Marys Medicine

Hepatmon.com

The Effect of Pioglitazone and Metformin on Liver Function Tests, Insulin Resistance, and Liver Fat Content in Nonalcoholic Fatty Liver Disease: A Ran- domized Double Blinded Clinical TrialMohsen Razavizade 1, Raika Jamali 2, 3, *, Abbas Arj 1, Seyyed Mohammad Matini 1, Alireza Moraveji 4, Effat Taherkhani 11 Internal Medicine Ward, Shahid Beheshti Kashan Hospital, Kashan University of Medical Sciences, Kashan, IR Iran 2 Research Development Center, Sina Hospital, Tehran University of Medical Sciences, Tehran, IR Iran 3 Students Scientific Research Center, Tehran University of Medical Sciences, Tehran, IR Iran 4 Department of Community Medicine, Kashan University of Medical Sciences, Kashan, IR Iran *Corresponding author: Raika Jamali, Students Scientific Research Center, Research Development Center, Sina Hospital, Tehran University of Medical Sci-ences, Tehran, IR Iran. Tel: +98-2163120000, Fax: +98-2163120001, E-mail: [email protected].
A B S T R A C T
Background: Non-alcoholic fatty liver disease (NAFLD) is considered as the hepatic manifestation of insulin resistance (IR) syndrome. The
effect of insulin sensitizers on liver function tests and metabolic indices in NAFLD patients is a matter of debate.
Objectives: The aim of study was to compare the effects of two different insulin sensitizers, pioglitazone, and metformin, on liver function
tests (LFT), lipid profile, homeostasis model assessment-IR (HOMA-IR) index, and liver fat content (LFC) in NAFLD patients.
Materials and Methods: This double blind clinical trial was performed on patients who were referred to a gastroenterology clinic with
evidence of fatty liver in ultrasonography. After excluding other causes, participants with persistent elevated alanine aminotransferase (ALT) levels and "NAFLD liver fat score" greater than -0.64 were presumed to have NAFLD and were enrolled. They were randomly assigned to take metformin (1 g/day) or pioglitazone (30 mg/day) for four months. Fasting serum glucose (FSG), ALT, aspartate aminotransferase (AST), alkaline phosphatase (ALP), triglyceride, cholesterol (CHOL), high and low density lipoprotein (HDL, LDL), HOMA-IR, and LFC were checked at the baseline, two and four months post-treatment. LFC was measured by a validated formula.
Results: Eighty patients (68 males) with mean age of 35.27 (± 7.98) were included. After 2 months, LFT was improved significantly in the
pioglitazone group and did not change in the metformin group. After four months, both medications significantly decreased serum levels of LFT, FSG, CHOL, LDL, HOMA-IR, and LFC, and increased serum level of HDL. No statistically significant differences were seen between the two treatment groups with regard to the changes of laboratory parameters and LFC from baseline to four months post-treatment.
Conclusions: During the four months, the use of metformin (1 g/day) and pioglitazone (30 mg/day) were safe and might have equally affected
LFT, HOMA-IR, lipid profile, and LFC in NAFLD patients.
Keywords: Fatty Liver; Insulin Resistance; Metformin; Pioglitazone Copyright 2013, Kowsar Corp.; Published by Kowsar Corp.
Article type: Research Article; Received: 25 Nov 2012; Revised: 01 Feb 2013; Accepted: 10 Mar 2013; Epub: 21 May 2013 Implication for health policy/practice/research/medical education: The aim of study is to compare the effect of two different insulin sensitizers on liver function tests (LFT), lipid profile, homeosta- sis model assessment-IR (HOMA-IR) index, and liver fat content (LFC) in NAFLD patients. Patients were randomly assigned to take metformin (1 g/day) or pioglitazone (30 mg/day). After four months, both medications led to a significantly decreased serum level of LFT, total serum cholesterol, LDL, HOMA-IR, and LFC, and an increased serum level of HDL. No statistically significant differ- ences were seen between two treatment groups with regard to the changes of laboratory parameters and LFC from baseline to four months post-treatment. This study concluded that use of metformin and pioglitazone are safe and might equally affect LFT, HOMA- IR, lipid profile, and LFC in NAFLD patients in four months.
Please cite this paper as: Razavizade M, Jamali R, Arj A, Matini SM, Moraveji A, Taherkhani E. The Effect of Pioglitazone and Metformin on Liver Function Tests, Insulin Resistance, and Liver Fat Content in Nonalcoholic Fatty Liver Disease: A Randomized Double Blinded Clinical Trial. Hepat Mon. 2013;13(5):e9270. DOI: 10.5812/hepatmon.9270 Copyright 2013, Kowsar Corp.; Published by Kowsar Corp.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which per- mits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Razavizade M et al. Pioglitazone and Metformin in Fatty Liver and liver fat content (LFC) in a sample of NAFLD patients.
Non-alcoholic fatty liver disease (NAFLD) is a common 3. Materials and Methods
cause of chronic hepatitis that can lead to cirrhosis and hepatocellular carcinoma (1, 2). It is already considered as the hepatic manifestation of insulin resistance (meta- 3.1. Ethical Considerations bolic) syndrome (3-5). The prevalence of NAFLD is rising This study was carried out according to the ethical worldwide due to the epidemic of obesity (6-8). Weight standards for human experimentation (Helsinki Decla- loss and obtaining an ideal body weight is a documented ration). The purpose of the study was explained to the treatment option for NAFLD patients who are obese or participants and an informed written consent was taken.
overweight (9). However, the effects of pharmacologi- cal therapies in improvement of liver function tests and 3.2. Liver Ultrasonography metabolic features of NAFLD patients need further inves- tigations. Insulin resistance seems to predispose lipid In this study, the radiologists compared the echogenici- accumulation within the liver and progresses to fibrosis ty of the right lobe of the and right kidney (in the sagittal in NAFLD (10). At present, prescription of insulin sensitiz- view) for the detection of fatty liver. The existence of fat in ing drugs (like pioglitazone and metformin) has come liver parenchyma scatters the beam of ultrasound more to interest as a source to decrease insulin resistance in than a non-fatty liver; therefore, the fatty liver appears NAFLD patients. The improvement of glycemic control, hyperechogenic (25). Although ultrasonography might lipid profile, and insulin sensitivity following the pre- have some limitations for the grading of NAFLD, its avail- scription of these medications in diabetic patients are ability makes it an appropriate tool for NAFLD screening well established. However, their effects on improvement (26). The radiologist was not informed about the clinical of liver function tests, lipid profile, and insulin resistance and laboratory data.
in NAFLD patients remain controversial. Pioglitazone is a peroxisome proliferator activated receptor (PPAR)- 3.3. Subjects gamma agonist that reduces insulin resistance in liver, Patients more than 18 years old with evidence of fatty liv- muscle, and adipose tissue (11). The possible mechanism er in ultrasonography were enrolled in this study. These responsible for insulin sensitivity by pioglitazone is pri- patients were referred to the gastroenterology clinic of a marily promoting fatty acid uptake to adipose tissue that general hospital from January 2011 to January 2012 (Step eventually results in decreasing serum fatty acids (11). 1). Patients with evidence of the following criteria were This process is regulated by adiponectin (ADP) in adipose excluded from the study: alcohol use (more than 20 gram tissue (12). The ability of Pioglitazone to rise serum levels per day in men and 10 gram per day in women), type 1 of ADP and to reduce serum levels of aminotransferase diabetes mellitus, heart disease (ischemic or congestive), has been demonstrated by several trials (13-18). Weight hepatic disease (viral hepatitis, autoimmune hepatitis, gain and fat redistribution from the central area to the wilson disease, hemochromatosis, liver mass lesion), lower body was the most common side effect of this renal disease (serum creatinine concentration of > 1.5 medication (13, 15, 17). Moreover, withdrawn due to hepa- mg/dl), any severe systemic co-morbidities, neoplasm, totoxicity was reported following pioglitazone treatment using any medication during the past 3 months, previ- (14). Metformin is a biguanide drug that improves insulin ous treatment (with thiazolidinediones, biguanides, or sensitivity in the liver and skeletal muscle (19). Antihyper- insulin), and pregnant or lactating women (Step 2). The glycemic effect of metformin is mainly due to decreased serum aminotransferase levels ≥ 40 U/L were consid- gluconeogenesis and a slight effect on glycogenolysis ered elevated (27). All participants with elevated serum (20). Several pilot trials reported that administration of aminotransferase levels in the first blood sample were re- Metformin resulted in reduced insulin resistance and im- checked in one month (lead-in phase). Participants with proved aminotransferase levels without weight gain in elevated aminotransferase levels in the second assess- NAFLD patients (21-23). On the other hand, another study ment were considered as having persistent elevated se- showed no improvement in aminotransferase levels after rum aminotransferase levels. This group of participants administration of metformin (24). Review of the litera- were presumed to have NAFLD and were included in the ture showed controversial results with regards to the ef- study if their "NAFLD liver fat score" value was greater fects of insulin sensitizing medications on liver function than -0.64 (Step 3) (7).
tests, and lipid profile in NAFLD patients.
3.4. NAFLD Liver fat Score and Liver fat Content This randomized double blind clinical trial was de- Performing liver biopsy for the determination of NAFLD signed to compare the effects of pioglitazone and metfor- has some limitations. Many patients due to its invasive- min on liver function tests, lipid profile, HOMA-IR index, ness, cost, and possible complications do not accept this Hepat Mon. 2013;13(5):e9270 Pioglitazone and Metformin in Fatty Liver Razavizade M et al. method. We used "NAFLD liver fat score" instead of liver ated by a clinical epidemiologist who was not aware of biopsy to determine NAFLD in our study (28). This score the treatment modalities using a computer software pro- was calculated as below: gram (Microsoft Office Excel 2007, Microsoft Corp, Red- "NAFLD liver fat score = (- 2.89) + 1.18 * metabolic syn- mond, WA, USA). An investigator who was not involved in drome (yes = 1 / no = 0) + 0.45 * type 2 diabetes (yes = 2 / no data collection and treatment, performed the enrollment = 0) + 0.15 * fasting serum insulin (mU/L) + 0.04 * fasting of patients and their assignments into treatment groups.
serum AST (U/L) – 0.94 * (AST/ALT)".
The values greater than - 0.64 had a sensitivity of 86% 3.9. Intervention and specificity of 71% for the prediction of NAFLD.
Lifestyle modification was the basis of treatment in We used the formula that applied to the same variables this study and was provided to all participants. It con- for the measurement of LFC, as below: sisted of providing a calorie-restricted diet to obtain an "Liver fat content (%) = 10 (-0.805 + 0.282 * metabolic syn- ideal body weight. The protocol used for diet in our study drome (yes = 1 / no = 0) + 0.078 * type 2 diabetes (yes =2 / was based on "Guidelines for the diagnosis and manage- no =0) + 0.525 * log fasting serum insulin (mU/L) + 0.521 * ment of nonalcoholic fatty liver disease: update 2010" log fasting serum AST (U/L) – 0.454 * log (AST/ALT)".
(9). Rapid weight loss was avoided since it could deterio- The previous study had showed the validity of this equa- rate serum aminotransferase levels. A dietitian who was tion for the prediction of LFC considering proton mag- blinded to the study protocol checked the participants netic resonance spectroscopy (PMRS) as the gold stan- and controlled their daily calorie intake during the run dard (28). There was a correlation between LFC identified in period. Eighty patients were randomized to receive ei- by PMRS and LFC calculated by the above formula (r = 0.7, ther 30 mg/day of pioglitazone or 1 g/day of metformin P < 0.0001) (28).
for four months. To minimize the gastrointestinal side 3.5. Study Design effects, metformin was taken at a dose of 500 mg/day. If the patients tolerated metformin, the dose increased This study was designed as a double blind randomized gradually to one gram per day. A study coordinator that controlled clinical trial (IRCT.IR ID: IRCT201105026361N1). was not aware of patient's data gave the medication to During the four-month run in period, a dietitian inter- them every month in a sealed envelope (double blind de- viewed the participants and recommended that they sign). The participants were requested to bring back the should not alter the daily calorie content of their diets.
empty bottles of medications at the follow up visits every month. Compliance (by means of pill count) and adverse 3.6. Outcome Measures effects were checked at follow up visits.
The primary outcome measure was improvement in the serum aminotransferase concentration from baseline 3.10. Study Measurements to the end of treatment at four months. The secondary Height (meter) and weight (kilogram) of the partici- outcome measures were changes in other biochemical pants were measured and body mass index (BMI) was parameters (liver function tests, lipid profile, and HOMA- calculated. Obese subjects were defined if their BMI was IR) and LFC from baseline to the end of treatment at four equal or greater than 30 kg/m2. Laboratory parameters were assessed at the general hospital at baseline, two and four months during the study period. After an over- 3.7. Sample Size Calculation night fast, sera of the participants were tested for fasting serum glucose (FSG), alanine aminotransferase (ALT), A statistical power analysis was utilized to determine aspartate aminotransferase (AST), alkaline phosphatase the sample size. Based on a 50% response to pioglitazone (ALP), triglyceride (TG), cholesterol (CHOL), low-density according to the previous study, and power calculation lipoprotein (LDL) and high density lipoprotein (HDL) by (α = 0.8, β = 0.05), a total sample size of 66 patients was enzymatic methods using Erba Mannheim auto analyzer determined to detect one U/L inter-group difference in XL-640 (Erba Diagnostics Mannheim, Germany). ALT, AST, serum aminotransferase concentration (16). To allow for and ALP levels were reported as unit per liter (U/L) and a possible 20% dropout rate, 80 patients (40 in each treat- FSG, TG, CHOL, LDL and HDL levels were reported as milli- ment group) were recruited in this double blind random- grams per deciliter (mg/dl). Serum insulin concentration ized clinical trial.
was measured using a commercially available kit (Bio- vendor, Brno, Czech Republic). The kit for determination of insulin (IR-insulin) in serum was based on a sandwich According to a predefined computer-generated block enzyme immunoassay. The procedure was performed randomization table with a 1:1 allocation, each of the pa- according to the manufacturer's manual, as follows. 96- tients was assigned to pioglitazone or metformin treat- well plate was coated with guinea pig anti human insulin ment groups. A random allocation sequence was gener- antibody and insulin standard or samples were added to Hepat Mon. 2013;13(5):e9270 Razavizade M et al. Pioglitazone and Metformin in Fatty Liver the wells for their immunoreactions. After incubation or discontinuation of the medications did not occur dur- and plate washing, biotinylated guinea pig anti human ing the study period. Pill counts during the follow up vis- insulin antibody was introduced to the wells and the an- its discovered a good adherence to therapy with a mean tibody - antigen - labeled antibody complex was formed consumption of 89% of expected tablets (range from 84 on the surface of the well. After rinsing out the exces- sive labeled antibody, HRP labeled streptavidin (SA-HRP) were added to bind to the labeled antibody. Finally, HRP enzyme activity was determined by o-Phenylenediamine Table 1. Patients Characteristics According to the Treatment
dihydrochloride (OPD) and the concentration of insulin was calculated. Quantitative measurement of insulin Metformin Pioglitazone P value
resistance (IR) was performed using homeostasis model Age, ya, Mean ± SD
assessment-IR (HOMA-IR = fasting serum insulin x fasting Gender, No.
serum glucose/22.5) (29).
To minimize the laboratory errors, the whole assay was performed by the same operator from the beginning to the end, and room temperature, air humidity, incubator Obese, No.
temperature were strictly controlled. All the measure- ments were performed in duplicate. The intra-assay and Diabetes mellitus,
inter-assay coefficient variations were less than 10% and 12% respectively.
Weight at baseline,
Kga, Mean ± SD
3.11. Statistical Analysis Weight at 2 months, 78.67 ± 6.73
Kga, Mean ± SD
Data was summarized as means ± SD for continuous variables. The Kolmogorov-Smirnov test was used to Weight at 4 months, 77.59 ± 7.90
Kga, Mean ± SD
evaluate the normal distribution of the continuous vari- ables. Two-sample t-test was used to compare the mean Body mass index at 27.93 ± 2.28
values of continuous variables (laboratory, LFC, and an- baseline, Kg/m2a,
thropometric) between the metformin and pioglitazone Mean ± SD
groups. The chi-square test was used to compare categori- Body mass index at
cal variables between the treatment groups. The changes 2 months, Kg/m2a,
in laboratory and LFC mean values between baseline, 2 Mean ± SD
and 4 months study period in each treatment group were Body mass index at
calculated by paired t-test. The differences in laboratory 4 months, Kg/m2a,
and LFC mean value changes between (metformin and Mean ± SD
pioglitazone) groups were tested by two-sample t-test. a The values of age, weight, and body mass index are expressed as mean Non-parametric methods were used for non-normally ± standard deviation distributed values. The statistical analyses were per- formed using SPSS version 17 (SPSS, Chicago, IL, USA). The probability of the difference between the dependent and 4.3. Study Findings independent variables were considered significant if a Serum AST, ALT, ALP, FSG, TG, CHOL, and LDL concentra- two-tailed P value was less than 0.05.
tions were normally distributed at the baseline. (Z score = 1.6, 1.15, 1.23, 1.57, 1.7, 0.55, and 0.76 respectively, all P val- 4. Results
ues > 0.05) Patients' characteristics in treatment groups are shown in Table 1. The mean values of serum AST, ALT, 4.1. Patients Enrolled ALP, TG, CHOL, LDL, HDL, FSG, HOMA-IR, and LFC at base- Between January 2011 to January 2012, 93 patients sus- line, two and four months post-treatment according to pected of having NAFLD were evaluated. Eighty patients treatment groups are provided in Table 2. The mean age, with mean age of 35.27 (± 7.98) years were enrolled in the weight, BMI, laboratory values, LFC, and gender were sim- study. Reasons for exclusion were patient's unwilling- ilar between the treatment groups at baseline and dur- ness to participate in the study (n = 7), normalization of ing the study period. The comparisons of body weight, ALT during the lead-in phase (n = 4), renal failure (n = 1), laboratory parameters, and LFC within the treatment and pregnancy (n = 1). The frequency of participants with groups along the study period are shown in Table 3. Mean diabetes mellitus was six and twelve participants had im- body weight, serum AST, ALT, ALP, FSG, CHOL, LDL, HOMA- paired fasting glucose.
IR values, and LFC were significantly higher, but mean serum HDL level was lower at baseline than four months 4.2. Medication Adverse Effects and Compliance post-treatment in both treatment groups. The compari- Significant side effects that need a decrease in the dose son of mean body weight, laboratory parameters, and Hepat Mon. 2013;13(5):e9270 Pioglitazone and Metformin in Fatty Liver Razavizade M et al. LFC changes from baseline to two and four months post- and LFC from baseline to four months post-treatment treatment, and from two to four months post-treatment were not significantly different between the treatment according to the treatment groups are provided in Table 4. The changes in mean body weight, laboratory values, Table 2. Mean Values (Standard Deviation) for Laboratory Parameters and Liver fat Content at Baseline, 2 and 4 Months Post-Treat-
ment According to Treatment Group
Metformin, Mean ± SD
Pioglitazone, Mean ± SD
Pioglitazone, Mean ± SD
Aspartate Aminotransferase, U/L
Alanine Aminotransferase, U/L
Alkaline Phosphatase, U/L
Low Density Lipoprotein, mg/dl
High Density Lipoprotein, mg/dl
Fasting Plasma Glucose, mg/dl
Liver Fat Content, %
a Abbreviations: HOMA-IR: homeostasis model assessment-insulin resistance Hepat Mon. 2013;13(5):e9270 Razavizade M et al. Pioglitazone and Metformin in Fatty Liver Table 3. Intra-Group Comparisons of Mean Body Weight, Laboratory Parameters and Liver fat ContentDuring the Study Period
Baseline vs. 2 months, P value
Baseline vs. 4 months, P value 2 months vs. 4 months, P value
Body Weight, kg
Aspartate Aminotransferase, U/L
Alanine Aminotransferase, U/L
Alkaline Phosphatase, U/L
Low Density Lipoprotein, mg/dl
High Density Lipoprotein, mg/dl
Fasting Plasma Glucose, mg/dl
Liver Fat Content, %
a Abbreviations: HOMA-IR: homeostasis model assessment-insulin resistance administration of metformin (1.5 g/day) was effective for reduction of aminotransferase levels (31). However, met- The results of the present study showed that both treat- formin was not more effective than lifestyle intervention ment modalities were effective in reduction of serum in reducing aminotransferase levels. In a pilot open label levels of AST and ALT. Previous pilot trials demonstrated trial on fifteen NAFLD patients, Nair et al showed that met- that metformin (1-1.5 g/day) was effective for reduction of formin (20 mg/kg) reduced ALT and HOMA-IR in the first ALT levels (21 - 23). In a twelve-month clinical trial in fifty- three months of study (24). However, after three months, five NAFLD patients, it was shown that metformin (2 g/ there was no further decrease in insulin resistance and a day) was better than a prescriptive weight reducing diet rebound increase in ALT was observed. Improvement of for reduction of ALT levels (30). In a 24-month pilot study AST and ALT following thiazolidinediones was reported on sixty children with NAFLD, Nobili et al. reported that in the previous trials. The results of a one year open-label Hepat Mon. 2013;13(5):e9270 Pioglitazone and Metformin in Fatty Liver Razavizade M et al. Table 4. Comparison of Body Weight, Laboratory Parameters and Liver fat Content Changes (mean ± SD) From Baseline to 2 and 4
Months Post Treatment, and From 2 to 4 Months Post Treatment According to the Treatment Groups
Changes From Baseline to 2
Changes From Baseline to 4
Changes From 2 to 4 Months
Body weight, kg
0.03 1.18 ± 3.70 0.05 0.79 ± 1.76 0.07 13.74 ± 27.1 10.82 ± 17.06 0.56 3.14 ± 20.01 transferase, U/L
Alanine aminotrans- 23.51 ± 22.84 6.97 ± 11.82
0.03 37.52 ±40.70 21.75 ± 38.30 0.07 14.01 ± 31.95 2.22 ± 11.82 ferase, U/L
Alkaline phospha-
2.92 ± 10.94 0.00 11.42 ± 13.59 tase, U/L
Triglyceride, mg/dl
0.26 3.16 ± 68.56 0.92 -5.20 ± 43.24 25.9 ± 16.31 0.37 -4.28 ± 14.36 -7.5 ± 16.41 Low density lipopro- 3.92 ± 12.54
-2.38 ± 18.65 0.33 0.58 2.6 ± 11.83 10.13 ± 38.52 0.24 tein, mg/dl
High density lipo-
0.60 -7.08 ± 6.89 0.15 -5.39 ± 7.22 protein, mg/dl
Fasting plasma
0.30 5.35 ± 5.32 0.19 0.42 ± 7.39 Liver fat content, %
a Abbreviations: HOMA-IR: Homeostasis model assessment-insulin resistance b Negative values represent the increase of the parameter at that interval study on sixty-three NAFLD patients showed that admin- or type two diabetes mellitus showed that pioglitazone istration of rosiglitazone (4 mg/day for the first month (45 mg/day) was more effective than the placebo in reduc- and 8 mg/day thereafter) was more effective than placebo tion of serum ALT levels after six months (17). Bajaj et al in reduction of aminotransferase level (32). Promrat et al. studied the effect of pioglitazone (45 mg/day) in fourteen investigated the effect of pioglitazone (30 mg/day) on patients with type 2 diabetes mellitus (33). After sixteen eighteen NAFLD patients for forty-eight weeks (13). The weeks, the mean ALT level was significantly lower than mean ALT level was significantly lower at the end of study the baseline value (22 ± 2 vs. 28 ± 3 U/L; P = 0.02). Tiikkain- (40 ± 25 U/L) compared to the baseline value (99 ± 71 U/L) en et al. compared the effects of rosiglitazone (8 mg/day) (P <0.001). Sanyal et al. evaluated the ALT level in twenty and metformin (2 g/day) on serum ALT levels in twenty non-diabetic NAFLD patients who were assigned to vita- type 2 diabetic patients (34). After four months, serum min E (400 IU/ day) alone or the combination of vitamin ALT level decreased in the rosiglitazone group but re- E (400 IU/ day) and pioglitazone (30 mg/day) (14). In this mained unchanged in the metformin group. The effect of six-month pilot study, serum ALT levels were normalized insulin sensitizers on the improvement of serum amino- for all patients of both treatment groups. Lutchman et al. transferase levels seems encouraging in the previous evaluated the effect of pioglitazone (30 mg/day) in twen- studies. The existence of controversy in their effect on se- ty-one NAFLD patients for forty-eight weeks (15). After for- rum aminotransferase levels and their sustained effect ty-eight weeks of pioglitazone therapy, the mean serum after discontinuation are the major drawbacks that need ALT level was significantly lower than the baseline values complementary investigations. We suggest that the dif- (75.7 ± 34.7 vs. 34 ± 12.7 U/L; P < 0.001). Forty-eight weeks ferences in the results of the above mentioned trials after discontinuation of pioglitazone, the mean serum might be due to the variations in the studies duration, ALT level significantly increased (69.5 ± 38.7 U/L). This patient's heterogeneity (considering weight, gender, dai- study concluded that in NAFLD patients, long-term ad- ly physical activity, and the level of IR), and the medica- ministration of pioglitazone might be necessary to main- tion doses. In this study, the serum ALP was decreased tain the obtained results following treatment. In the significantly from baseline to four months post-treat- study of Aithal et al., seventy-four non-diabetic NAFLD ment in both treatment groups. An improvement of se- patients were assigned to take either pioglitazone (30 rum ALP level was reported following pioglitazone treat- mg/day) or placebo for one year (16). In the pioglitazone ment (13 , 14). These observations are in concordance with group, the mean serum ALT level was significantly lower our study. However, there was an increase in serum ALP at the end of the study compared to the baseline value level following metformin treatment in the study by Nair (55.9 ± 25.7 vs. 93.6 ± 61.3 U/L; P < 0.001). A study on fifty- et al. (24). It should be noted that total amounts of ALP five patients with NAFLD and impaired glucose tolerance cannot be considered as a specific marker for liver dis- Hepat Mon. 2013;13(5):e9270 Razavizade M et al. Pioglitazone and Metformin in Fatty Liver ease. The hepatic isoenzyme is a more reliable marker for this study. This shortcoming might explain the absence the detection of liver disease. Therefore, the rise of total of weight gain in the pioglitazone group of our study. Ob- serum ALP following administration of metformin in the taining an ideal body weight is the mainstay of treatment above-mentioned study might be related to conditions in NAFLD. Hypocaloric diet and administration of appro- unrelated to NAFLD. Documentation of this observation priate daily physical activity help the NAFLD patient re- with concomitant checking of ALP hepatic isoenzyme or duce their body weight. In this study, a dietitian super- gamma glutamyl transpeptidase seems a reasonable ap- vised participants' daily calorie intake. The dietitian was proach. In this study, the serum metabolic parameters blinded to the treatment groups and controlled daily (FSG, HOMA-IR, CHOL, LDL, and HDL) and LFC were im- calorie intake in follow up visits. The protocol used for proved significantly from baseline to four months post- diet in our study was based on "Guidelines for the diagno- treatment in both treatment groups. These findings are sis and management of nonalcoholic fatty liver disease: in accordance with the previous trials (13 - 17 , 21 - 24). In update 2010" (9). Hypo caloric diets resulted in weight the present study, serum TG level did not change signifi- reduction in both treatment groups from baseline to the cantly in both treatment groups during the study period. end of study (Table 3). Meanwhile, body weight reduction This finding is in accordance with the results of previous was not statistically different between treatment groups trials that reported serum TG level did not change follow- (Table 4). The above findings show that hypocaloric diets ing pioglitazone prescription in NAFLD patients (13 , 16 , resulted in a significant reduction of body weight in par- 17). The results of previous trials on the effect of metfor- ticipants in each treatment groups; Moreover, body min on serum TG level in NAFLD patients are somewhat weight reduction that was due to the hypocaloric diet, controversial. Magalotti et al. studied the effect of metfor- was not statistically different in treatment groups. There- min (1.5 g/day) in eleven NAFLD patients for six months fore, it can be concluded that diet was not different in (21). The mean serum TG level did not change significant- treatment groups and might not have biased the results. ly from baseline (176 ± 104 mg/dl) to the end of treatment We should note the role of metformin in NAFLD based on (142 ± 61 mg/dl). Marchesini et al. studied the effect of the recent guideline (2012) from the American Associa- metformin (1.5 g/day) in twenty NAFLD patients for four tion for the Study of Liver Diseases, American College of months (23). The mean serum TG level did not change sig- Gastroenterology, and the American Gastroenterological nificantly from baseline (1.91 ± 1.12 mmol/L) to the end of Association. It has been mentioned that metformin does treatment (1.92 ± 1.06 mmol/L) in the metformin group. not have significant effects on liver histology. It is not rec- In the study of Nair et al, serum TG level did not change ommended as a specific treatment for liver disease in significantly following a twelve month administration of adults with NAFLD (35). It is obvious that our study was metformin for NAFLD patients (24). Nobili et al. showed designed and conducted (in 2011) before the release of that metformin (1.5 g/day) reduced serum TG levels in this guideline.
NAFLD patients after 12 months (31). The differences in the results of the above-mentioned trials might be due to the 5.1. The Limitations of the Study variations in the studies duration, patients' heterogene- Serum aminotransferase levels seem to have fluctua- ity (considering weight, gender, daily physical activity, tions in the natural course of disease (even without any and the level of IR), medication doses, and possible con- therapy) in NAFLD patients (36). Therefore, the lack of comitant familial hyperlipidemia. To the best of our control group was a limitation of study. Serum ALT value knowledge, this is the first clinical trial on NAFLD pa- has limitations in predicting NAFLD (37). Although the tients that compared the effect of pioglitazone and met- participants were presumed to have NASH according to formin on liver function tests (including AST, ALT, and the study protocol, the diagnosis of NASH was not con- ALP), metabolic profile (including FSG, HOMA-IR, and lip- firmed by liver histology. Therefore, the absence of liver id profile), and LFC. No statistically significant differences biopsy was another limitation of this study in document- were seen between the treatment groups with regard to ing NASH patients. Exercise and weight reduction are the the changes of laboratory parameters and LFC from base- factors that influence the treatment of NAFLD. The lack of line to four months post-treatment. It seems that metfor- control on participants' daily exercise in this study might min and pioglitazone might equally affect the liver func- be considered as another limitation. The duration of tion tests, HOMA-IR, CHOL, LDL, HDL, and LFC in NAFLD study period was rather short, the sample size was small, patients. Weight gain after pioglitazone treatment was and the sustained effects of these medications were not reported in previous studies (13 , 15 - 17). In this study, evaluated after discontinuation of treatment.
there was no significant weight change in the partici- pants receiving pioglitazone during the study period. 5.2. Recommendation for Future Studies Our finding was similar to that of Sanyal et al. that report- ed BMI did not increase significantly in subjects receiving Since pioglitazone and metformin reduce the IR by dif- pioglitazone (14). Although the daily calorie intake of the ferent mechanisms, combined therapy with these drugs participants was checked by a dietitian during the follow may show more improvement in aminotransferase con- up visits, but daily physical activity was not monitored in centrations than either of them alone. Comparing the Hepat Mon. 2013;13(5):e9270 Pioglitazone and Metformin in Fatty Liver Razavizade M et al. effects of combination therapy with pioglitazone and 6. Kelishadi R, Poursafa P. Obesity and air pollution: global risk fac- metformin versus pioglitazone alone on histologic and tors for pediatric non-alcoholic fatty liver disease. Hepat Mon. biochemical changes in NAFLD patients with longer fol- 7. Jamali R, Khonsari M, Merat S, Khoshnia M, Jafari E, Bahram Kalh- low up duration together with well-established control ori A, et al. Persistent alanine aminotransferase elevation among groups is recommended.
the general Iranian population: prevalence and causes. World J 5.3. Summary 8. Dai HF, Shen Z, Yu CH, Zhang XC, Li YM. Epidemiology of fatty liver in an islander population of China: a population-based This double blind randomized clinical trial was per- case-control study. Hepatobiliary Pancreat Dis Int. 2008;7(4):373-8.
9. Fan JG, Jia JD, Li YM, Wang BY, Lu LG, Shi JP, et al. Guidelines for the formed on eighty NAFLD patients who were randomly diagnosis and management of nonalcoholic fatty liver disease: assigned to metformin and pioglitazone treatments. update 2010: (published in Chinese on Chinese Journal of Hepa- Administration of either medication significantly led tology 2010; 18:163-166). J Dig Dis. 2011;12(1):38-44.
10. Paradis V, Perlemuter G, Bonvoust F, Dargere D, Parfait B, Vidaud to a decrease in the serum AST, ALT, ALP, FSG, CHOL, LDL, M, et al. High glucose and hyperinsulinemia stimulate connec- HOMA-IR, and LFC, and an increase in the serum HDL. tive tissue growth factor expression: a potential mechanism in- No statistically significant differences were observed volved in progression to fibrosis in nonalcoholic steatohepatitis. between the treatment groups with regards to changes Hepatology. 2001;34(4 Pt 1):738-44.
11. Gross B, Staels B. PPAR agonists: multimodal drugs for the treat- of laboratory parameters and LFC from baseline to four ment of type-2 diabetes. Best Pract Res Clin Endocrinol Metab. months post-treatment. It seems that metformin and pioglitazone might equally affect liver function tests, 12. Maeda N, Takahashi M, Funahashi T, Kihara S, Nishizawa H, Kishi- HOMA-IR, CHOL, LDL, HDL, and LFC in NAFLD patients. da K, et al. PPARgamma ligands increase expression and plasma concentrations of adiponectin, an adipose-derived protein. Dia- These medications could be suggested as a safe treat- ment option for the management of NAFLD patients.
13. Promrat K, Lutchman G, Uwaifo GI, Freedman RJ, Soza A, Heller T, et al. A pilot study of pioglitazone treatment for nonalcoholic 14. Sanyal AJ, Mofrad PS, Contos MJ, Sargeant C, Luketic VA, Sterling This study was supported by the research funds of RK, et al. A pilot study of vitamin E versus vitamin E and piogli- tazone for the treatment of nonalcoholic steatohepatitis. Clin Kashan University of Medical Sciences (No: 29-5-1-2851). The authors extend their gratitude to Dr Arsia Jamali, 15. Lutchman G, Modi A, Kleiner DE, Promrat K, Heller T, Ghany M, Dr Vafa Rahimi movaghar, and Dr Neda Moslemi from et al. The effects of discontinuing pioglitazone in patients with Tehran University of Medical Sciences, for reviewing the 16. 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. Gastro- 17. Belfort R, Harrison SA, Brown K, Darland C, Finch J, Hardies J, et al. All of the authors have contributed to different parts of A placebo-controlled trial of pioglitazone in subjects with nonal- the research.
coholic steatohepatitis. N Engl J Med. 2006;355(22):2297-307.
18. Al-Gharabally A, O'Brien CB, Acosta RC. A pilot study of piogli- tazone for the treatment of non-alcoholic fatty liver disease. Hep- at Mon. 2007;7(3):131-7.
19. Scarpello JH, Howlett HC. Metformin therapy and clinical uses. There is no conflict of interest.
Diab Vasc Dis Res. 2008;5(3):157-67.
20. Natali A, Ferrannini E. Effects of metformin and thiazolidinedio- nes on suppression of hepatic glucose production and stimula- tion of glucose uptake in type 2 diabetes: a systematic review. This study was financially supported by Kashan Univer- 21. Magalotti D, Marchesini G, Ramilli S, Berzigotti A, Bianchi G, sity of Medical Sciences.
Zoli M. Splanchnic haemodynamics in non-alcoholic fatty liver disease: effect of a dietary/pharmacological treatment. A pilot study. Dig Liver Dis. 2004;36(6):406-11.
22. Schwimmer JB, Middleton MS, Deutsch R, Lavine JE. A phase 2 1. Duan XY, Qiao L, Fan JG. Clinical features of nonalcoholic fatty clinical trial of metformin as a treatment for non-diabetic pae- liver disease-associated hepatocellular carcinoma. Hepatobiliary diatric non-alcoholic steatohepatitis. Aliment Pharmacol Ther. Pancreat Dis Int. 2012;11(1):18-27.
2. Butt AS, Abbas Z, Jafri W. Hepatocellular carcinoma in pakistan: 23. Marchesini G, Brizi M, Bianchi G, Tomassetti S, Zoli M, Mel- where do we stand? Hepat Mon. 2012;12(10 HCC).
chionda N. Metformin in non-alcoholic steatohepatitis. Lancet. 3. Fan JG, Peng YD. Metabolic syndrome and non-alcoholic fatty liver disease: Asian definitions and Asian studies. Hepatobiliary 24. Nair S, Diehl AM, Wiseman M, Farr GH, Jr, Perrillo RP. Metformin Pancreat Dis Int. 2007;6(6):572-8.
in the treatment of non-alcoholic steatohepatitis: a pilot open 4. Jin HB, Gu ZY, Yu CH, Li YM. Association of nonalcoholic fatty liver label trial. Aliment Pharmacol Ther. 2004;20(1):23-8.
disease with type 2 diabetes: clinical features and independent 25. Ghamar-Chehreh ME, Khedmat H, Amini M, Taheri S. Predictive risk factors in diabetic fatty liver patients. Hepatobiliary Pancreat Factors for Ultrasonographic Grading of Nonalcoholic Fatty Liv- Dis Int. 2005;4(3):389-92.
er Disease. Hepat Mon. 2012;12(11).
5. Gharouni M, Rashidi A. Association between fatty liver and coro- 26. Razavizade M, Jamali R, Arj A, Talari H. Serum parameters predict nary artery disease: Yet to explore. Hepat Mon. 2007;2007(4, Au-
the severity of ultrasonographic findings in non-alcoholic fatty liver disease. Hepatobiliary Pancreat Dis Int. 2012;11(5):513-20.
Hepat Mon. 2013;13(5):e9270 Razavizade M et al. Pioglitazone and Metformin in Fatty Liver 27. Jamali R, Pourshams A, Amini S, Deyhim MR, Rezvan H, Male- 33. Bajaj M, Suraamornkul S, Pratipanawatr T, Hardies LJ, kzadeh R. The upper normal limit of serum alanine amino- Pratipanawatr W, Glass L, et al. Pioglitazone reduces hepatic fat transferase in Golestan Province, northeast Iran. Arch Iran Med. content and augments splanchnic glucose uptake in patients with type 2 diabetes. Diabetes. 2003;52(6):1364-70.
28. Kotronen A, Peltonen M, Hakkarainen A, Sevastianova K, Berg- 34. Tiikkainen M, Hakkinen AM, Korsheninnikova E, Nyman T, Maki- holm R, Johansson LM, et al. Prediction of non-alcoholic fatty mattila S, Yki-Jarvinen H. Effects of rosiglitazone and metformin liver disease and liver fat using metabolic and genetic factors. on liver fat content, hepatic insulin resistance, insulin clearance, and gene expression in adipose tissue in patients with type 2 dia- 29. Wallace TM, Levy JC, Matthews DR. Use and abuse of HOMA mod- eling. Diabetes Care. 2004;27(6):1487-95.
35. Chalasani N, Younossi Z, Lavine JE, Diehl AM, Brunt EM, Cusi K, 30. Bugianesi E, Gentilcore E, Manini R, Natale S, Vanni E, Villanova N, et al. The diagnosis and management of non-alcoholic fatty liver et al. A randomized controlled trial of metformin versus vitamin disease: Practice guideline by the American Association for the E or prescriptive diet in nonalcoholic fatty liver disease. Am J Gas- Study of Liver Diseases, American College of Gastroenterology, and the American Gastroenterological Association. Am J Gastro- 31. Nobili V, Manco M, Ciampalini P, Alisi A, Devito R, Bugianesi E, et al. Metformin use in children with nonalcoholic fatty liver 36. Ipekci SH, Basaranoglu M, Sonsuz A. The fluctuation of serum disease: an open-label, 24-month, observational pilot study. Clin levels of aminotransferase in patients with nonalcoholic steato- hepatitis. J Clin Gastroenterol. 2003;36(4):371.
32. Ratziu V, Giral P, Jacqueminet S, Charlotte F, Hartemann-Heurtier 37. Khosravi S, Alavian SM, Zare A, Daryani NE, Fereshtehnejad SM, A, Serfaty L, et al. Rosiglitazone for nonalcoholic steatohepatitis: Keramati MR, et al. Non-alcoholic fatty liver disease and corre- one-year results of the randomized placebo-controlled Fatty Liv- lation of serum alanin aminotransferase level with histopatho- er Improvement with Rosiglitazone Therapy (FLIRT) Trial. Gastro- logic findings. Hepat Mon. 2011;11(6):452-8.
Hepat Mon. 2013;13(5):e9270

Source: http://hepatmon.com/19866.pdf