgements We sincerely thank InteKrin Therapeutics Inc. (Redwood City, CA) for financially supporting this systematic review. However, the InteKrin Therapeutics Inc. had no role in the design of this systematic review, data collection, data analysis, data interpretation or the writing of the manuscript.

Conflict of Interest SAP and CSM received consulting fees from InteKrin Therapeutics Inc. and CSM is a shareholder of Coherus Biosciences Inc.

Appendix A. Supplementary Data Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.metabol.2016.05.013.

REFERENCES

[1] Fazel Y, Koenig AB, Sayiner M, Goodman ZD, Younossi ZM. Epidemiology and natural history of nonalcoholic fatty liver disease. Metabolism 2016. http://dx.doi.org/10.1016/j.metabol.2016.01.012. [2] Polyzos SA, Kountouras J, Zavos C. Nonalcoholic fatty liver disease: the pathogenetic roles of insulin resistance and adipocytokines. Curr Mol Med 2009;72:299–314. [3] Rinella ME. Nonalcoholic fatty liver disease: a systematic review. JAMA 2015;313:2263–73. [4] Chalasani N, Younossi Z, Lavine JE, Diehl AM, Brunt EM, Cusi K, et al. 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. Hepatology 2012;55:2005–23. [5] Polyzos SA, Kountouras J, Mantzoros CS. Adipokines in nonalcoholic fatty liver disease. Metabolism 2016. http://dx. doi.org/10.1016/j.metabol.2015.11.006. [6] Polyzos SA, Kountouras J, Zavos C, Tsiaousi E. The role of adiponectin in the pathogenesis and treatment of nonalcoholic fatty liver disease. Diabetes Obes Metab 2010;12:365–83. [7] Polyzos SA, Toulis KA, Goulis DG, Zavos C, Kountouras J. Serum total adiponectin in nonalcoholic fatty liver disease: a systematic review and meta-analysis. Metabolism 2011;60:313–26. [8] Polyzos SA, Kountouras J, Zavos C. Adiponectin as a potential therapeutic agent for nonalcoholic steatohepatitis. Hepatol Res 2010;40:446–7. [9] Polyzos SA, Kountouras J, Zavos C. Adiponectin in non-alcoholic fatty liver disease treatment: therapeutic perspectives and unresolved dilemmas. Int J Clin Pract 2011;65:373–4. [10] Xu A, Wang Y, Keshaw H, Xu LY, Lam KS, Cooper GJ. The fatderived hormone adiponectin alleviates alcoholic and nonalcoholic fatty liver diseases in mice. J Clin Invest 2003;112:91–100.

1305

[11] Fukushima J, Kamada Y, Matsumoto H, Yoshida Y, Ezaki H, Takemura T, et al. Adiponectin prevents progression of steatohepatitis in mice by regulating oxidative stress and Kupffer cell phenotype polarization. Hepatol Res 2009;39: 724–38. [12] Simpson F, Whitehead JP. Adiponectin-it's all about the modifications. Int J Biochem Cell Biol 2010;42:785–8. [13] Sauer S. Ligands for the nuclear peroxisome proliferator-activated receptor gamma. Trends Pharmacol Sci 2015;36:688–704. [14] Polyzos SA, Kountouras J, Zavos C, Deretzi G. Nonalcoholic fatty liver disease: multimodal treatment options for a pathogenetically multiple-hit disease. J Clin Gastroenterol 2012;46:272–84. [15] Singh S, Khera R, Allen AM, Murad MH, Loomba R. Comparative effectiveness of pharmacological interventions for nonalcoholic steatohepatitis: a systematic review and network meta-analysis. Hepatology 2015;62:1417–32. [16] Musso G, Cassader M, Rosina F, Gambino R. Impact of current treatments on liver disease, glucose metabolism and cardiovascular risk in non-alcoholic fatty liver disease (NAFLD): a systematic review and meta-analysis of randomised trials. Diabetologia 2012;55:885–904. [17] Mahady SE, Webster AC, Walker S, Sanyal A, George J. The role of thiazolidinediones in non-alcoholic steatohepatitis - a systematic review and meta analysis. J Hepatol 2011;55:1383–90. [18] Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Ann Intern Med 2009;151:264–9. [19] Stroup DF, Berlin JA, Morton SC, Olkin I, Williamson GD, Rennie D, et al. Meta-analysis of observational studies in epidemiology: a proposal for reporting. Meta-analysis Of Observational Studies in Epidemiology (MOOSE) group. JAMA 2000;283:2008–12. [20] Cui KQ, Zhao XW, Zhang Y, Kang XH, Meng J, Chen XL. Efficacy of rosiglitazone in treatment of nonalcoholic fatty liver disease and its relations with adiponectin. World Chin J Digestology 2006;14:1326–9. [21] Gupta AK, Bray GA, Greenway FL, Martin CK, Johnson WD, Smith SR. Pioglitazone, but not metformin, reduces liver fat in type-2 diabetes mellitus independent of weight changes. J Diabetes Complications 2010;24:289–96. [22] Zhu C, Zhou H, Han Y, Ling L, Huang G. Application of adiponectin, TNF-α and ferrtin in type 2 diabetes with intrahepatic lipid infiltration. Nuclear Techniques 2008;31:356–9. [23] Yang L, Song MQ, Zhang QL, Shou L, Zang SF, Yang YL. Effect of piglitazone and metformin on retinol-binding protein-4 and adiponectin in patients with type 2 diabetes mellitus complicated with non-alcohol fatty acid liver diseases. Zhongguo Yi Xue Ke Xue Yuan Xue Bao 2014;36:309–12. [24] Saryusz-Wolska M, Szymanska-Garbacz E, Jablkowski M, Bialkowska J, Pawlowski M, Kwiecinska E, et al. Rosiglitazone treatment in nondiabetic subjects with nonalcoholic fatty liver disease. Pol Arch Med Wewn 2011;121:61–6. [25] Aithal GP, Thomas JA, Kaye PV, Lawson A, Ryder SD, Spendlove I, et al. Randomized, placebo-controlled trial of pioglitazone in nondiabetic subjects with nonalcoholic steatohepatitis. Gastroenterology 2008;135:1176–84. [26] Gastaldelli A, Harrison SA, Belfort-Aguilar R, Hardies LJ, Balas B, Schenker S, et al. Importance of changes in adipose tissue insulin resistance to histological response during thiazolidinedione treatment of patients with nonalcoholic steatohepatitis. Hepatology 2009;50:1087–93. [27] Gastaldelli A, Harrison S, Belfort-Aguiar R, Hardies J, Balas B, Schenker S, et al. Pioglitazone in the treatment of NASH: the role of adiponectin. Aliment Pharmacol Ther 2010;32:769–75. [28] Lutchman G, Modi A, Kleiner DE, Promrat K, Heller T, Ghany M, et al. The effects of discontinuing pioglitazone in patients with nonalcoholic steatohepatitis. Hepatology 2007;46:424–9.

1306

ME TAB O L IS M CL I N ICA L A N D EX PE R IM EN T AL 6 5 ( 2 0 16 ) 12 9 7 – 13 0 6

[29] Belfort R, Harrison SA, Brown K, Darland C, Finch J, Hardies J, et al. A placebo-controlled trial of pioglitazone in subjects with nonalcoholic steatohepatitis. N Engl J Med 2006;355: 2297–307. [30] Lutchman G, Promrat K, Kleiner DE, Heller T, Ghany MG, Yanovski JA, et al. Changes in serum adipokine levels during pioglitazone treatment for nonalcoholic steatohepatitis: relationship to histological improvement. Clin Gastroenterol Hepatol 2006;4:1048–52. [31] Ratziu V, Giral P, Jacqueminet S, Charlotte F, Hartemann-Heurtier A, Serfaty L, et al. Rosiglitazone for nonalcoholic steatohepatitis: one-year results of the randomized placebo-controlled Fatty Liver Improvement with Rosiglitazone Therapy (FLIRT) Trial. Gastroenterology 2008;135:100–10. [32] Sharma BC, Kumar A, Garg V, Reddy RS, Sakhuja P, Sarin SK. A randomized controlled trial comparing efficacy of pentoxifylline and pioglitazone on metabolic factors and liver histology in patients with non-alcoholic steatohepatitis. J Clin Exp Hepatol 2012;2:333–7. [33] Brunt EM, Janney CG, Di Bisceglie AM, Neuschwander-Tetri BA, Bacon BR. Nonalcoholic steatohepatitis: a proposal for grading and staging the histological lesions. Am J Gastroenterol 1999;94:2467–74. [34] Kleiner DE, Brunt EM, Van NM, Behling C, Contos MJ, Cummings OW, et al. Design and validation of a histological scoring system for nonalcoholic fatty liver disease. Hepatology 2005;41:1313–21. [35] Promrat K, Lutchman G, Uwaifo GI, Freedman RJ, Soza A, Heller T, et al. A pilot study of pioglitazone treatment for nonalcoholic steatohepatitis. Hepatology 2004;39:188–96. [36] Angulo P, Kleiner DE, Dam-Larsen S, Adams LA, Bjornsson ES, Charatcharoenwitthaya P, et al. Liver fibrosis, but no other histologic features, is associated with long-term outcomes of patients with nonalcoholic fatty liver disease. Gastroenterology 2015;149:389–97. [37] Nathan DM. Diabetes: Advances in Diagnosis and Treatment. JAMA 2015;314:1052–62. [38] Dunn TN, Akiyama T, Lee HW, Kim JB, Knotts TA, Smith SR, et al. Evaluation of the synuclein-gamma (SNCG) gene as a PPARgamma target in murine adipocytes, dorsal root ganglia somatosensory neurons, and human adipose tissue. PLoS One 2015;10, e0115830. [39] Imajo K, Fujita K, Yoneda M, Nozaki Y, Ogawa Y, Shinohara Y, et al. Hyperresponsivity to low-dose endotoxin during progression to nonalcoholic steatohepatitis is regulated by leptin-mediated signaling. Cell Metab 2012;16:44–54. [40] Polyzos SA, Kountouras J, Mantzoros CS. Leptin in nonalcoholic fatty liver disease: a narrative review. Metabolism 2015;64:60–78. [41] Polyzos SA, Kountouras J, Zavos C, Deretzi G. The potential adverse role of leptin resistance in nonalcoholic fatty liver

[42]

[43]

[44]

[45]

[46]

[47]

[48]

[49]

[50]

[51]

[52]

[53]

disease: a hypothesis based on critical review of literature. J Clin Gastroenterol 2011;45:50–4. Polyzos SA, Aronis KN, Kountouras J, Raptis DD, Vasiloglou MF, Mantzoros CS. Circulating leptin in non-alcoholic fatty liver disease: a systematic review and meta-analysis. Diabetologia 2015;59:30–43. Pereira RI, Leitner JW, Erickson C, Draznin B. Pioglitazone acutely stimulates adiponectin secretion from mouse and human adipocytes via activation of the phosphatidylinositol 3'-kinase. Life Sci 2008;83:638–43. Tishinsky JM, Ma DW, Robinson LE. Eicosapentaenoic acid and rosiglitazone increase adiponectin in an additive and PPARgamma-dependent manner in human adipocytes. Obesity 2011;19:262–8. Higgins LS, Mantzoros CS. The development of INT131 as a Selective PPARgamma Modulator: approach to a safer insulin sensitizer. PPAR Res 2008;2008, 936906. Nissen SE, Wolski K. Rosiglitazone revisited: an updated meta-analysis of risk for myocardial infarction and cardiovascular mortality. Arch Intern Med 2010;170:1191–201. Lewis JD, Ferrara A, Peng T, Hedderson M, Bilker WB, Quesenberry Jr CP, et al. Risk of bladder cancer among diabetic patients treated with pioglitazone: interim report of a longitudinal cohort study. Diabetes Care 2011;34:916–22. Dunn FL, Higgins LS, Fredrickson J, DePaoli AM. Selective modulation of PPARgamma activity can lower plasma glucose without typical thiazolidinedione side-effects in patients with type 2 diabetes. J Diabetes Complications 2011; 25:151–8. DePaoli AM, Higgins LS, Henry RR, Mantzoros C, Dunn FL. Can a selective PPARgamma modulator improve glycemic control in patients with type 2 diabetes with fewer side effects compared with pioglitazone? Diabetes Care 2014;37:1918–23. Liu S, Wu HJ, Zhang ZQ, Chen Q, Liu B, Wu JP, et al. The ameliorating effect of rosiglitazone on experimental nonalcoholic steatohepatitis is associated with regulating adiponectin receptor expression in rats. Eur J Pharmacol 2011;650:384–9. Nan YM, Han F, Kong LB, Zhao SX, Wang RQ, Wu WJ, et al. Adenovirus-mediated peroxisome proliferator activated receptor gamma overexpression prevents nutritional fibrotic steatohepatitis in mice. Scand J Gastroenterol 2011;46:358–69. Nascimbeni F, Aron-Wisnewsky J, Pais R, Tordjman J, Poitou C, Charlotte F, et al. Statins, antidiabetic medications and liver histology in patients with diabetes with non-alcoholic fatty liver disease. BMJ Open Gastro 2016;3, e000075. Egger M, Juni P, Bartlett C, Holenstein F, Sterne J. How important are comprehensive literature searches and the assessment of trial quality in systematic reviews? Empirical study. Health Technol Assess 2003;7:1–76.